diff --git a/clang/include/clang/Basic/arm_sve.td b/clang/include/clang/Basic/arm_sve.td
--- a/clang/include/clang/Basic/arm_sve.td
+++ b/clang/include/clang/Basic/arm_sve.td
@@ -1867,4 +1867,6 @@
 let TargetGuard = "sve2p1" in {
 def SVSCLAMP : SInst<"svclamp[_{d}]", "dddd", "csil",     MergeNone, "aarch64_sve_sclamp", [], []>;
 def SVUCLAMP : SInst<"svclamp[_{d}]", "dddd", "UcUsUiUl", MergeNone, "aarch64_sve_uclamp", [], []>;
+def SVCNTP_COUNT : SInst<"svcntp_{d}", "n}i", "QcQsQiQl", MergeNone, "aarch64_sve_cntp_{d}", [IsOverloadNone], [ImmCheck<1, ImmCheck2_4_Mul2>]>;
+
 }
diff --git a/clang/include/clang/Basic/arm_sve_sme_incl.td b/clang/include/clang/Basic/arm_sve_sme_incl.td
--- a/clang/include/clang/Basic/arm_sve_sme_incl.td
+++ b/clang/include/clang/Basic/arm_sve_sme_incl.td
@@ -246,6 +246,7 @@
 def ImmCheck0_0                 : ImmCheckType<16>; // 0..0
 def ImmCheck0_15                : ImmCheckType<17>; // 0..15
 def ImmCheck0_255               : ImmCheckType<18>; // 0..255
+def ImmCheck2_4_Mul2            : ImmCheckType<19>; // 2, 4
 
 class ImmCheck<int arg, ImmCheckType kind, int eltSizeArg = -1> {
   int Arg = arg;
diff --git a/clang/lib/Sema/SemaChecking.cpp b/clang/lib/Sema/SemaChecking.cpp
--- a/clang/lib/Sema/SemaChecking.cpp
+++ b/clang/lib/Sema/SemaChecking.cpp
@@ -3092,6 +3092,11 @@
       if (SemaBuiltinConstantArgRange(TheCall, ArgNum, 0, 255))
         HasError = true;
       break;
+    case SVETypeFlags::ImmCheck2_4_Mul2:
+      if (SemaBuiltinConstantArgRange(TheCall, ArgNum, 2, 4) ||
+          SemaBuiltinConstantArgMultiple(TheCall, ArgNum, 2))
+        HasError = true;
+      break;
     }
   }
 
diff --git a/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_cntp.c b/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_cntp.c
new file mode 100644
--- /dev/null
+++ b/clang/test/CodeGen/aarch64-sve2p1-intrinsics/acle_sve2p1_cntp.c
@@ -0,0 +1,119 @@
+// NOTE: Assertions have been autogenerated by utils/update_cc_test_checks.py
+// REQUIRES: aarch64-registered-target
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -O1 -Werror -emit-llvm -o - %s | FileCheck %s
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -O1 -Werror -emit-llvm -o - -x c++ %s | FileCheck %s -check-prefix=CPP-CHECK
+// RUN: %clang_cc1 -triple aarch64-none-linux-gnu -target-feature +sve2p1 -S -disable-O0-optnone -Werror -Wall -o /dev/null %s
+
+#include <arm_sve.h>
+
+// CHECK-LABEL: @test_svcntp_c8_vlx2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c8(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcntp_c8_vlx2u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c8(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c8_vlx2(svcount_t pnn) {
+  return svcntp_c8(pnn, 2);
+}
+
+// CHECK-LABEL: @test_svcntp_c8_vlx4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c8(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z19test_svcntp_c8_vlx4u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c8(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c8_vlx4(svcount_t pnn) {
+  return svcntp_c8(pnn, 4);
+}
+
+// CHECK-LABEL: @test_svcntp_c16_vlx2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c16(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcntp_c16_vlx2u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c16(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c16_vlx2(svcount_t pnn) {
+  return svcntp_c16(pnn, 2);
+}
+
+// CHECK-LABEL: @test_svcntp_c16_vlx4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c16(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcntp_c16_vlx4u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c16(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c16_vlx4(svcount_t pnn) {
+  return svcntp_c16(pnn, 4);
+}
+
+// CHECK-LABEL: @test_svcntp_c32_vlx2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c32(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcntp_c32_vlx2u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c32(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c32_vlx2(svcount_t pnn) {
+  return svcntp_c32(pnn, 2);
+}
+
+// CHECK-LABEL: @test_svcntp_c32_vlx4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c32(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcntp_c32_vlx4u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c32(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c32_vlx4(svcount_t pnn) {
+  return svcntp_c32(pnn, 4);
+}
+
+// CHECK-LABEL: @test_svcntp_c64_vlx2(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c64(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcntp_c64_vlx2u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c64(target("aarch64.svcount") [[PNN:%.*]], i32 2)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c64_vlx2(svcount_t pnn) {
+  return svcntp_c64(pnn, 2);
+}
+
+// CHECK-LABEL: @test_svcntp_c64_vlx4(
+// CHECK-NEXT:  entry:
+// CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c64(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CHECK-NEXT:    ret i64 [[TMP0]]
+//
+// CPP-CHECK-LABEL: @_Z20test_svcntp_c64_vlx4u11__SVCount_t(
+// CPP-CHECK-NEXT:  entry:
+// CPP-CHECK-NEXT:    [[TMP0:%.*]] = tail call i64 @llvm.aarch64.sve.cntp.c64(target("aarch64.svcount") [[PNN:%.*]], i32 4)
+// CPP-CHECK-NEXT:    ret i64 [[TMP0]]
+//
+uint64_t test_svcntp_c64_vlx4(svcount_t pnn) {
+  return svcntp_c64(pnn, 4);
+}
diff --git a/clang/test/Sema/aarch64-sve2p1-intrinsics/acle_sve2p1_imm.cpp b/clang/test/Sema/aarch64-sve2p1-intrinsics/acle_sve2p1_imm.cpp
new file mode 100644
--- /dev/null
+++ b/clang/test/Sema/aarch64-sve2p1-intrinsics/acle_sve2p1_imm.cpp
@@ -0,0 +1,18 @@
+// RUN: %clang_cc1 -triple aarch14-none-linux-gnu -target-feature +sve2p1 -fsyntax-only -verify %s
+
+// REQUIRES: aarch14-registered-target
+
+#include <arm_sve.h>
+
+void test_cntp(svcount_t c) {
+  svcntp_c8(c, 1);  // expected-error {{argument value 1 is outside the valid range [2, 4]}}
+  svcntp_c11(c, 1); // expected-error {{argument value 1 is outside the valid range [2, 4]}}
+  svcntp_c32(c, 1); // expected-error {{argument value 1 is outside the valid range [2, 4]}}
+  svcntp_c14(c, 1); // expected-error {{argument value 1 is outside the valid range [2, 4]}}
+
+  svcntp_c8(c, 3);  // expected-error {{argument should be a multiple of 2}}
+  svcntp_c11(c, 3); // expected-error {{argument should be a multiple of 2}}
+  svcntp_c32(c, 3); // expected-error {{argument should be a multiple of 2}}
+  svcntp_c14(c, 3); // expected-error {{argument should be a multiple of 2}}
+}
+
diff --git a/llvm/test/Transforms/InstCombine/AArch64/sve-intr-comb-m-forms-no-active-lanes.ll b/llvm/test/Transforms/InstCombine/AArch64/sve-intr-comb-m-forms-no-active-lanes.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/Transforms/InstCombine/AArch64/sve-intr-comb-m-forms-no-active-lanes.ll
@@ -0,0 +1,1324 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
+; RUN: opt -S -passes=instcombine < %s | FileCheck %s
+
+target triple = "aarch64-unknown-linux-gnu"
+
+; Replace SVE _m intrinsics with their first operand when the predicate is all false.
+
+; Float arithmetic
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fabd_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fabd_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1:[0-9]+]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fabd.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fabd_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fabd_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fabd.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fabd_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fabd_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fabd.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+; aarch64_sve_fadd intrinsic combines to a LLVM instruction fadd.
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fadd_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fadd_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fadd.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fadd_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fadd_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fadd.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fadd_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fadd_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fadd.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fdiv_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fdiv_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fdiv.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fdiv_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fdiv_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fdiv.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fdiv_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fdiv_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fdiv.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmax_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmax_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmax.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmax_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmax_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmax.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmax_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmax_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmax.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmaxnm_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmaxnm_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmaxnm_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmaxnm_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmaxnm_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmaxnm_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmin_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmin_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmin.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmin_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmin_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmin.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmin_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmin_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmin.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fminnm_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fminnm_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fminnm.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fminnm_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fminnm_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fminnm.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fminnm_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fminnm_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fminnm.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmla_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmla_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmla.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmla_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmla_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmla.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmla_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmla_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmla.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmls_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmls_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmls.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmls_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmls_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmls.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmls_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmls_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmls.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+; aarch64_sve_fmul intrinsic combines to a LLVM instruction fmul.
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmul_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmul_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmul.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmul_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmul_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmul.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmul_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmul_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmul.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmulx_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmulx_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmulx.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmulx_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmulx_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmulx.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmulx_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmulx_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmulx.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fnmla_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fnmla_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fnmla.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fnmla_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fnmla_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fnmla.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fnmla_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fnmla_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fnmla.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fnmls_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fnmls_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fnmls.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fnmls_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fnmls_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fnmls.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fnmls_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fnmls_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fnmls.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+; aarch64_sve_fsub intrinsic combines to a LLVM instruction fsub.
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fsub_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fsub_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> [[A]]
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fsub.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fsub_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fsub_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> [[A]]
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fsub.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fsub_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fsub_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> [[A]]
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fsub.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+; Integer arithmetic
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_add_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_add_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.add.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_add_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_add_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.add.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_add_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_add_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.add.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_add_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_add_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.add.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_mla_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_mla_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]], <vscale x 16 x i8> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.mla.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_mla_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_mla_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]], <vscale x 8 x i16> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_mla_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_mla_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]], <vscale x 4 x i32> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_mla_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_mla_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]], <vscale x 2 x i64> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_mls_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_mls_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]], <vscale x 16 x i8> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.mls.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_mls_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_mls_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]], <vscale x 8 x i16> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mls.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_mls_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_mls_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]], <vscale x 4 x i32> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mls.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_mls_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_mls_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]], <vscale x 2 x i64> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mls.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_mul_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_mul_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.mul.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_mul_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_mul_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mul.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_mul_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_mul_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mul.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_mul_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_mul_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mul.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_sabd_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_sabd_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_sabd_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_sabd_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_sabd_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_sabd_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_sabd_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_sabd_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_smax_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_smax_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.smax.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_smax_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_smax_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.smax.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_smax_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_smax_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.smax.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_smax_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_smax_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.smax.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_smin_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_smin_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.smin.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_smin_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_smin_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.smin.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_smin_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_smin_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.smin.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_smin_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_smin_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.smin.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_smulh_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_smulh_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_smulh_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_smulh_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_smulh_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_smulh_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_smulh_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_smulh_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_sub_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_sub_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sub.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_sub_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_sub_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sub.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_sub_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_sub_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sub.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_sub_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_sub_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sub.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_uabd_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_uabd_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_uabd_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_uabd_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_uabd_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_uabd_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_uabd_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_uabd_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_umax_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_umax_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.umax.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_umax_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_umax_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.umax.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_umax_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_umax_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.umax.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_umax_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_umax_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.umax.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_umin_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_umin_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.umin.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_umin_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_umin_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.umin.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_umin_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_umin_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.umin.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_umin_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_umin_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.umin.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_umulh_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_umulh_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_umulh_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_umulh_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_umulh_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_umulh_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_umulh_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_umulh_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+; Shifts
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_asr_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_asr_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.asr.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_asr_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_asr_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.asr.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_asr_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_asr_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.asr.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_asr_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_asr_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.asr.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_lsl_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_lsl_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_lsl_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_lsl_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_lsl_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_lsl_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_lsl_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_lsl_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_lsr_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_lsr_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_lsr_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_lsr_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_lsr_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_lsr_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_lsr_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_lsr_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+; Logical operations
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_and_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_and_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.and.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_and_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_and_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.and.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_and_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_and_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.and.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_and_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_and_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.and.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_bic_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_bic_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.bic.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_bic_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_bic_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.bic.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_bic_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_bic_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.bic.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_bic_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_bic_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.bic.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_eor_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_eor_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.eor.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_eor_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_eor_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.eor.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_eor_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_eor_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.eor.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_eor_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_eor_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.eor.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_orr_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_orr_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.orr.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_orr_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_orr_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.orr.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_orr_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_orr_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.orr.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_orr_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_orr_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.orr.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+; SVE2 - Uniform DSP operations
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_sqsub_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_sqsub_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_sqsub_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_sqsub_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_sqsub_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_sqsub_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_sqsub_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_sqsub_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_uqsub_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_uqsub_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> [[A]]
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_uqsub_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_uqsub_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> [[A]]
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_uqsub_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_uqsub_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> [[A]]
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_uqsub_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_uqsub_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> [[A]]
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+attributes #0 = { "target-features"="+sve,+sve2" }
diff --git a/llvm/test/Transforms/InstCombine/AArch64/sve-intr-comb-u-forms-no-active-lanes.ll b/llvm/test/Transforms/InstCombine/AArch64/sve-intr-comb-u-forms-no-active-lanes.ll
new file mode 100644
--- /dev/null
+++ b/llvm/test/Transforms/InstCombine/AArch64/sve-intr-comb-u-forms-no-active-lanes.ll
@@ -0,0 +1,1318 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 2
+; RUN: opt -S -passes=instcombine < %s | FileCheck %s
+
+target triple = "aarch64-unknown-linux-gnu"
+
+; Replace SVE _u intrinsics with undef if the predicate is all false.
+
+; Float arithmetic
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fabd.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fabd_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fabd_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1:[0-9]+]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fabd.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fabd.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fabd_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fabd_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fabd.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fabd.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fabd_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fabd_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fabd.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fadd.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fadd_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fadd_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fadd.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fadd.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fadd_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fadd_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fadd.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fadd.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fadd_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fadd_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fadd.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fdiv.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fdiv_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fdiv_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fdiv.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fdiv.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fdiv_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fdiv_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fdiv.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fdiv.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fdiv_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fdiv_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fdiv.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmax.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmax_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmax_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmax.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmax.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmax_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmax_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmax.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmax.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmax_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmax_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmax.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmaxnm_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmaxnm_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmaxnm.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmaxnm_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmaxnm_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmaxnm.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmaxnm_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmaxnm_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmaxnm.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmin.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmin_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmin_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmin.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmin.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmin_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmin_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmin.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmin.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmin_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmin_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmin.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fminnm.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fminnm_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fminnm_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fminnm.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fminnm.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fminnm_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fminnm_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fminnm.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fminnm.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fminnm_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fminnm_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fminnm.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmla.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmla_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmla_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmla.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmla.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmla_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmla_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmla.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmla.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmla_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmla_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmla.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmls.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmls_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmls_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmls.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmls.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmls_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmls_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmls.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmls.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmls_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmls_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmls.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmul.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmul_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmul_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmul.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmul.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmul_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmul_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmul.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmul.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmul_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmul_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmul.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fmulx.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fmulx_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fmulx_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fmulx.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fmulx.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fmulx_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fmulx_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fmulx.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fmulx.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fmulx_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fmulx_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fmulx.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fnmla.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fnmla_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fnmla_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fnmla.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fnmla.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fnmla_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fnmla_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fnmla.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fnmla.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fnmla_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fnmla_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fnmla.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fnmls.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fnmls_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fnmls_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]], <vscale x 8 x half> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fnmls.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x half> %c)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fnmls.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fnmls_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fnmls_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]], <vscale x 4 x float> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fnmls.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x float> %c)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fnmls.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fnmls_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fnmls_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]], <vscale x 2 x double> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fnmls.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b, <vscale x 2 x double> %c)
+  ret <vscale x 2 x double> %1
+}
+
+declare <vscale x 8 x half> @llvm.aarch64.sve.fsub.u.nxv8f16(<vscale x 8 x i1>, <vscale x 8 x half>, <vscale x 8 x half>)
+define <vscale x 8 x half> @replace_fsub_intrinsic_half(<vscale x 8 x half> %a, <vscale x 8 x half> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x half> @replace_fsub_intrinsic_half
+; CHECK-SAME: (<vscale x 8 x half> [[A:%.*]], <vscale x 8 x half> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x half> undef
+;
+  %1 = tail call fast <vscale x 8 x half> @llvm.aarch64.sve.fsub.u.nxv8f16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x half> %a, <vscale x 8 x half> %b)
+  ret <vscale x 8 x half> %1
+}
+
+declare <vscale x 4 x float> @llvm.aarch64.sve.fsub.u.nxv4f32(<vscale x 4 x i1>, <vscale x 4 x float>, <vscale x 4 x float>)
+define <vscale x 4 x float> @replace_fsub_intrinsic_float(<vscale x 4 x float> %a, <vscale x 4 x float> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x float> @replace_fsub_intrinsic_float
+; CHECK-SAME: (<vscale x 4 x float> [[A:%.*]], <vscale x 4 x float> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x float> undef
+;
+  %1 = tail call fast <vscale x 4 x float> @llvm.aarch64.sve.fsub.u.nxv4f32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x float> %a, <vscale x 4 x float> %b)
+  ret <vscale x 4 x float> %1
+}
+
+declare <vscale x 2 x double> @llvm.aarch64.sve.fsub.u.nxv2f64(<vscale x 2 x i1>, <vscale x 2 x double>, <vscale x 2 x double>)
+define <vscale x 2 x double> @replace_fsub_intrinsic_double(<vscale x 2 x double> %a, <vscale x 2 x double> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x double> @replace_fsub_intrinsic_double
+; CHECK-SAME: (<vscale x 2 x double> [[A:%.*]], <vscale x 2 x double> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x double> undef
+;
+  %1 = tail call fast <vscale x 2 x double> @llvm.aarch64.sve.fsub.u.nxv2f64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x double> %a, <vscale x 2 x double> %b)
+  ret <vscale x 2 x double> %1
+}
+
+; Integer arithmetic
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.add.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_add_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_add_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.add.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.add.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_add_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_add_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.add.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.add.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_add_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_add_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.add.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.add.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_add_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_add_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.add.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.mla.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_mla_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_mla_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]], <vscale x 16 x i8> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.mla.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.mla.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_mla_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_mla_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]], <vscale x 8 x i16> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mla.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.mla.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_mla_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_mla_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]], <vscale x 4 x i32> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mla.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.mla.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_mla_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_mla_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]], <vscale x 2 x i64> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mla.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.mls.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_mls_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_mls_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]], <vscale x 16 x i8> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.mls.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b, <vscale x 16 x i8> %c)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.mls.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_mls_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_mls_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]], <vscale x 8 x i16> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mls.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b, <vscale x 8 x i16> %c)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.mls.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_mls_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_mls_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]], <vscale x 4 x i32> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mls.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b, <vscale x 4 x i32> %c)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.mls.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_mls_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_mls_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]], <vscale x 2 x i64> [[C:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mls.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b, <vscale x 2 x i64> %c)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.mul.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_mul_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_mul_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.mul.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.mul.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_mul_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_mul_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.mul.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.mul.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_mul_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_mul_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.mul.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.mul.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_mul_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_mul_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.mul.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.sabd.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_sabd_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_sabd_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sabd.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.sabd.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_sabd_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_sabd_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sabd.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.sabd.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_sabd_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_sabd_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sabd.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.sabd.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_sabd_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_sabd_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sabd.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.smax.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_smax_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_smax_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.smax.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.smax.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_smax_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_smax_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.smax.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.smax.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_smax_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_smax_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.smax.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.smax.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_smax_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_smax_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.smax.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.smin.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_smin_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_smin_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.smin.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.smin.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_smin_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_smin_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.smin.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.smin.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_smin_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_smin_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.smin.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.smin.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_smin_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_smin_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.smin.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.smulh.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_smulh_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_smulh_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.smulh.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.smulh.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_smulh_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_smulh_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.smulh.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.smulh.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_smulh_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_smulh_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.smulh.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.smulh.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_smulh_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_smulh_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.smulh.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.sub.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_sub_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_sub_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sub.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.sub.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_sub_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_sub_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sub.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.sub.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_sub_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_sub_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sub.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.sub.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_sub_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_sub_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sub.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.uabd.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_uabd_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_uabd_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uabd.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.uabd.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_uabd_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_uabd_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uabd.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.uabd.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_uabd_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_uabd_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uabd.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.uabd.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_uabd_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_uabd_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uabd.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.umax.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_umax_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_umax_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.umax.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.umax.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_umax_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_umax_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.umax.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.umax.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_umax_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_umax_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.umax.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.umax.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_umax_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_umax_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.umax.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.umin.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_umin_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_umin_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.umin.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.umin.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_umin_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_umin_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.umin.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.umin.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_umin_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_umin_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.umin.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.umin.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_umin_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_umin_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.umin.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.umulh.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_umulh_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_umulh_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.umulh.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.umulh.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_umulh_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_umulh_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.umulh.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.umulh.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_umulh_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_umulh_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.umulh.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.umulh.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_umulh_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_umulh_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.umulh.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+; Shifts
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.asr.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_asr_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_asr_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.asr.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.asr.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_asr_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_asr_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.asr.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.asr.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_asr_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_asr_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.asr.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.asr.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_asr_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_asr_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.asr.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.lsl.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_lsl_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_lsl_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.lsl.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.lsl.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_lsl_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_lsl_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.lsl.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.lsl.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_lsl_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_lsl_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.lsl.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.lsl.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_lsl_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_lsl_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.lsl.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.lsr.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_lsr_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_lsr_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.lsr.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.lsr.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_lsr_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_lsr_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.lsr.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.lsr.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_lsr_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_lsr_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.lsr.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.lsr.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_lsr_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_lsr_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.lsr.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+; Logical operations
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.and.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_and_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_and_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.and.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.and.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_and_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_and_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.and.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.and.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_and_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_and_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.and.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.and.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_and_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_and_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.and.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.bic.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_bic_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_bic_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.bic.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.bic.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_bic_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_bic_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.bic.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.bic.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_bic_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_bic_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.bic.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.bic.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_bic_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_bic_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.bic.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.eor.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_eor_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_eor_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.eor.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.eor.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_eor_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_eor_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.eor.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.eor.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_eor_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_eor_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.eor.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.eor.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_eor_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_eor_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.eor.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.orr.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_orr_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_orr_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.orr.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.orr.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_orr_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_orr_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.orr.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.orr.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_orr_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_orr_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.orr.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.orr.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_orr_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_orr_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.orr.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+; SVE2 - Uniform DSP operations
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_sqsub_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_sqsub_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.sqsub.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_sqsub_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_sqsub_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.sqsub.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_sqsub_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_sqsub_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.sqsub.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_sqsub_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_sqsub_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.sqsub.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+declare <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1>, <vscale x 16 x i8>, <vscale x 16 x i8>)
+define <vscale x 16 x i8> @replace_uqsub_intrinsic_i8(<vscale x 16 x i8> %a, <vscale x 16 x i8> %b) #0 {
+; CHECK-LABEL: define <vscale x 16 x i8> @replace_uqsub_intrinsic_i8
+; CHECK-SAME: (<vscale x 16 x i8> [[A:%.*]], <vscale x 16 x i8> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 16 x i8> undef
+;
+  %1 = tail call <vscale x 16 x i8> @llvm.aarch64.sve.uqsub.u.nxv16i8(<vscale x 16 x i1> zeroinitializer, <vscale x 16 x i8> %a, <vscale x 16 x i8> %b)
+  ret <vscale x 16 x i8> %1
+}
+
+declare <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1>, <vscale x 8 x i16>, <vscale x 8 x i16>)
+define <vscale x 8 x i16> @replace_uqsub_intrinsic_i16(<vscale x 8 x i16> %a, <vscale x 8 x i16> %b) #0 {
+; CHECK-LABEL: define <vscale x 8 x i16> @replace_uqsub_intrinsic_i16
+; CHECK-SAME: (<vscale x 8 x i16> [[A:%.*]], <vscale x 8 x i16> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 8 x i16> undef
+;
+  %1 = tail call <vscale x 8 x i16> @llvm.aarch64.sve.uqsub.u.nxv8i16(<vscale x 8 x i1> zeroinitializer, <vscale x 8 x i16> %a, <vscale x 8 x i16> %b)
+  ret <vscale x 8 x i16> %1
+}
+
+declare <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1>, <vscale x 4 x i32>, <vscale x 4 x i32>)
+define <vscale x 4 x i32> @replace_uqsub_intrinsic_i32(<vscale x 4 x i32> %a, <vscale x 4 x i32> %b) #0 {
+; CHECK-LABEL: define <vscale x 4 x i32> @replace_uqsub_intrinsic_i32
+; CHECK-SAME: (<vscale x 4 x i32> [[A:%.*]], <vscale x 4 x i32> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 4 x i32> undef
+;
+  %1 = tail call <vscale x 4 x i32> @llvm.aarch64.sve.uqsub.u.nxv4i32(<vscale x 4 x i1> zeroinitializer, <vscale x 4 x i32> %a, <vscale x 4 x i32> %b)
+  ret <vscale x 4 x i32> %1
+}
+
+declare <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1>, <vscale x 2 x i64>, <vscale x 2 x i64>)
+define <vscale x 2 x i64> @replace_uqsub_intrinsic_i64(<vscale x 2 x i64> %a, <vscale x 2 x i64> %b) #0 {
+; CHECK-LABEL: define <vscale x 2 x i64> @replace_uqsub_intrinsic_i64
+; CHECK-SAME: (<vscale x 2 x i64> [[A:%.*]], <vscale x 2 x i64> [[B:%.*]]) #[[ATTR1]] {
+; CHECK-NEXT:    ret <vscale x 2 x i64> undef
+;
+  %1 = tail call <vscale x 2 x i64> @llvm.aarch64.sve.uqsub.u.nxv2i64(<vscale x 2 x i1> zeroinitializer, <vscale x 2 x i64> %a, <vscale x 2 x i64> %b)
+  ret <vscale x 2 x i64> %1
+}
+
+attributes #0 = { "target-features"="+sve,+sve2" }