diff --git a/clang/include/clang/Basic/riscv_vector.td b/clang/include/clang/Basic/riscv_vector.td
--- a/clang/include/clang/Basic/riscv_vector.td
+++ b/clang/include/clang/Basic/riscv_vector.td
@@ -1854,7 +1854,7 @@
 
 // 12.13. Vector Single-Width Integer Multiply-Add Instructions
 let UnMaskedPolicyScheme = HasPolicyOperand,
-    IsPrototypeDefaultTU = true in {
+    IsPrototypeDefaultTU = false in {
 defm vmacc  : RVVIntTerBuiltinSet;
 defm vnmsac : RVVIntTerBuiltinSet;
 defm vmadd  : RVVIntTerBuiltinSet;
@@ -1967,7 +1967,7 @@
 
 // 14.6. Vector Single-Width Floating-Point Fused Multiply-Add Instructions
 let UnMaskedPolicyScheme = HasPolicyOperand,
-    IsPrototypeDefaultTU = true in {
+    IsPrototypeDefaultTU = false in {
 defm vfmacc  : RVVFloatingTerBuiltinSet;
 defm vfnmacc : RVVFloatingTerBuiltinSet;
 defm vfmsac  : RVVFloatingTerBuiltinSet;
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmacc.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmacc_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmacc_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmacc_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmacc_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmacc_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmacc_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmacc_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmacc_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmacc_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmacc_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmacc_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmacc_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmacc_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmacc_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmacc_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmacc_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmacc_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmacc_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmadd.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmadd.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmadd.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmadd_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmadd_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmadd_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmadd_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmadd_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmadd_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmadd_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmadd_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmadd_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmadd_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmadd_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmadd_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmadd_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmadd_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmadd_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmadd_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmadd_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmadd_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsac.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsac_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsac_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsac_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsac_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsac_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsac_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsac_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsac_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsac_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsac_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsac_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsac_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsac_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsac_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsac_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsac_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsac_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsac_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsub.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsub.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfmsub.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsub_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsub_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsub_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsub_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsub_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsub_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsub_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsub_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsub_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsub_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsub_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsub_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsub_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsub_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsub_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsub_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsub_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsub_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmacc.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmacc_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmacc_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmacc_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmacc_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmacc_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmacc_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmacc_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmacc_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmacc_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmacc_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmacc_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmacc_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmacc_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmacc_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmacc_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmacc_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmacc_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmacc_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmadd.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmadd.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmadd.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmadd_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmadd_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmadd_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmadd_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmadd_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmadd_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmadd_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmadd_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmadd_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmadd_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmadd_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmadd_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmadd_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmadd_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmadd_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmadd_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmadd_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmadd_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsac.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsac_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsac_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsac_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsac_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsac_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsac_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsac_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsac_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsac_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsac_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsac_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsac_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsac_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsac_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsac_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsac_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsac_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsac_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsub.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsub.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfnmsub.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsub_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsub_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsub_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsub_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsub_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsub_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsub_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsub_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsub_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -97,7 +97,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsub_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -107,7 +107,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsub_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -117,7 +117,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsub_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -127,7 +127,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsub_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -137,7 +137,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsub_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -147,7 +147,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsub_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -157,7 +157,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsub_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -167,7 +167,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsub_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsub_vf_f64m8(vfloat64m8_t acc, double op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmacc.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmacc_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmacc_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmacc_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmacc_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmacc_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmacc_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmacc_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmacc_vf_f64m8(vfloat64m8_t acc, float op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwmsac.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmsac_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmsac_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmsac_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmsac_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmsac_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmsac_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmsac_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmsac_vf_f64m8(vfloat64m8_t acc, float op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmacc.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmacc_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmacc_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmacc_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmacc_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmacc_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmacc_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmacc_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmacc_vf_f64m8(vfloat64m8_t acc, float op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vfwnmsac.c
@@ -7,7 +7,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmsac_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -17,7 +17,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmsac_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -27,7 +27,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmsac_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -37,7 +37,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmsac_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -47,7 +47,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmsac_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -57,7 +57,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmsac_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -67,7 +67,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmsac_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -77,7 +77,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmsac_vf_f64m8(vfloat64m8_t acc, float op1,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmacc.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmacc_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmacc_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmacc_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmacc_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmacc_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmacc_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmacc_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmacc_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmacc_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmacc_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmacc_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmacc_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmacc_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmacc_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmacc_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmacc_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmacc_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmacc_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmacc_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmacc_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmacc_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmacc_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmacc_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmacc_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmacc_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmacc_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmacc_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmacc_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmacc_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmacc_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmacc_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmacc_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmacc_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmacc_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmacc_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmacc_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmacc_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmacc_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmacc_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmacc_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmacc_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmacc_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmacc_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmacc_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmacc_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmacc_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmacc_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmacc_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmacc_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmacc_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmacc_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmacc_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmacc_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmacc_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmacc_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmacc_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmacc_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmacc_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmacc_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmacc_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmacc_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmacc_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmacc_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmacc_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmacc_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmacc_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmacc_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmacc_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmacc_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmacc_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmacc_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmacc_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmacc_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmacc_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmacc_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmacc_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmacc_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmacc_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmacc_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmacc_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmacc_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmacc_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmacc_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmacc_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmacc_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmacc_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmacc_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmacc_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmadd.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmadd.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vmadd.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmadd_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmadd_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmadd_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmadd_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmadd_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmadd_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmadd_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmadd_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmadd_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmadd_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmadd_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmadd_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmadd_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmadd_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmadd_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmadd_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmadd_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmadd_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmadd_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmadd_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmadd_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmadd_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmadd_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmadd_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmadd_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmadd_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmadd_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmadd_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmadd_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmadd_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmadd_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmadd_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmadd_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmadd_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmadd_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmadd_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmadd_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmadd_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmadd_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmadd_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmadd_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmadd_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmadd_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmadd_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmadd_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmadd_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmadd_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmadd_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmadd_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmadd_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmadd_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmadd_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmadd_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmadd_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmadd_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmadd_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmadd_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmadd_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmadd_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmadd_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmadd_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmadd_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmadd_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmadd_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmadd_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmadd_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmadd_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmadd_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmadd_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmadd_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmadd_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmadd_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmadd_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmadd_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmadd_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmadd_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmadd_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmadd_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmadd_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmadd_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmadd_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmadd_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmadd_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmadd_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmadd_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmadd_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmadd_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmadd_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsac.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsac_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsac_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsac_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsac_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsac_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsac_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsac_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsac_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsac_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsac_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsac_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsac_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsac_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsac_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsac_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsac_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsac_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsac_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsac_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsac_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsac_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsac_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsac_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsac_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsac_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsac_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsac_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsac_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsac_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsac_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsac_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsac_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsac_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsac_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsac_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsac_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsac_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsac_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsac_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsac_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsac_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsac_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsac_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsac_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsac_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsac_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsac_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsac_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsac_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsac_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsac_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsac_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsac_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsac_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsac_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsac_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsac_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsac_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsac_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsac_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsac_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsac_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsac_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsac_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsac_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsac_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsac_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsac_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsac_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsac_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsac_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsac_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsac_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsac_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsac_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsac_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsac_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsac_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsac_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsac_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsac_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsac_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsac_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsac_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsac_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsac_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsac_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsac_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsub.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsub.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vnmsub.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsub_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsub_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsub_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsub_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsub_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsub_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsub_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsub_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsub_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsub_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsub_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsub_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsub_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsub_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsub_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsub_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsub_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsub_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsub_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsub_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsub_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsub_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsub_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsub_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsub_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsub_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsub_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsub_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsub_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsub_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsub_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsub_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsub_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsub_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsub_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsub_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsub_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsub_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsub_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsub_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsub_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsub_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsub_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsub_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsub_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsub_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsub_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsub_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsub_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsub_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsub_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsub_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsub_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsub_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsub_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsub_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsub_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsub_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsub_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsub_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsub_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsub_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsub_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsub_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsub_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsub_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsub_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsub_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsub_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsub_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsub_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsub_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsub_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsub_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsub_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsub_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsub_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsub_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsub_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsub_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsub_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsub_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsub_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsub_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsub_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsub_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsub_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsub_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vwmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vwmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics-overloaded/vwmacc.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmacc_vv_i16mf4(vint16mf4_t acc, vint8mf8_t op1,
@@ -16,7 +16,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmacc_vx_i16mf4(vint16mf4_t acc, int8_t op1, vint8mf8_t op2,
@@ -26,7 +26,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmacc_vv_i16mf2(vint16mf2_t acc, vint8mf4_t op1,
@@ -36,7 +36,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmacc_vx_i16mf2(vint16mf2_t acc, int8_t op1, vint8mf4_t op2,
@@ -46,7 +46,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmacc_vv_i16m1(vint16m1_t acc, vint8mf2_t op1, vint8mf2_t op2,
@@ -56,7 +56,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmacc_vx_i16m1(vint16m1_t acc, int8_t op1, vint8mf2_t op2,
@@ -66,7 +66,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmacc_vv_i16m2(vint16m2_t acc, vint8m1_t op1, vint8m1_t op2,
@@ -76,7 +76,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmacc_vx_i16m2(vint16m2_t acc, int8_t op1, vint8m1_t op2,
@@ -86,7 +86,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmacc_vv_i16m4(vint16m4_t acc, vint8m2_t op1, vint8m2_t op2,
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmacc_vx_i16m4(vint16m4_t acc, int8_t op1, vint8m2_t op2,
@@ -106,7 +106,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmacc_vv_i16m8(vint16m8_t acc, vint8m4_t op1, vint8m4_t op2,
@@ -116,7 +116,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmacc_vx_i16m8(vint16m8_t acc, int8_t op1, vint8m4_t op2,
@@ -126,7 +126,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmacc_vv_i32mf2(vint32mf2_t acc, vint16mf4_t op1,
@@ -136,7 +136,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmacc_vx_i32mf2(vint32mf2_t acc, int16_t op1, vint16mf4_t op2,
@@ -146,7 +146,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmacc_vv_i32m1(vint32m1_t acc, vint16mf2_t op1,
@@ -156,7 +156,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmacc_vx_i32m1(vint32m1_t acc, int16_t op1, vint16mf2_t op2,
@@ -166,7 +166,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmacc_vv_i32m2(vint32m2_t acc, vint16m1_t op1, vint16m1_t op2,
@@ -176,7 +176,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmacc_vx_i32m2(vint32m2_t acc, int16_t op1, vint16m1_t op2,
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmacc_vv_i32m4(vint32m4_t acc, vint16m2_t op1, vint16m2_t op2,
@@ -196,7 +196,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmacc_vx_i32m4(vint32m4_t acc, int16_t op1, vint16m2_t op2,
@@ -206,7 +206,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmacc_vv_i32m8(vint32m8_t acc, vint16m4_t op1, vint16m4_t op2,
@@ -216,7 +216,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmacc_vx_i32m8(vint32m8_t acc, int16_t op1, vint16m4_t op2,
@@ -226,7 +226,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmacc_vv_i64m1(vint64m1_t acc, vint32mf2_t op1,
@@ -236,7 +236,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmacc_vx_i64m1(vint64m1_t acc, int32_t op1, vint32mf2_t op2,
@@ -246,7 +246,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmacc_vv_i64m2(vint64m2_t acc, vint32m1_t op1, vint32m1_t op2,
@@ -256,7 +256,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmacc_vx_i64m2(vint64m2_t acc, int32_t op1, vint32m1_t op2,
@@ -266,7 +266,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmacc_vv_i64m4(vint64m4_t acc, vint32m2_t op1, vint32m2_t op2,
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmacc_vx_i64m4(vint64m4_t acc, int32_t op1, vint32m2_t op2,
@@ -286,7 +286,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmacc_vv_i64m8(vint64m8_t acc, vint32m4_t op1, vint32m4_t op2,
@@ -296,7 +296,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmacc_vx_i64m8(vint64m8_t acc, int32_t op1, vint32m4_t op2,
@@ -306,7 +306,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vwmaccu_vv_u16mf4(vuint16mf4_t acc, vuint8mf8_t op1,
@@ -316,7 +316,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vwmaccu_vx_u16mf4(vuint16mf4_t acc, uint8_t op1,
@@ -326,7 +326,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vwmaccu_vv_u16mf2(vuint16mf2_t acc, vuint8mf4_t op1,
@@ -336,7 +336,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vwmaccu_vx_u16mf2(vuint16mf2_t acc, uint8_t op1,
@@ -346,7 +346,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vwmaccu_vv_u16m1(vuint16m1_t acc, vuint8mf2_t op1,
@@ -356,7 +356,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vwmaccu_vx_u16m1(vuint16m1_t acc, uint8_t op1, vuint8mf2_t op2,
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vwmaccu_vv_u16m2(vuint16m2_t acc, vuint8m1_t op1,
@@ -376,7 +376,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vwmaccu_vx_u16m2(vuint16m2_t acc, uint8_t op1, vuint8m1_t op2,
@@ -386,7 +386,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vwmaccu_vv_u16m4(vuint16m4_t acc, vuint8m2_t op1,
@@ -396,7 +396,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vwmaccu_vx_u16m4(vuint16m4_t acc, uint8_t op1, vuint8m2_t op2,
@@ -406,7 +406,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vwmaccu_vv_u16m8(vuint16m8_t acc, vuint8m4_t op1,
@@ -416,7 +416,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vwmaccu_vx_u16m8(vuint16m8_t acc, uint8_t op1, vuint8m4_t op2,
@@ -426,7 +426,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vwmaccu_vv_u32mf2(vuint32mf2_t acc, vuint16mf4_t op1,
@@ -436,7 +436,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vwmaccu_vx_u32mf2(vuint32mf2_t acc, uint16_t op1,
@@ -446,7 +446,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vwmaccu_vv_u32m1(vuint32m1_t acc, vuint16mf2_t op1,
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vwmaccu_vx_u32m1(vuint32m1_t acc, uint16_t op1,
@@ -466,7 +466,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vwmaccu_vv_u32m2(vuint32m2_t acc, vuint16m1_t op1,
@@ -476,7 +476,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vwmaccu_vx_u32m2(vuint32m2_t acc, uint16_t op1,
@@ -486,7 +486,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vwmaccu_vv_u32m4(vuint32m4_t acc, vuint16m2_t op1,
@@ -496,7 +496,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vwmaccu_vx_u32m4(vuint32m4_t acc, uint16_t op1,
@@ -506,7 +506,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vwmaccu_vv_u32m8(vuint32m8_t acc, vuint16m4_t op1,
@@ -516,7 +516,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vwmaccu_vx_u32m8(vuint32m8_t acc, uint16_t op1,
@@ -526,7 +526,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vwmaccu_vv_u64m1(vuint64m1_t acc, vuint32mf2_t op1,
@@ -536,7 +536,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vwmaccu_vx_u64m1(vuint64m1_t acc, uint32_t op1,
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vwmaccu_vv_u64m2(vuint64m2_t acc, vuint32m1_t op1,
@@ -556,7 +556,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vwmaccu_vx_u64m2(vuint64m2_t acc, uint32_t op1,
@@ -566,7 +566,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vwmaccu_vv_u64m4(vuint64m4_t acc, vuint32m2_t op1,
@@ -576,7 +576,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vwmaccu_vx_u64m4(vuint64m4_t acc, uint32_t op1,
@@ -586,7 +586,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vwmaccu_vv_u64m8(vuint64m8_t acc, vuint32m4_t op1,
@@ -596,7 +596,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vwmaccu_vx_u64m8(vuint64m8_t acc, uint32_t op1,
@@ -606,7 +606,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmaccsu_vv_i16mf4(vint16mf4_t acc, vint8mf8_t op1,
@@ -616,7 +616,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmaccsu_vx_i16mf4(vint16mf4_t acc, int8_t op1,
@@ -626,7 +626,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmaccsu_vv_i16mf2(vint16mf2_t acc, vint8mf4_t op1,
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmaccsu_vx_i16mf2(vint16mf2_t acc, int8_t op1,
@@ -646,7 +646,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmaccsu_vv_i16m1(vint16m1_t acc, vint8mf2_t op1,
@@ -656,7 +656,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmaccsu_vx_i16m1(vint16m1_t acc, int8_t op1, vuint8mf2_t op2,
@@ -666,7 +666,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmaccsu_vv_i16m2(vint16m2_t acc, vint8m1_t op1, vuint8m1_t op2,
@@ -676,7 +676,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmaccsu_vx_i16m2(vint16m2_t acc, int8_t op1, vuint8m1_t op2,
@@ -686,7 +686,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmaccsu_vv_i16m4(vint16m4_t acc, vint8m2_t op1, vuint8m2_t op2,
@@ -696,7 +696,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmaccsu_vx_i16m4(vint16m4_t acc, int8_t op1, vuint8m2_t op2,
@@ -706,7 +706,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmaccsu_vv_i16m8(vint16m8_t acc, vint8m4_t op1, vuint8m4_t op2,
@@ -716,7 +716,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmaccsu_vx_i16m8(vint16m8_t acc, int8_t op1, vuint8m4_t op2,
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmaccsu_vv_i32mf2(vint32mf2_t acc, vint16mf4_t op1,
@@ -736,7 +736,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmaccsu_vx_i32mf2(vint32mf2_t acc, int16_t op1,
@@ -746,7 +746,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmaccsu_vv_i32m1(vint32m1_t acc, vint16mf2_t op1,
@@ -756,7 +756,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmaccsu_vx_i32m1(vint32m1_t acc, int16_t op1, vuint16mf2_t op2,
@@ -766,7 +766,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmaccsu_vv_i32m2(vint32m2_t acc, vint16m1_t op1,
@@ -776,7 +776,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmaccsu_vx_i32m2(vint32m2_t acc, int16_t op1, vuint16m1_t op2,
@@ -786,7 +786,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmaccsu_vv_i32m4(vint32m4_t acc, vint16m2_t op1,
@@ -796,7 +796,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmaccsu_vx_i32m4(vint32m4_t acc, int16_t op1, vuint16m2_t op2,
@@ -806,7 +806,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmaccsu_vv_i32m8(vint32m8_t acc, vint16m4_t op1,
@@ -816,7 +816,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmaccsu_vx_i32m8(vint32m8_t acc, int16_t op1, vuint16m4_t op2,
@@ -826,7 +826,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmaccsu_vv_i64m1(vint64m1_t acc, vint32mf2_t op1,
@@ -836,7 +836,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmaccsu_vx_i64m1(vint64m1_t acc, int32_t op1, vuint32mf2_t op2,
@@ -846,7 +846,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmaccsu_vv_i64m2(vint64m2_t acc, vint32m1_t op1,
@@ -856,7 +856,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmaccsu_vx_i64m2(vint64m2_t acc, int32_t op1, vuint32m1_t op2,
@@ -866,7 +866,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmaccsu_vv_i64m4(vint64m4_t acc, vint32m2_t op1,
@@ -876,7 +876,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmaccsu_vx_i64m4(vint64m4_t acc, int32_t op1, vuint32m2_t op2,
@@ -886,7 +886,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmaccsu_vv_i64m8(vint64m8_t acc, vint32m4_t op1,
@@ -896,7 +896,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmaccsu_vx_i64m8(vint64m8_t acc, int32_t op1, vuint32m4_t op2,
@@ -906,7 +906,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccus.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccus.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmaccus_vx_i16mf4(vint16mf4_t acc, uint8_t op1,
@@ -916,7 +916,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccus.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccus.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmaccus_vx_i16mf2(vint16mf2_t acc, uint8_t op1,
@@ -926,7 +926,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccus.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccus.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmaccus_vx_i16m1(vint16m1_t acc, uint8_t op1, vint8mf2_t op2,
@@ -936,7 +936,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccus.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccus.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmaccus_vx_i16m2(vint16m2_t acc, uint8_t op1, vint8m1_t op2,
@@ -946,7 +946,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccus.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccus.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmaccus_vx_i16m4(vint16m4_t acc, uint8_t op1, vint8m2_t op2,
@@ -956,7 +956,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccus.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccus.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmaccus_vx_i16m8(vint16m8_t acc, uint8_t op1, vint8m4_t op2,
@@ -966,7 +966,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccus.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccus.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmaccus_vx_i32mf2(vint32mf2_t acc, uint16_t op1,
@@ -976,7 +976,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccus.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccus.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmaccus_vx_i32m1(vint32m1_t acc, uint16_t op1, vint16mf2_t op2,
@@ -986,7 +986,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccus.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccus.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmaccus_vx_i32m2(vint32m2_t acc, uint16_t op1, vint16m1_t op2,
@@ -996,7 +996,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccus.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccus.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmaccus_vx_i32m4(vint32m4_t acc, uint16_t op1, vint16m2_t op2,
@@ -1006,7 +1006,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccus.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccus.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmaccus_vx_i32m8(vint32m8_t acc, uint16_t op1, vint16m4_t op2,
@@ -1016,7 +1016,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccus.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccus.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmaccus_vx_i64m1(vint64m1_t acc, uint32_t op1, vint32mf2_t op2,
@@ -1026,7 +1026,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccus.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccus.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmaccus_vx_i64m2(vint64m2_t acc, uint32_t op1, vint32m1_t op2,
@@ -1036,7 +1036,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccus.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccus.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmaccus_vx_i64m4(vint64m4_t acc, uint32_t op1, vint32m2_t op2,
@@ -1046,7 +1046,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccus.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccus.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmaccus_vx_i64m8(vint64m8_t acc, uint32_t op1, vint32m4_t op2,
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmacc.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmacc_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmacc_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmacc_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmacc_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmacc_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmacc_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmacc_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmacc_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmacc_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmacc_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmacc_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmacc_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmacc_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmacc_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmacc_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmacc_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmacc_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmacc_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -377,7 +377,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmacc.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmacc.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmacc_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -386,7 +386,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmacc.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmacc.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmacc_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -395,7 +395,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmacc.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmacc.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmacc_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -404,7 +404,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmacc.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmacc.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmacc_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -413,7 +413,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmacc.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmacc.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmacc_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -422,7 +422,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmacc.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmacc.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmacc_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -431,7 +431,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmacc.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmacc.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmacc_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -440,7 +440,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmacc.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmacc.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmacc_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -449,7 +449,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmacc.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmacc.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmacc_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -458,7 +458,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmacc.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmacc.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmacc_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -467,7 +467,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmacc.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmacc.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmacc_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -476,7 +476,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmacc_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmacc.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmacc.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmacc_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmadd.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmadd.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmadd.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmadd_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmadd_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmadd_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmadd_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmadd_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmadd_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmadd_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmadd_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmadd_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmadd_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmadd_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmadd_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmadd_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmadd_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmadd_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmadd_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmadd_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmadd_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -377,7 +377,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmadd.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmadd.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmadd_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -386,7 +386,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmadd.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmadd.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmadd_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -395,7 +395,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmadd.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmadd.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmadd_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -404,7 +404,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmadd.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmadd.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmadd_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -413,7 +413,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmadd.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmadd.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmadd_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -422,7 +422,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmadd.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmadd.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmadd_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -431,7 +431,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmadd.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmadd.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmadd_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -440,7 +440,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmadd.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmadd.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmadd_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -449,7 +449,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmadd.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmadd.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmadd_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -458,7 +458,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmadd.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmadd.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmadd_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -467,7 +467,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmadd.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmadd.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmadd_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -476,7 +476,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmadd_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmadd.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmadd.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmadd_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsac.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsac_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsac_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsac_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsac_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsac_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsac_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsac_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsac_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsac_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsac_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsac_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsac_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsac_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsac_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsac_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsac_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsac_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsac_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -377,7 +377,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsac.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsac.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmsac_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -386,7 +386,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsac.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsac.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmsac_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -395,7 +395,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsac.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsac.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmsac_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -404,7 +404,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsac.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsac.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmsac_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -413,7 +413,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsac.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsac.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmsac_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -422,7 +422,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsac.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsac.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmsac_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -431,7 +431,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsac.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsac.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmsac_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -440,7 +440,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsac.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsac.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmsac_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -449,7 +449,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsac.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsac.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmsac_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -458,7 +458,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsac.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsac.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmsac_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -467,7 +467,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsac.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsac.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmsac_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -476,7 +476,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsac_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsac.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsac.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmsac_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsub.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsub.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfmsub.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsub_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfmsub_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsub_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfmsub_vf_f32m1(vfloat32m1_t acc, float op1, vfloat32m1_t op2,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsub_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfmsub_vf_f32m2(vfloat32m2_t acc, float op1, vfloat32m2_t op2,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsub_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfmsub_vf_f32m4(vfloat32m4_t acc, float op1, vfloat32m4_t op2,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsub_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfmsub_vf_f32m8(vfloat32m8_t acc, float op1, vfloat32m8_t op2,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsub_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfmsub_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsub_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfmsub_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsub_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfmsub_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsub_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfmsub_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -377,7 +377,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsub.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsub.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmsub_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -386,7 +386,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsub.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfmsub.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfmsub_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -395,7 +395,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsub.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsub.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmsub_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -404,7 +404,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsub.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfmsub.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfmsub_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -413,7 +413,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsub.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsub.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmsub_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -422,7 +422,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsub.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfmsub.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfmsub_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -431,7 +431,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsub.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsub.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmsub_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -440,7 +440,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsub.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfmsub.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfmsub_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -449,7 +449,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsub.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsub.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmsub_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -458,7 +458,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsub.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfmsub.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfmsub_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -467,7 +467,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsub.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsub.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmsub_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -476,7 +476,7 @@
 
 // CHECK-RV64-LABEL: @test_vfmsub_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsub.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfmsub.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfmsub_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmacc.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmacc_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmacc.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmacc_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmacc_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmacc.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmacc_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmacc_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmacc.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmacc_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmacc_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmacc.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmacc_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmacc_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmacc.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmacc_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmacc_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmacc.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmacc_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmacc_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmacc.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmacc_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmacc_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmacc.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmacc_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmacc_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmacc.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmacc_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -378,7 +378,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmacc.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmacc.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmacc_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -387,7 +387,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmacc.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmacc.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmacc_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -396,7 +396,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmacc.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmacc.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmacc_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -405,7 +405,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmacc.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmacc.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmacc_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -414,7 +414,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmacc.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmacc.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmacc_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -423,7 +423,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmacc.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmacc.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmacc_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -432,7 +432,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmacc.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmacc.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmacc_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -441,7 +441,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmacc.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmacc.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmacc_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -450,7 +450,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmacc.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmacc.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmacc_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -459,7 +459,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmacc.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmacc.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmacc_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -468,7 +468,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmacc.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmacc.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmacc_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -477,7 +477,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmacc_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmacc.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmacc.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmacc_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmadd.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmadd.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmadd.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmadd_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmadd.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmadd_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmadd_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmadd.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmadd_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmadd_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmadd.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmadd_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmadd_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmadd.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmadd_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmadd_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmadd.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmadd_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmadd_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmadd.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmadd_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmadd_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmadd.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmadd_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmadd_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmadd.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmadd_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmadd_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmadd.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmadd_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -378,7 +378,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmadd.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmadd.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmadd_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -387,7 +387,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmadd.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmadd.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmadd_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -396,7 +396,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmadd.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmadd.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmadd_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -405,7 +405,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmadd.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmadd.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmadd_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -414,7 +414,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmadd.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmadd.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmadd_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -423,7 +423,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmadd.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmadd.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmadd_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -432,7 +432,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmadd.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmadd.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmadd_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -441,7 +441,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmadd.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmadd.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmadd_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -450,7 +450,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmadd.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmadd.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmadd_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -459,7 +459,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmadd.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmadd.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmadd_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -468,7 +468,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmadd.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmadd.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmadd_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -477,7 +477,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmadd_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmadd.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmadd.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmadd_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsac.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsac_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsac.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsac_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsac_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsac.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsac_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsac_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsac.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsac_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsac_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsac.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsac_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsac_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsac.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsac_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsac_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsac.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsac_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsac_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsac.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsac_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsac_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsac.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsac_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsac_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsac.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsac_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -378,7 +378,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsac.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsac.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmsac_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -387,7 +387,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsac.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsac.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmsac_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -396,7 +396,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsac.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsac.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmsac_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -405,7 +405,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsac.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsac.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmsac_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -414,7 +414,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsac.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsac.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmsac_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -423,7 +423,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsac.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsac.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmsac_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -432,7 +432,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsac.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsac.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmsac_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -441,7 +441,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsac.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsac.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmsac_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -450,7 +450,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsac.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsac.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmsac_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -459,7 +459,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsac.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsac.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmsac_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -468,7 +468,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsac.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsac.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmsac_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -477,7 +477,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsac_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsac.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsac.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmsac_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsub.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsub.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfnmsub.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.nxv1f32.i64(<vscale x 1 x float> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsub_vv_f32mf2(vfloat32mf2_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfnmsub.nxv1f32.f32.i64(<vscale x 1 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfnmsub_vf_f32mf2(vfloat32mf2_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.nxv2f32.i64(<vscale x 2 x float> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsub_vv_f32m1(vfloat32m1_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfnmsub.nxv2f32.f32.i64(<vscale x 2 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfnmsub_vf_f32m1(vfloat32m1_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.nxv4f32.i64(<vscale x 4 x float> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsub_vv_f32m2(vfloat32m2_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfnmsub.nxv4f32.f32.i64(<vscale x 4 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfnmsub_vf_f32m2(vfloat32m2_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.nxv8f32.i64(<vscale x 8 x float> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsub_vv_f32m4(vfloat32m4_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfnmsub.nxv8f32.f32.i64(<vscale x 8 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfnmsub_vf_f32m4(vfloat32m4_t acc, float op1,
@@ -88,7 +88,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.nxv16f32.i64(<vscale x 16 x float> [[ACC:%.*]], <vscale x 16 x float> [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsub_vv_f32m8(vfloat32m8_t acc, vfloat32m8_t op1,
@@ -98,7 +98,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfnmsub.nxv16f32.f32.i64(<vscale x 16 x float> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 16 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfnmsub_vf_f32m8(vfloat32m8_t acc, float op1,
@@ -108,7 +108,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.nxv1f64.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x double> [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsub_vv_f64m1(vfloat64m1_t acc, vfloat64m1_t op1,
@@ -118,7 +118,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfnmsub.nxv1f64.f64.i64(<vscale x 1 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 1 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfnmsub_vf_f64m1(vfloat64m1_t acc, double op1,
@@ -128,7 +128,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.nxv2f64.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x double> [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsub_vv_f64m2(vfloat64m2_t acc, vfloat64m2_t op1,
@@ -138,7 +138,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfnmsub.nxv2f64.f64.i64(<vscale x 2 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 2 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfnmsub_vf_f64m2(vfloat64m2_t acc, double op1,
@@ -148,7 +148,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.nxv4f64.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x double> [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsub_vv_f64m4(vfloat64m4_t acc, vfloat64m4_t op1,
@@ -158,7 +158,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfnmsub.nxv4f64.f64.i64(<vscale x 4 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 4 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfnmsub_vf_f64m4(vfloat64m4_t acc, double op1,
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.nxv8f64.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x double> [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsub_vv_f64m8(vfloat64m8_t acc, vfloat64m8_t op1,
@@ -178,7 +178,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfnmsub.nxv8f64.f64.i64(<vscale x 8 x double> [[ACC:%.*]], double [[OP1:%.*]], <vscale x 8 x double> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfnmsub_vf_f64m8(vfloat64m8_t acc, double op1,
@@ -378,7 +378,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsub.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsub.nxv1f16.nxv1f16.i64(<vscale x 1 x half> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmsub_vv_f16mf4(vfloat16mf4_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -387,7 +387,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsub.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x half> @llvm.riscv.vfnmsub.nxv1f16.f16.i64(<vscale x 1 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x half> [[TMP0]]
 //
 vfloat16mf4_t test_vfnmsub_vf_f16mf4(vfloat16mf4_t vd, _Float16 rs1, vfloat16mf4_t vs2, size_t vl) {
@@ -396,7 +396,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsub.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsub.nxv2f16.nxv2f16.i64(<vscale x 2 x half> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmsub_vv_f16mf2(vfloat16mf2_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -405,7 +405,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsub.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x half> @llvm.riscv.vfnmsub.nxv2f16.f16.i64(<vscale x 2 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x half> [[TMP0]]
 //
 vfloat16mf2_t test_vfnmsub_vf_f16mf2(vfloat16mf2_t vd, _Float16 rs1, vfloat16mf2_t vs2, size_t vl) {
@@ -414,7 +414,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsub.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsub.nxv4f16.nxv4f16.i64(<vscale x 4 x half> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmsub_vv_f16m1(vfloat16m1_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -423,7 +423,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsub.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x half> @llvm.riscv.vfnmsub.nxv4f16.f16.i64(<vscale x 4 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x half> [[TMP0]]
 //
 vfloat16m1_t test_vfnmsub_vf_f16m1(vfloat16m1_t vd, _Float16 rs1, vfloat16m1_t vs2, size_t vl) {
@@ -432,7 +432,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsub.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsub.nxv8f16.nxv8f16.i64(<vscale x 8 x half> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmsub_vv_f16m2(vfloat16m2_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -441,7 +441,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsub.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x half> @llvm.riscv.vfnmsub.nxv8f16.f16.i64(<vscale x 8 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x half> [[TMP0]]
 //
 vfloat16m2_t test_vfnmsub_vf_f16m2(vfloat16m2_t vd, _Float16 rs1, vfloat16m2_t vs2, size_t vl) {
@@ -450,7 +450,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsub.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsub.nxv16f16.nxv16f16.i64(<vscale x 16 x half> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmsub_vv_f16m4(vfloat16m4_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -459,7 +459,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsub.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x half> @llvm.riscv.vfnmsub.nxv16f16.f16.i64(<vscale x 16 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x half> [[TMP0]]
 //
 vfloat16m4_t test_vfnmsub_vf_f16m4(vfloat16m4_t vd, _Float16 rs1, vfloat16m4_t vs2, size_t vl) {
@@ -468,7 +468,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vv_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsub.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsub.nxv32f16.nxv32f16.i64(<vscale x 32 x half> [[VD:%.*]], <vscale x 32 x half> [[VS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmsub_vv_f16m8(vfloat16m8_t vd, vfloat16m8_t vs1, vfloat16m8_t vs2, size_t vl) {
@@ -477,7 +477,7 @@
 
 // CHECK-RV64-LABEL: @test_vfnmsub_vf_f16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsub.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x half> @llvm.riscv.vfnmsub.nxv32f16.f16.i64(<vscale x 32 x half> [[VD:%.*]], half [[RS1:%.*]], <vscale x 32 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x half> [[TMP0]]
 //
 vfloat16m8_t test_vfnmsub_vf_f16m8(vfloat16m8_t vd, _Float16 rs1, vfloat16m8_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmacc.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmacc_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmacc_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmacc_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmacc_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmacc_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmacc_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmacc_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmacc_vf_f64m8(vfloat64m8_t acc, float op1,
@@ -172,7 +172,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwmacc_vv_f32mf2(vfloat32mf2_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -181,7 +181,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmacc.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwmacc_vf_f32mf2(vfloat32mf2_t vd, _Float16 vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -190,7 +190,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwmacc_vv_f32m1(vfloat32m1_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -199,7 +199,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmacc.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwmacc_vf_f32m1(vfloat32m1_t vd, _Float16 vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -208,7 +208,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwmacc_vv_f32m2(vfloat32m2_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -217,7 +217,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmacc.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwmacc_vf_f32m2(vfloat32m2_t vd, _Float16 vs1, vfloat16m1_t vs2, size_t vl) {
@@ -226,7 +226,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwmacc_vv_f32m4(vfloat32m4_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -235,7 +235,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmacc.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwmacc_vf_f32m4(vfloat32m4_t vd, _Float16 vs1, vfloat16m2_t vs2, size_t vl) {
@@ -244,7 +244,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwmacc_vv_f32m8(vfloat32m8_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -253,7 +253,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmacc_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmacc.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwmacc_vf_f32m8(vfloat32m8_t vd, _Float16 vs1, vfloat16m4_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwmsac.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmsac_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwmsac_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmsac_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwmsac_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmsac_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwmsac_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmsac_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwmsac_vf_f64m8(vfloat64m8_t acc, float op1,
@@ -172,7 +172,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwmsac_vv_f32mf2(vfloat32mf2_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -181,7 +181,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwmsac.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwmsac_vf_f32mf2(vfloat32mf2_t vd, _Float16 vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -190,7 +190,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwmsac_vv_f32m1(vfloat32m1_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -199,7 +199,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwmsac.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwmsac_vf_f32m1(vfloat32m1_t vd, _Float16 vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -208,7 +208,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwmsac_vv_f32m2(vfloat32m2_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -217,7 +217,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwmsac.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwmsac_vf_f32m2(vfloat32m2_t vd, _Float16 vs1, vfloat16m1_t vs2, size_t vl) {
@@ -226,7 +226,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwmsac_vv_f32m4(vfloat32m4_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -235,7 +235,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwmsac.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwmsac_vf_f32m4(vfloat32m4_t vd, _Float16 vs1, vfloat16m2_t vs2, size_t vl) {
@@ -244,7 +244,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwmsac_vv_f32m8(vfloat32m8_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -253,7 +253,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwmsac_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwmsac.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwmsac_vf_f32m8(vfloat32m8_t vd, _Float16 vs1, vfloat16m4_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmacc.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmacc_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmacc.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmacc_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmacc_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmacc.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmacc_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmacc_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmacc.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmacc_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmacc_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmacc.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmacc_vf_f64m8(vfloat64m8_t acc, float op1,
@@ -172,7 +172,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwnmacc_vv_f32mf2(vfloat32mf2_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -181,7 +181,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmacc.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwnmacc_vf_f32mf2(vfloat32mf2_t vd, _Float16 vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -190,7 +190,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwnmacc_vv_f32m1(vfloat32m1_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -199,7 +199,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmacc.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwnmacc_vf_f32m1(vfloat32m1_t vd, _Float16 vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -208,7 +208,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwnmacc_vv_f32m2(vfloat32m2_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -217,7 +217,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmacc.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwnmacc_vf_f32m2(vfloat32m2_t vd, _Float16 vs1, vfloat16m1_t vs2, size_t vl) {
@@ -226,7 +226,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwnmacc_vv_f32m4(vfloat32m4_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -235,7 +235,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmacc.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwnmacc_vf_f32m4(vfloat32m4_t vd, _Float16 vs1, vfloat16m2_t vs2, size_t vl) {
@@ -244,7 +244,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwnmacc_vv_f32m8(vfloat32m8_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -253,7 +253,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmacc_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmacc.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwnmacc_vf_f32m8(vfloat32m8_t vd, _Float16 vs1, vfloat16m4_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vfwnmsac.c
@@ -8,7 +8,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.nxv1f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], <vscale x 1 x float> [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmsac_vv_f64m1(vfloat64m1_t acc, vfloat32mf2_t op1,
@@ -18,7 +18,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x double> @llvm.riscv.vfwnmsac.nxv1f64.f32.nxv1f32.i64(<vscale x 1 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 1 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x double> [[TMP0]]
 //
 vfloat64m1_t test_vfwnmsac_vf_f64m1(vfloat64m1_t acc, float op1,
@@ -28,7 +28,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.nxv2f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], <vscale x 2 x float> [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmsac_vv_f64m2(vfloat64m2_t acc, vfloat32m1_t op1,
@@ -38,7 +38,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x double> @llvm.riscv.vfwnmsac.nxv2f64.f32.nxv2f32.i64(<vscale x 2 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 2 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x double> [[TMP0]]
 //
 vfloat64m2_t test_vfwnmsac_vf_f64m2(vfloat64m2_t acc, float op1,
@@ -48,7 +48,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.nxv4f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], <vscale x 4 x float> [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmsac_vv_f64m4(vfloat64m4_t acc, vfloat32m2_t op1,
@@ -58,7 +58,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x double> @llvm.riscv.vfwnmsac.nxv4f64.f32.nxv4f32.i64(<vscale x 4 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 4 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x double> [[TMP0]]
 //
 vfloat64m4_t test_vfwnmsac_vf_f64m4(vfloat64m4_t acc, float op1,
@@ -68,7 +68,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.nxv8f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], <vscale x 8 x float> [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmsac_vv_f64m8(vfloat64m8_t acc, vfloat32m4_t op1,
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x double> @llvm.riscv.vfwnmsac.nxv8f64.f32.nxv8f32.i64(<vscale x 8 x double> [[ACC:%.*]], float [[OP1:%.*]], <vscale x 8 x float> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x double> [[TMP0]]
 //
 vfloat64m8_t test_vfwnmsac_vf_f64m8(vfloat64m8_t acc, float op1,
@@ -172,7 +172,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.nxv1f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], <vscale x 1 x half> [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwnmsac_vv_f32mf2(vfloat32mf2_t vd, vfloat16mf4_t vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -181,7 +181,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x float> @llvm.riscv.vfwnmsac.nxv1f32.f16.nxv1f16.i64(<vscale x 1 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 1 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x float> [[TMP0]]
 //
 vfloat32mf2_t test_vfwnmsac_vf_f32mf2(vfloat32mf2_t vd, _Float16 vs1, vfloat16mf4_t vs2, size_t vl) {
@@ -190,7 +190,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.nxv2f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], <vscale x 2 x half> [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwnmsac_vv_f32m1(vfloat32m1_t vd, vfloat16mf2_t vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -199,7 +199,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x float> @llvm.riscv.vfwnmsac.nxv2f32.f16.nxv2f16.i64(<vscale x 2 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 2 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x float> [[TMP0]]
 //
 vfloat32m1_t test_vfwnmsac_vf_f32m1(vfloat32m1_t vd, _Float16 vs1, vfloat16mf2_t vs2, size_t vl) {
@@ -208,7 +208,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.nxv4f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], <vscale x 4 x half> [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwnmsac_vv_f32m2(vfloat32m2_t vd, vfloat16m1_t vs1, vfloat16m1_t vs2, size_t vl) {
@@ -217,7 +217,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x float> @llvm.riscv.vfwnmsac.nxv4f32.f16.nxv4f16.i64(<vscale x 4 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 4 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x float> [[TMP0]]
 //
 vfloat32m2_t test_vfwnmsac_vf_f32m2(vfloat32m2_t vd, _Float16 vs1, vfloat16m1_t vs2, size_t vl) {
@@ -226,7 +226,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.nxv8f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], <vscale x 8 x half> [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwnmsac_vv_f32m4(vfloat32m4_t vd, vfloat16m2_t vs1, vfloat16m2_t vs2, size_t vl) {
@@ -235,7 +235,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x float> @llvm.riscv.vfwnmsac.nxv8f32.f16.nxv8f16.i64(<vscale x 8 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 8 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x float> [[TMP0]]
 //
 vfloat32m4_t test_vfwnmsac_vf_f32m4(vfloat32m4_t vd, _Float16 vs1, vfloat16m2_t vs2, size_t vl) {
@@ -244,7 +244,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vv_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.nxv16f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], <vscale x 16 x half> [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwnmsac_vv_f32m8(vfloat32m8_t vd, vfloat16m4_t vs1, vfloat16m4_t vs2, size_t vl) {
@@ -253,7 +253,7 @@
 
 // CHECK-RV64-LABEL: @test_vfwnmsac_vf_f32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x float> @llvm.riscv.vfwnmsac.nxv16f32.f16.nxv16f16.i64(<vscale x 16 x float> [[VD:%.*]], half [[VS1:%.*]], <vscale x 16 x half> [[VS2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x float> [[TMP0]]
 //
 vfloat32m8_t test_vfwnmsac_vf_f32m8(vfloat32m8_t vd, _Float16 vs1, vfloat16m4_t vs2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vmacc.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmacc_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmacc_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmacc_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmacc_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmacc_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmacc_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmacc_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmacc_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmacc_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmacc_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmacc_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmacc_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmacc_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmacc_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmacc_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmacc_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmacc_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmacc_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmacc_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmacc_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmacc_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmacc_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmacc_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmacc_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmacc_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmacc_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmacc_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmacc_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmacc_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmacc_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmacc_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmacc_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmacc_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmacc_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmacc_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmacc_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmacc_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmacc_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmacc_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmacc_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmacc_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmacc_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmacc_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmacc_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmacc_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmacc.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmacc_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmacc_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmacc.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmacc_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmacc_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmacc.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmacc_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmacc_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmacc.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmacc_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmacc_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmacc.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmacc_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmacc_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmacc.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmacc_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmacc_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmacc.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmacc_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmacc_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmacc.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmacc_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmacc_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmacc.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmacc_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmacc_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmacc.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmacc_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmacc_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmacc.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmacc_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmacc_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmacc.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmacc_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmacc_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmacc.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmacc_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmacc_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmacc.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmacc_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmacc_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmacc.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmacc_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmacc_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmacc.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmacc_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmacc_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmacc.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmacc_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmacc_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmacc.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmacc_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmacc_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmacc.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmacc_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmacc_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmacc.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmacc_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmacc_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmacc.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmacc_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmacc_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vmacc_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmacc.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmacc_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vmadd.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vmadd.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vmadd.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vmadd.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmadd_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vmadd_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmadd_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vmadd_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmadd_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vmadd_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmadd_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vmadd_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmadd_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vmadd_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmadd_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vmadd_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmadd_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vmadd_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmadd_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vmadd_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmadd_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vmadd_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmadd_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vmadd_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmadd_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vmadd_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmadd_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vmadd_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmadd_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vmadd_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmadd_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vmadd_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmadd_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vmadd_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmadd_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vmadd_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmadd_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vmadd_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmadd_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vmadd_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmadd_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vmadd_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmadd_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vmadd_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmadd_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vmadd_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmadd_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vmadd_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmadd_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vmadd.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vmadd_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmadd_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vmadd.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vmadd_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmadd_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vmadd.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vmadd_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmadd_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vmadd.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vmadd_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmadd_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vmadd.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vmadd_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmadd_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vmadd.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vmadd_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmadd_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vmadd.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vmadd_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmadd_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vmadd.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vmadd_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmadd_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vmadd.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vmadd_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmadd_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vmadd.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vmadd_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmadd_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vmadd.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vmadd_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmadd_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vmadd.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vmadd_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmadd_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vmadd.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vmadd_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmadd_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vmadd.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vmadd_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmadd_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vmadd.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vmadd_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmadd_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vmadd.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vmadd_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmadd_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vmadd.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vmadd_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmadd_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vmadd.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vmadd_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmadd_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vmadd.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vmadd_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmadd_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vmadd.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vmadd_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmadd_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vmadd.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vmadd_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmadd_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vmadd_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vmadd.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vmadd_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsac.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsac.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsac.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsac.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsac_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsac_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsac_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsac_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsac_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsac_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsac_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsac_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsac_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsac_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsac_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsac_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsac_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsac_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsac_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsac_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsac_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsac_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsac_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsac_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsac_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsac_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsac_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsac_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsac_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsac_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsac_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsac_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsac_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsac_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsac_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsac_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsac_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsac_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsac_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsac_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsac_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsac_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsac_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsac_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsac_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsac_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsac_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsac_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsac_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsac.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsac_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsac_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsac.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsac_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsac_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsac.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsac_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsac_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsac.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsac_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsac_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsac.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsac_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsac_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsac.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsac_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsac_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsac.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsac_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsac_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsac.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsac_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsac_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsac.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsac_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsac_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsac.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsac_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsac_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsac.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsac_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsac_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsac.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsac_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsac_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsac.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsac_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsac_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsac.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsac_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsac_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsac.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsac_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsac_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsac.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsac_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsac_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsac.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsac_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsac_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsac.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsac_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsac_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsac.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsac_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsac_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsac.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsac_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsac_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsac.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsac_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsac_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsac_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsac.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsac_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsub.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsub.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsub.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vnmsub.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsub_vv_i8mf8(vint8mf8_t acc, vint8mf8_t op1, vint8mf8_t op2, size_t vl) {
@@ -15,7 +15,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vint8mf8_t test_vnmsub_vx_i8mf8(vint8mf8_t acc, int8_t op1, vint8mf8_t op2, size_t vl) {
@@ -24,7 +24,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsub_vv_i8mf4(vint8mf4_t acc, vint8mf4_t op1, vint8mf4_t op2, size_t vl) {
@@ -33,7 +33,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vint8mf4_t test_vnmsub_vx_i8mf4(vint8mf4_t acc, int8_t op1, vint8mf4_t op2, size_t vl) {
@@ -42,7 +42,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsub_vv_i8mf2(vint8mf2_t acc, vint8mf2_t op1, vint8mf2_t op2, size_t vl) {
@@ -51,7 +51,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vint8mf2_t test_vnmsub_vx_i8mf2(vint8mf2_t acc, int8_t op1, vint8mf2_t op2, size_t vl) {
@@ -60,7 +60,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsub_vv_i8m1(vint8m1_t acc, vint8m1_t op1, vint8m1_t op2, size_t vl) {
@@ -69,7 +69,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vint8m1_t test_vnmsub_vx_i8m1(vint8m1_t acc, int8_t op1, vint8m1_t op2, size_t vl) {
@@ -78,7 +78,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsub_vv_i8m2(vint8m2_t acc, vint8m2_t op1, vint8m2_t op2, size_t vl) {
@@ -87,7 +87,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vint8m2_t test_vnmsub_vx_i8m2(vint8m2_t acc, int8_t op1, vint8m2_t op2, size_t vl) {
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsub_vv_i8m4(vint8m4_t acc, vint8m4_t op1, vint8m4_t op2, size_t vl) {
@@ -105,7 +105,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vint8m4_t test_vnmsub_vx_i8m4(vint8m4_t acc, int8_t op1, vint8m4_t op2, size_t vl) {
@@ -114,7 +114,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsub_vv_i8m8(vint8m8_t acc, vint8m8_t op1, vint8m8_t op2, size_t vl) {
@@ -123,7 +123,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vint8m8_t test_vnmsub_vx_i8m8(vint8m8_t acc, int8_t op1, vint8m8_t op2, size_t vl) {
@@ -132,7 +132,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsub_vv_i16mf4(vint16mf4_t acc, vint16mf4_t op1, vint16mf4_t op2, size_t vl) {
@@ -141,7 +141,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vnmsub_vx_i16mf4(vint16mf4_t acc, int16_t op1, vint16mf4_t op2, size_t vl) {
@@ -150,7 +150,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsub_vv_i16mf2(vint16mf2_t acc, vint16mf2_t op1, vint16mf2_t op2, size_t vl) {
@@ -159,7 +159,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vnmsub_vx_i16mf2(vint16mf2_t acc, int16_t op1, vint16mf2_t op2, size_t vl) {
@@ -168,7 +168,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsub_vv_i16m1(vint16m1_t acc, vint16m1_t op1, vint16m1_t op2, size_t vl) {
@@ -177,7 +177,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vnmsub_vx_i16m1(vint16m1_t acc, int16_t op1, vint16m1_t op2, size_t vl) {
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsub_vv_i16m2(vint16m2_t acc, vint16m2_t op1, vint16m2_t op2, size_t vl) {
@@ -195,7 +195,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vnmsub_vx_i16m2(vint16m2_t acc, int16_t op1, vint16m2_t op2, size_t vl) {
@@ -204,7 +204,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsub_vv_i16m4(vint16m4_t acc, vint16m4_t op1, vint16m4_t op2, size_t vl) {
@@ -213,7 +213,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vnmsub_vx_i16m4(vint16m4_t acc, int16_t op1, vint16m4_t op2, size_t vl) {
@@ -222,7 +222,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsub_vv_i16m8(vint16m8_t acc, vint16m8_t op1, vint16m8_t op2, size_t vl) {
@@ -231,7 +231,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vnmsub_vx_i16m8(vint16m8_t acc, int16_t op1, vint16m8_t op2, size_t vl) {
@@ -240,7 +240,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsub_vv_i32mf2(vint32mf2_t acc, vint32mf2_t op1, vint32mf2_t op2, size_t vl) {
@@ -249,7 +249,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vnmsub_vx_i32mf2(vint32mf2_t acc, int32_t op1, vint32mf2_t op2, size_t vl) {
@@ -258,7 +258,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsub_vv_i32m1(vint32m1_t acc, vint32m1_t op1, vint32m1_t op2, size_t vl) {
@@ -267,7 +267,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vnmsub_vx_i32m1(vint32m1_t acc, int32_t op1, vint32m1_t op2, size_t vl) {
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsub_vv_i32m2(vint32m2_t acc, vint32m2_t op1, vint32m2_t op2, size_t vl) {
@@ -285,7 +285,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vnmsub_vx_i32m2(vint32m2_t acc, int32_t op1, vint32m2_t op2, size_t vl) {
@@ -294,7 +294,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsub_vv_i32m4(vint32m4_t acc, vint32m4_t op1, vint32m4_t op2, size_t vl) {
@@ -303,7 +303,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vnmsub_vx_i32m4(vint32m4_t acc, int32_t op1, vint32m4_t op2, size_t vl) {
@@ -312,7 +312,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsub_vv_i32m8(vint32m8_t acc, vint32m8_t op1, vint32m8_t op2, size_t vl) {
@@ -321,7 +321,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vnmsub_vx_i32m8(vint32m8_t acc, int32_t op1, vint32m8_t op2, size_t vl) {
@@ -330,7 +330,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsub_vv_i64m1(vint64m1_t acc, vint64m1_t op1, vint64m1_t op2, size_t vl) {
@@ -339,7 +339,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vnmsub_vx_i64m1(vint64m1_t acc, int64_t op1, vint64m1_t op2, size_t vl) {
@@ -348,7 +348,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsub_vv_i64m2(vint64m2_t acc, vint64m2_t op1, vint64m2_t op2, size_t vl) {
@@ -357,7 +357,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vnmsub_vx_i64m2(vint64m2_t acc, int64_t op1, vint64m2_t op2, size_t vl) {
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsub_vv_i64m4(vint64m4_t acc, vint64m4_t op1, vint64m4_t op2, size_t vl) {
@@ -375,7 +375,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vnmsub_vx_i64m4(vint64m4_t acc, int64_t op1, vint64m4_t op2, size_t vl) {
@@ -384,7 +384,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsub_vv_i64m8(vint64m8_t acc, vint64m8_t op1, vint64m8_t op2, size_t vl) {
@@ -393,7 +393,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vnmsub_vx_i64m8(vint64m8_t acc, int64_t op1, vint64m8_t op2, size_t vl) {
@@ -402,7 +402,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.nxv1i8.i64(<vscale x 1 x i8> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsub_vv_u8mf8(vuint8mf8_t acc, vuint8mf8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -411,7 +411,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8mf8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i8> @llvm.riscv.vnmsub.nxv1i8.i8.i64(<vscale x 1 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i8> [[TMP0]]
 //
 vuint8mf8_t test_vnmsub_vx_u8mf8(vuint8mf8_t acc, uint8_t op1, vuint8mf8_t op2, size_t vl) {
@@ -420,7 +420,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.nxv2i8.i64(<vscale x 2 x i8> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsub_vv_u8mf4(vuint8mf4_t acc, vuint8mf4_t op1, vuint8mf4_t op2, size_t vl) {
@@ -429,7 +429,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i8> @llvm.riscv.vnmsub.nxv2i8.i8.i64(<vscale x 2 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i8> [[TMP0]]
 //
 vuint8mf4_t test_vnmsub_vx_u8mf4(vuint8mf4_t acc, uint8_t op1, vuint8mf4_t op2, size_t vl) {
@@ -438,7 +438,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.nxv4i8.i64(<vscale x 4 x i8> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsub_vv_u8mf2(vuint8mf2_t acc, vuint8mf2_t op1, vuint8mf2_t op2, size_t vl) {
@@ -447,7 +447,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i8> @llvm.riscv.vnmsub.nxv4i8.i8.i64(<vscale x 4 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i8> [[TMP0]]
 //
 vuint8mf2_t test_vnmsub_vx_u8mf2(vuint8mf2_t acc, uint8_t op1, vuint8mf2_t op2, size_t vl) {
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.nxv8i8.i64(<vscale x 8 x i8> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsub_vv_u8m1(vuint8m1_t acc, vuint8m1_t op1, vuint8m1_t op2, size_t vl) {
@@ -465,7 +465,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i8> @llvm.riscv.vnmsub.nxv8i8.i8.i64(<vscale x 8 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i8> [[TMP0]]
 //
 vuint8m1_t test_vnmsub_vx_u8m1(vuint8m1_t acc, uint8_t op1, vuint8m1_t op2, size_t vl) {
@@ -474,7 +474,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.nxv16i8.i64(<vscale x 16 x i8> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsub_vv_u8m2(vuint8m2_t acc, vuint8m2_t op1, vuint8m2_t op2, size_t vl) {
@@ -483,7 +483,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i8> @llvm.riscv.vnmsub.nxv16i8.i8.i64(<vscale x 16 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i8> [[TMP0]]
 //
 vuint8m2_t test_vnmsub_vx_u8m2(vuint8m2_t acc, uint8_t op1, vuint8m2_t op2, size_t vl) {
@@ -492,7 +492,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.nxv32i8.i64(<vscale x 32 x i8> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsub_vv_u8m4(vuint8m4_t acc, vuint8m4_t op1, vuint8m4_t op2, size_t vl) {
@@ -501,7 +501,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i8> @llvm.riscv.vnmsub.nxv32i8.i8.i64(<vscale x 32 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i8> [[TMP0]]
 //
 vuint8m4_t test_vnmsub_vx_u8m4(vuint8m4_t acc, uint8_t op1, vuint8m4_t op2, size_t vl) {
@@ -510,7 +510,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.nxv64i8.i64(<vscale x 64 x i8> [[ACC:%.*]], <vscale x 64 x i8> [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsub_vv_u8m8(vuint8m8_t acc, vuint8m8_t op1, vuint8m8_t op2, size_t vl) {
@@ -519,7 +519,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u8m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 64 x i8> @llvm.riscv.vnmsub.nxv64i8.i8.i64(<vscale x 64 x i8> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 64 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 64 x i8> [[TMP0]]
 //
 vuint8m8_t test_vnmsub_vx_u8m8(vuint8m8_t acc, uint8_t op1, vuint8m8_t op2, size_t vl) {
@@ -528,7 +528,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.nxv1i16.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsub_vv_u16mf4(vuint16mf4_t acc, vuint16mf4_t op1, vuint16mf4_t op2, size_t vl) {
@@ -537,7 +537,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vnmsub.nxv1i16.i16.i64(<vscale x 1 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vnmsub_vx_u16mf4(vuint16mf4_t acc, uint16_t op1, vuint16mf4_t op2, size_t vl) {
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.nxv2i16.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsub_vv_u16mf2(vuint16mf2_t acc, vuint16mf2_t op1, vuint16mf2_t op2, size_t vl) {
@@ -555,7 +555,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vnmsub.nxv2i16.i16.i64(<vscale x 2 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vnmsub_vx_u16mf2(vuint16mf2_t acc, uint16_t op1, vuint16mf2_t op2, size_t vl) {
@@ -564,7 +564,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.nxv4i16.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsub_vv_u16m1(vuint16m1_t acc, vuint16m1_t op1, vuint16m1_t op2, size_t vl) {
@@ -573,7 +573,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vnmsub.nxv4i16.i16.i64(<vscale x 4 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vnmsub_vx_u16m1(vuint16m1_t acc, uint16_t op1, vuint16m1_t op2, size_t vl) {
@@ -582,7 +582,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.nxv8i16.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsub_vv_u16m2(vuint16m2_t acc, vuint16m2_t op1, vuint16m2_t op2, size_t vl) {
@@ -591,7 +591,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vnmsub.nxv8i16.i16.i64(<vscale x 8 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vnmsub_vx_u16m2(vuint16m2_t acc, uint16_t op1, vuint16m2_t op2, size_t vl) {
@@ -600,7 +600,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.nxv16i16.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsub_vv_u16m4(vuint16m4_t acc, vuint16m4_t op1, vuint16m4_t op2, size_t vl) {
@@ -609,7 +609,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vnmsub.nxv16i16.i16.i64(<vscale x 16 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vnmsub_vx_u16m4(vuint16m4_t acc, uint16_t op1, vuint16m4_t op2, size_t vl) {
@@ -618,7 +618,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.nxv32i16.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i16> [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsub_vv_u16m8(vuint16m8_t acc, vuint16m8_t op1, vuint16m8_t op2, size_t vl) {
@@ -627,7 +627,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vnmsub.nxv32i16.i16.i64(<vscale x 32 x i16> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 32 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vnmsub_vx_u16m8(vuint16m8_t acc, uint16_t op1, vuint16m8_t op2, size_t vl) {
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.nxv1i32.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsub_vv_u32mf2(vuint32mf2_t acc, vuint32mf2_t op1, vuint32mf2_t op2, size_t vl) {
@@ -645,7 +645,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vnmsub.nxv1i32.i32.i64(<vscale x 1 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vnmsub_vx_u32mf2(vuint32mf2_t acc, uint32_t op1, vuint32mf2_t op2, size_t vl) {
@@ -654,7 +654,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.nxv2i32.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsub_vv_u32m1(vuint32m1_t acc, vuint32m1_t op1, vuint32m1_t op2, size_t vl) {
@@ -663,7 +663,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vnmsub.nxv2i32.i32.i64(<vscale x 2 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vnmsub_vx_u32m1(vuint32m1_t acc, uint32_t op1, vuint32m1_t op2, size_t vl) {
@@ -672,7 +672,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.nxv4i32.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsub_vv_u32m2(vuint32m2_t acc, vuint32m2_t op1, vuint32m2_t op2, size_t vl) {
@@ -681,7 +681,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vnmsub.nxv4i32.i32.i64(<vscale x 4 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vnmsub_vx_u32m2(vuint32m2_t acc, uint32_t op1, vuint32m2_t op2, size_t vl) {
@@ -690,7 +690,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.nxv8i32.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsub_vv_u32m4(vuint32m4_t acc, vuint32m4_t op1, vuint32m4_t op2, size_t vl) {
@@ -699,7 +699,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vnmsub.nxv8i32.i32.i64(<vscale x 8 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vnmsub_vx_u32m4(vuint32m4_t acc, uint32_t op1, vuint32m4_t op2, size_t vl) {
@@ -708,7 +708,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.nxv16i32.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i32> [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsub_vv_u32m8(vuint32m8_t acc, vuint32m8_t op1, vuint32m8_t op2, size_t vl) {
@@ -717,7 +717,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vnmsub.nxv16i32.i32.i64(<vscale x 16 x i32> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 16 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vnmsub_vx_u32m8(vuint32m8_t acc, uint32_t op1, vuint32m8_t op2, size_t vl) {
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.nxv1i64.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i64> [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsub_vv_u64m1(vuint64m1_t acc, vuint64m1_t op1, vuint64m1_t op2, size_t vl) {
@@ -735,7 +735,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vnmsub.nxv1i64.i64.i64(<vscale x 1 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 1 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vnmsub_vx_u64m1(vuint64m1_t acc, uint64_t op1, vuint64m1_t op2, size_t vl) {
@@ -744,7 +744,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.nxv2i64.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i64> [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsub_vv_u64m2(vuint64m2_t acc, vuint64m2_t op1, vuint64m2_t op2, size_t vl) {
@@ -753,7 +753,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vnmsub.nxv2i64.i64.i64(<vscale x 2 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 2 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vnmsub_vx_u64m2(vuint64m2_t acc, uint64_t op1, vuint64m2_t op2, size_t vl) {
@@ -762,7 +762,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.nxv4i64.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i64> [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsub_vv_u64m4(vuint64m4_t acc, vuint64m4_t op1, vuint64m4_t op2, size_t vl) {
@@ -771,7 +771,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vnmsub.nxv4i64.i64.i64(<vscale x 4 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 4 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vnmsub_vx_u64m4(vuint64m4_t acc, uint64_t op1, vuint64m4_t op2, size_t vl) {
@@ -780,7 +780,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.nxv8i64.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i64> [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsub_vv_u64m8(vuint64m8_t acc, vuint64m8_t op1, vuint64m8_t op2, size_t vl) {
@@ -789,7 +789,7 @@
 
 // CHECK-RV64-LABEL: @test_vnmsub_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vnmsub.nxv8i64.i64.i64(<vscale x 8 x i64> [[ACC:%.*]], i64 [[OP1:%.*]], <vscale x 8 x i64> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vnmsub_vx_u64m8(vuint64m8_t acc, uint64_t op1, vuint64m8_t op2, size_t vl) {
diff --git a/clang/test/CodeGen/RISCV/rvv-intrinsics/vwmacc.c b/clang/test/CodeGen/RISCV/rvv-intrinsics/vwmacc.c
--- a/clang/test/CodeGen/RISCV/rvv-intrinsics/vwmacc.c
+++ b/clang/test/CodeGen/RISCV/rvv-intrinsics/vwmacc.c
@@ -6,7 +6,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmacc_vv_i16mf4(vint16mf4_t acc, vint8mf8_t op1,
@@ -16,7 +16,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmacc.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmacc_vx_i16mf4(vint16mf4_t acc, int8_t op1, vint8mf8_t op2,
@@ -26,7 +26,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmacc_vv_i16mf2(vint16mf2_t acc, vint8mf4_t op1,
@@ -36,7 +36,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmacc.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmacc_vx_i16mf2(vint16mf2_t acc, int8_t op1, vint8mf4_t op2,
@@ -46,7 +46,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmacc_vv_i16m1(vint16m1_t acc, vint8mf2_t op1, vint8mf2_t op2,
@@ -56,7 +56,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmacc.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmacc_vx_i16m1(vint16m1_t acc, int8_t op1, vint8mf2_t op2,
@@ -66,7 +66,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmacc_vv_i16m2(vint16m2_t acc, vint8m1_t op1, vint8m1_t op2,
@@ -76,7 +76,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmacc.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmacc_vx_i16m2(vint16m2_t acc, int8_t op1, vint8m1_t op2,
@@ -86,7 +86,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmacc_vv_i16m4(vint16m4_t acc, vint8m2_t op1, vint8m2_t op2,
@@ -96,7 +96,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmacc.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmacc_vx_i16m4(vint16m4_t acc, int8_t op1, vint8m2_t op2,
@@ -106,7 +106,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmacc_vv_i16m8(vint16m8_t acc, vint8m4_t op1, vint8m4_t op2,
@@ -116,7 +116,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmacc.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmacc_vx_i16m8(vint16m8_t acc, int8_t op1, vint8m4_t op2,
@@ -126,7 +126,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmacc_vv_i32mf2(vint32mf2_t acc, vint16mf4_t op1,
@@ -136,7 +136,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmacc.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmacc_vx_i32mf2(vint32mf2_t acc, int16_t op1, vint16mf4_t op2,
@@ -146,7 +146,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmacc_vv_i32m1(vint32m1_t acc, vint16mf2_t op1,
@@ -156,7 +156,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmacc.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmacc_vx_i32m1(vint32m1_t acc, int16_t op1, vint16mf2_t op2,
@@ -166,7 +166,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmacc_vv_i32m2(vint32m2_t acc, vint16m1_t op1, vint16m1_t op2,
@@ -176,7 +176,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmacc.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmacc_vx_i32m2(vint32m2_t acc, int16_t op1, vint16m1_t op2,
@@ -186,7 +186,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmacc_vv_i32m4(vint32m4_t acc, vint16m2_t op1, vint16m2_t op2,
@@ -196,7 +196,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmacc.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmacc_vx_i32m4(vint32m4_t acc, int16_t op1, vint16m2_t op2,
@@ -206,7 +206,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmacc_vv_i32m8(vint32m8_t acc, vint16m4_t op1, vint16m4_t op2,
@@ -216,7 +216,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmacc.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmacc_vx_i32m8(vint32m8_t acc, int16_t op1, vint16m4_t op2,
@@ -226,7 +226,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmacc_vv_i64m1(vint64m1_t acc, vint32mf2_t op1,
@@ -236,7 +236,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmacc.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmacc_vx_i64m1(vint64m1_t acc, int32_t op1, vint32mf2_t op2,
@@ -246,7 +246,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmacc_vv_i64m2(vint64m2_t acc, vint32m1_t op1, vint32m1_t op2,
@@ -256,7 +256,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmacc.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmacc_vx_i64m2(vint64m2_t acc, int32_t op1, vint32m1_t op2,
@@ -266,7 +266,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmacc_vv_i64m4(vint64m4_t acc, vint32m2_t op1, vint32m2_t op2,
@@ -276,7 +276,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmacc.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmacc_vx_i64m4(vint64m4_t acc, int32_t op1, vint32m2_t op2,
@@ -286,7 +286,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmacc_vv_i64m8(vint64m8_t acc, vint32m4_t op1, vint32m4_t op2,
@@ -296,7 +296,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmacc_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmacc.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmacc_vx_i64m8(vint64m8_t acc, int32_t op1, vint32m4_t op2,
@@ -306,7 +306,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vwmaccu_vv_u16mf4(vuint16mf4_t acc, vuint8mf8_t op1,
@@ -316,7 +316,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vuint16mf4_t test_vwmaccu_vx_u16mf4(vuint16mf4_t acc, uint8_t op1,
@@ -326,7 +326,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vwmaccu_vv_u16mf2(vuint16mf2_t acc, vuint8mf4_t op1,
@@ -336,7 +336,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vuint16mf2_t test_vwmaccu_vx_u16mf2(vuint16mf2_t acc, uint8_t op1,
@@ -346,7 +346,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vwmaccu_vv_u16m1(vuint16m1_t acc, vuint8mf2_t op1,
@@ -356,7 +356,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vuint16m1_t test_vwmaccu_vx_u16m1(vuint16m1_t acc, uint8_t op1, vuint8mf2_t op2,
@@ -366,7 +366,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vwmaccu_vv_u16m2(vuint16m2_t acc, vuint8m1_t op1,
@@ -376,7 +376,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vuint16m2_t test_vwmaccu_vx_u16m2(vuint16m2_t acc, uint8_t op1, vuint8m1_t op2,
@@ -386,7 +386,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vwmaccu_vv_u16m4(vuint16m4_t acc, vuint8m2_t op1,
@@ -396,7 +396,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vuint16m4_t test_vwmaccu_vx_u16m4(vuint16m4_t acc, uint8_t op1, vuint8m2_t op2,
@@ -406,7 +406,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vwmaccu_vv_u16m8(vuint16m8_t acc, vuint8m4_t op1,
@@ -416,7 +416,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vuint16m8_t test_vwmaccu_vx_u16m8(vuint16m8_t acc, uint8_t op1, vuint8m4_t op2,
@@ -426,7 +426,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vwmaccu_vv_u32mf2(vuint32mf2_t acc, vuint16mf4_t op1,
@@ -436,7 +436,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vuint32mf2_t test_vwmaccu_vx_u32mf2(vuint32mf2_t acc, uint16_t op1,
@@ -446,7 +446,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vwmaccu_vv_u32m1(vuint32m1_t acc, vuint16mf2_t op1,
@@ -456,7 +456,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vuint32m1_t test_vwmaccu_vx_u32m1(vuint32m1_t acc, uint16_t op1,
@@ -466,7 +466,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vwmaccu_vv_u32m2(vuint32m2_t acc, vuint16m1_t op1,
@@ -476,7 +476,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vuint32m2_t test_vwmaccu_vx_u32m2(vuint32m2_t acc, uint16_t op1,
@@ -486,7 +486,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vwmaccu_vv_u32m4(vuint32m4_t acc, vuint16m2_t op1,
@@ -496,7 +496,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vuint32m4_t test_vwmaccu_vx_u32m4(vuint32m4_t acc, uint16_t op1,
@@ -506,7 +506,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vwmaccu_vv_u32m8(vuint32m8_t acc, vuint16m4_t op1,
@@ -516,7 +516,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vuint32m8_t test_vwmaccu_vx_u32m8(vuint32m8_t acc, uint16_t op1,
@@ -526,7 +526,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vwmaccu_vv_u64m1(vuint64m1_t acc, vuint32mf2_t op1,
@@ -536,7 +536,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vuint64m1_t test_vwmaccu_vx_u64m1(vuint64m1_t acc, uint32_t op1,
@@ -546,7 +546,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vwmaccu_vv_u64m2(vuint64m2_t acc, vuint32m1_t op1,
@@ -556,7 +556,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vuint64m2_t test_vwmaccu_vx_u64m2(vuint64m2_t acc, uint32_t op1,
@@ -566,7 +566,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vwmaccu_vv_u64m4(vuint64m4_t acc, vuint32m2_t op1,
@@ -576,7 +576,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vuint64m4_t test_vwmaccu_vx_u64m4(vuint64m4_t acc, uint32_t op1,
@@ -586,7 +586,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vv_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vwmaccu_vv_u64m8(vuint64m8_t acc, vuint32m4_t op1,
@@ -596,7 +596,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccu_vx_u64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vuint64m8_t test_vwmaccu_vx_u64m8(vuint64m8_t acc, uint32_t op1,
@@ -606,7 +606,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.nxv1i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], <vscale x 1 x i8> [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmaccsu_vv_i16mf4(vint16mf4_t acc, vint8mf8_t op1,
@@ -616,7 +616,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccsu.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmaccsu_vx_i16mf4(vint16mf4_t acc, int8_t op1,
@@ -626,7 +626,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.nxv2i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], <vscale x 2 x i8> [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmaccsu_vv_i16mf2(vint16mf2_t acc, vint8mf4_t op1,
@@ -636,7 +636,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccsu.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmaccsu_vx_i16mf2(vint16mf2_t acc, int8_t op1,
@@ -646,7 +646,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.nxv4i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], <vscale x 4 x i8> [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmaccsu_vv_i16m1(vint16m1_t acc, vint8mf2_t op1,
@@ -656,7 +656,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccsu.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmaccsu_vx_i16m1(vint16m1_t acc, int8_t op1, vuint8mf2_t op2,
@@ -666,7 +666,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], <vscale x 8 x i8> [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmaccsu_vv_i16m2(vint16m2_t acc, vint8m1_t op1, vuint8m1_t op2,
@@ -676,7 +676,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccsu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmaccsu_vx_i16m2(vint16m2_t acc, int8_t op1, vuint8m1_t op2,
@@ -686,7 +686,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], <vscale x 16 x i8> [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmaccsu_vv_i16m4(vint16m4_t acc, vint8m2_t op1, vuint8m2_t op2,
@@ -696,7 +696,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccsu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmaccsu_vx_i16m4(vint16m4_t acc, int8_t op1, vuint8m2_t op2,
@@ -706,7 +706,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], <vscale x 32 x i8> [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmaccsu_vv_i16m8(vint16m8_t acc, vint8m4_t op1, vuint8m4_t op2,
@@ -716,7 +716,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccsu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmaccsu_vx_i16m8(vint16m8_t acc, int8_t op1, vuint8m4_t op2,
@@ -726,7 +726,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.nxv1i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], <vscale x 1 x i16> [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmaccsu_vv_i32mf2(vint32mf2_t acc, vint16mf4_t op1,
@@ -736,7 +736,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccsu.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmaccsu_vx_i32mf2(vint32mf2_t acc, int16_t op1,
@@ -746,7 +746,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.nxv2i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], <vscale x 2 x i16> [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmaccsu_vv_i32m1(vint32m1_t acc, vint16mf2_t op1,
@@ -756,7 +756,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccsu.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmaccsu_vx_i32m1(vint32m1_t acc, int16_t op1, vuint16mf2_t op2,
@@ -766,7 +766,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], <vscale x 4 x i16> [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmaccsu_vv_i32m2(vint32m2_t acc, vint16m1_t op1,
@@ -776,7 +776,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccsu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmaccsu_vx_i32m2(vint32m2_t acc, int16_t op1, vuint16m1_t op2,
@@ -786,7 +786,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], <vscale x 8 x i16> [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmaccsu_vv_i32m4(vint32m4_t acc, vint16m2_t op1,
@@ -796,7 +796,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccsu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmaccsu_vx_i32m4(vint32m4_t acc, int16_t op1, vuint16m2_t op2,
@@ -806,7 +806,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], <vscale x 16 x i16> [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmaccsu_vv_i32m8(vint32m8_t acc, vint16m4_t op1,
@@ -816,7 +816,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccsu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmaccsu_vx_i32m8(vint32m8_t acc, int16_t op1, vuint16m4_t op2,
@@ -826,7 +826,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.nxv1i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], <vscale x 1 x i32> [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmaccsu_vv_i64m1(vint64m1_t acc, vint32mf2_t op1,
@@ -836,7 +836,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccsu.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmaccsu_vx_i64m1(vint64m1_t acc, int32_t op1, vuint32mf2_t op2,
@@ -846,7 +846,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], <vscale x 2 x i32> [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmaccsu_vv_i64m2(vint64m2_t acc, vint32m1_t op1,
@@ -856,7 +856,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccsu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmaccsu_vx_i64m2(vint64m2_t acc, int32_t op1, vuint32m1_t op2,
@@ -866,7 +866,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], <vscale x 4 x i32> [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmaccsu_vv_i64m4(vint64m4_t acc, vint32m2_t op1,
@@ -876,7 +876,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccsu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmaccsu_vx_i64m4(vint64m4_t acc, int32_t op1, vuint32m2_t op2,
@@ -886,7 +886,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vv_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], <vscale x 8 x i32> [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmaccsu_vv_i64m8(vint64m8_t acc, vint32m4_t op1,
@@ -896,7 +896,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccsu_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccsu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmaccsu_vx_i64m8(vint64m8_t acc, int32_t op1, vuint32m4_t op2,
@@ -906,7 +906,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16mf4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccus.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i16> @llvm.riscv.vwmaccus.nxv1i16.i8.nxv1i8.i64(<vscale x 1 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 1 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i16> [[TMP0]]
 //
 vint16mf4_t test_vwmaccus_vx_i16mf4(vint16mf4_t acc, uint8_t op1,
@@ -916,7 +916,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccus.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i16> @llvm.riscv.vwmaccus.nxv2i16.i8.nxv2i8.i64(<vscale x 2 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 2 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i16> [[TMP0]]
 //
 vint16mf2_t test_vwmaccus_vx_i16mf2(vint16mf2_t acc, uint8_t op1,
@@ -926,7 +926,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccus.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i16> @llvm.riscv.vwmaccus.nxv4i16.i8.nxv4i8.i64(<vscale x 4 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 4 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i16> [[TMP0]]
 //
 vint16m1_t test_vwmaccus_vx_i16m1(vint16m1_t acc, uint8_t op1, vint8mf2_t op2,
@@ -936,7 +936,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccus.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.vwmaccus.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 8 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
 //
 vint16m2_t test_vwmaccus_vx_i16m2(vint16m2_t acc, uint8_t op1, vint8m1_t op2,
@@ -946,7 +946,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccus.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.vwmaccus.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 16 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
 //
 vint16m4_t test_vwmaccus_vx_i16m4(vint16m4_t acc, uint8_t op1, vint8m2_t op2,
@@ -956,7 +956,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i16m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccus.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.vwmaccus.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[ACC:%.*]], i8 [[OP1:%.*]], <vscale x 32 x i8> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
 //
 vint16m8_t test_vwmaccus_vx_i16m8(vint16m8_t acc, uint8_t op1, vint8m4_t op2,
@@ -966,7 +966,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32mf2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccus.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i32> @llvm.riscv.vwmaccus.nxv1i32.i16.nxv1i16.i64(<vscale x 1 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 1 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i32> [[TMP0]]
 //
 vint32mf2_t test_vwmaccus_vx_i32mf2(vint32mf2_t acc, uint16_t op1,
@@ -976,7 +976,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccus.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i32> @llvm.riscv.vwmaccus.nxv2i32.i16.nxv2i16.i64(<vscale x 2 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 2 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i32> [[TMP0]]
 //
 vint32m1_t test_vwmaccus_vx_i32m1(vint32m1_t acc, uint16_t op1, vint16mf2_t op2,
@@ -986,7 +986,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccus.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.vwmaccus.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 4 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
 //
 vint32m2_t test_vwmaccus_vx_i32m2(vint32m2_t acc, uint16_t op1, vint16m1_t op2,
@@ -996,7 +996,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccus.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.vwmaccus.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 8 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
 //
 vint32m4_t test_vwmaccus_vx_i32m4(vint32m4_t acc, uint16_t op1, vint16m2_t op2,
@@ -1006,7 +1006,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i32m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccus.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.vwmaccus.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[ACC:%.*]], i16 [[OP1:%.*]], <vscale x 16 x i16> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
 //
 vint32m8_t test_vwmaccus_vx_i32m8(vint32m8_t acc, uint16_t op1, vint16m4_t op2,
@@ -1016,7 +1016,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m1(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccus.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 1 x i64> @llvm.riscv.vwmaccus.nxv1i64.i32.nxv1i32.i64(<vscale x 1 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 1 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 1 x i64> [[TMP0]]
 //
 vint64m1_t test_vwmaccus_vx_i64m1(vint64m1_t acc, uint32_t op1, vint32mf2_t op2,
@@ -1026,7 +1026,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m2(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccus.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.vwmaccus.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 2 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
 //
 vint64m2_t test_vwmaccus_vx_i64m2(vint64m2_t acc, uint32_t op1, vint32m1_t op2,
@@ -1036,7 +1036,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m4(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccus.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.vwmaccus.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 4 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
 //
 vint64m4_t test_vwmaccus_vx_i64m4(vint64m4_t acc, uint32_t op1, vint32m2_t op2,
@@ -1046,7 +1046,7 @@
 
 // CHECK-RV64-LABEL: @test_vwmaccus_vx_i64m8(
 // CHECK-RV64-NEXT:  entry:
-// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccus.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 0)
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.vwmaccus.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[ACC:%.*]], i32 [[OP1:%.*]], <vscale x 8 x i32> [[OP2:%.*]], i64 [[VL:%.*]], i64 1)
 // CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
 //
 vint64m8_t test_vwmaccus_vx_i64m8(vint64m8_t acc, uint32_t op1, vint32m4_t op2,