[RISCV] Add intrinsics for vfmv.f.s and vfmv.s.f

[craig.topper]

Summary

Similar to D93365, but for floating point. No need for special ISD opcodes though. We can directly isel these from intrinsics. I had to use anyfloat_ty instead of anyvector_ty in the intrinsics to make LLVMVectorElementType not crash when imported into the -gen-dag-isel tablegen backend.

[HsiangKai]

LGTM.

[frasercrmck]

llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td

1378

Is this the opposite of ToFPR32? Should we thus define FromFPR32? At any rate we can presumably defvar inst = !cast<Instruction>... to reduce code duplication.

[craig.topper]

Use defvar to simplify code.

Not sure if makes sense to add FromFPR32. This might be the only instruction that goes that direction.

[frasercrmck]

In D93426#2459871, @craig.topper wrote:

Use defvar to simplify code.

Not sure if makes sense to add FromFPR32. This might be the only instruction that goes that direction.

Fair enough. We can create a helper out of it if and when it's required.

llvm/include/llvm/IR/IntrinsicsRISCV.td

def int_riscv_vmv_x_s : Intrinsic<[LLVMVectorElementType<0>],
                                     [llvm_anyint_ty],
                                     [IntrNoMem]>, RISCVVIntrinsic;
   def int_riscv_vmv_s_x : Intrinsic<[llvm_anyint_ty],
                                     [LLVMMatchType<0>, LLVMVectorElementType<0>,
                                      llvm_anyint_ty],
                                     [IntrNoMem]>, RISCVVIntrinsic {
     let ExtendOperand = 2;
   }
+
+  def int_riscv_vfmv_f_s : Intrinsic<[LLVMVectorElementType<0>],
+                                     [llvm_anyfloat_ty],
+                                     [IntrNoMem]>, RISCVVIntrinsic;
+  def int_riscv_vfmv_s_f : Intrinsic<[llvm_anyfloat_ty],
+                                     [LLVMMatchType<0>, LLVMVectorElementType<0>,
+                                      llvm_anyint_ty],
+                                     [IntrNoMem]>, RISCVVIntrinsic;
 } // TargetPrefix = "riscv"

llvm/lib/Target/RISCV/RISCVInstrInfoV.td

 } // hasSideEffects = 0, mayLoad = 0, mayStore = 0
 } // Predicates = [HasStdExtV]
 
 let Predicates = [HasStdExtV, HasStdExtF] in {
 let hasSideEffects = 0, mayLoad = 0, mayStore = 0, vm = 1 in {
 // Floating-Point Scalar Move Instructions
 def VFMV_F_S : RVInstV<0b010000, 0b00000, OPFVV, (outs FPR32:$vd),
                       (ins VR:$vs2), "vfmv.f.s", "$vd, $vs2">;
-def VFMV_S_F : RVInstV2<0b010000, 0b00000, OPFVF, (outs VR:$vd),
-                      (ins FPR32:$rs1), "vfmv.s.f", "$vd, $rs1">;
+let Constraints = "$vd = $vd_wb" in
+def VFMV_S_F : RVInstV2<0b010000, 0b00000, OPFVF, (outs VR:$vd_wb),
+                      (ins VR:$vd, FPR32:$rs1), "vfmv.s.f", "$vd, $rs1">;
 
 } // hasSideEffects = 0, mayLoad = 0, mayStore = 0, vm = 1
 } // Predicates = [HasStdExtV, HasStdExtF]
 
 let Predicates = [HasStdExtV] in {
 // Vector Slide Instructions
 let Constraints = "@earlyclobber $vd", RVVConstraint = SlideUp in {
 defm VSLIDEUP_V : VALU_IV_X_I<"vslideup", 0b001110, uimm5>;

llvm/lib/Target/RISCV/RISCVInstrInfoVPseudos.td

let VLMul = m.value in {
       let SEWIndex = 2, BaseInstr = VMV_X_S in
       def PseudoVMV_X_S # "_" # m.MX: Pseudo<(outs GPR:$rd),
                                              (ins m.vrclass:$rs2, ixlenimm:$sew),
                                              []>, RISCVVPseudo;
       let VLIndex = 3, SEWIndex = 4, BaseInstr = VMV_S_X,
           Constraints = "$rd = $rs1" in
       def PseudoVMV_S_X # "_" # m.MX: Pseudo<(outs m.vrclass:$rd),
                                              (ins m.vrclass:$rs1, GPR:$rs2,
                                                   GPR:$vl, ixlenimm:$sew),
                                              []>, RISCVVPseudo;
     }
   }
 }
+} // Predicates = [HasStdExtV]
+
+//===----------------------------------------------------------------------===//
+// 17.2. Floating-Point Scalar Move Instructions
+//===----------------------------------------------------------------------===//
+
+let Predicates = [HasStdExtV, HasStdExtF] in {
+let mayLoad = 0, mayStore = 0, hasSideEffects = 0, usesCustomInserter = 1,
+    Uses = [VL, VTYPE] in {
+  foreach m = MxList.m in {
+    let VLMul = m.value in {
+      let SEWIndex = 2, BaseInstr = VFMV_F_S in
+      def PseudoVFMV_F_S # "_" # m.MX : Pseudo<(outs FPR32:$rd),
+                                               (ins m.vrclass:$rs2,
+                                                    ixlenimm:$sew),
+                                               []>, RISCVVPseudo;
+      let VLIndex = 3, SEWIndex = 4, BaseInstr = VFMV_S_F,
+          Constraints = "$rd = $rs1" in
+      def PseudoVFMV_S_F # "_" # m.MX : Pseudo<(outs m.vrclass:$rd),
+                                               (ins m.vrclass:$rs1, FPR32:$rs2,
+                                                    GPR:$vl, ixlenimm:$sew),
+                                               []>, RISCVVPseudo;
+    }
   }
+}
+} // Predicates = [HasStdExtV, HasStdExtF]
 
 //===----------------------------------------------------------------------===//
 // Patterns.
 //===----------------------------------------------------------------------===//
 let Predicates = [HasStdExtV] in {
 
 // Whole-register vector patterns.
 
foreach vti = AllIntegerVectors in {
   def : Pat<(riscv_vmv_x_s (vti.Vector vti.RegClass:$rs2)),
             (!cast<Instruction>("PseudoVMV_X_S_" # vti.LMul.MX) $rs2, vti.SEW)>;
   def : Pat<(vti.Vector (int_riscv_vmv_s_x (vti.Vector vti.RegClass:$rs1),
                                            GPR:$rs2, GPR:$vl)),
             (!cast<Instruction>("PseudoVMV_S_X_" # vti.LMul.MX)
              (vti.Vector $rs1), $rs2, (NoX0 GPR:$vl), vti.SEW)>;
 }
 } // Predicates = [HasStdExtV]
+
+//===----------------------------------------------------------------------===//
+// 17.2. Floating-Point Scalar Move Instructions
+//===----------------------------------------------------------------------===//
+
+let Predicates = [HasStdExtV, HasStdExtF] in {
+foreach fvti = AllFloatVectors in {
+  defvar instr = !cast<Instruction>("PseudoVFMV_F_S_" # fvti.LMul.MX);
+  def : Pat<(fvti.Scalar (int_riscv_vfmv_f_s (fvti.Vector fvti.RegClass:$rs2))),
+             // Floating point instructions with a scalar result will always
+             // generate the result in a register of class FPR32. When dealing
+             // with the f64 variant of a pattern we need to promote the FPR32
+             // subregister generated by the instruction to the FPR64 base
+             // register expected by the type in the pattern
+             !cond(!eq(!cast<string>(fvti.ScalarRegClass),
+                       !cast<string>(FPR64)):
+                      (SUBREG_TO_REG (i32 -1),
+                                     (instr $rs2, fvti.SEW), sub_32),
+                   !eq(!cast<string>(fvti.ScalarRegClass),
+                       !cast<string>(FPR16)):
+                      (EXTRACT_SUBREG (instr $rs2, fvti.SEW), sub_16),
+                   !eq(1, 1):
+                      (instr $rs2, fvti.SEW))>;
+
+  def : Pat<(fvti.Vector (int_riscv_vfmv_s_f (fvti.Vector fvti.RegClass:$rs1),
+                         (fvti.Scalar fvti.ScalarRegClass:$rs2), GPR:$vl)),
+            (!cast<Instruction>("PseudoVFMV_S_F_" # fvti.LMul.MX)
+             (fvti.Vector $rs1), ToFPR32<fvti.Scalar, fvti.ScalarRegClass, "rs2">.ret,
+             (NoX0 GPR:$vl), fvti.SEW)>;
+}
+} // Predicates = [HasStdExtV, HasStdExtF]

llvm/test/CodeGen/RISCV/rvv/vfmv.f.s.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv64 -mattr=+d,+experimental-v,+experimental-zfh -target-abi lp64d -verify-machineinstrs < %s | FileCheck %s
+; RUN: llc -mtriple=riscv32 -mattr=+d,+experimental-v,+experimental-zfh -target-abi ilp32d -verify-machineinstrs < %s | FileCheck %s
+
+declare half @llvm.riscv.vfmv.f.s.nxv1f16(<vscale x 1 x half>)
+
+define half @intrinsic_vfmv.f.s_s_nxv1f16(<vscale x 1 x half> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv1f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e16,mf4,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h killed $f10_f
+; CHECK-NEXT:    ret
+entry:
+  %a = call half @llvm.riscv.vfmv.f.s.nxv1f16(<vscale x 1 x half> %0)
+  ret half %a
+}
+
+declare half @llvm.riscv.vfmv.f.s.nxv2f16(<vscale x 2 x half>)
+
+define half @intrinsic_vfmv.f.s_s_nxv2f16(<vscale x 2 x half> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv2f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e16,mf2,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h killed $f10_f
+; CHECK-NEXT:    ret
+entry:
+  %a = call half @llvm.riscv.vfmv.f.s.nxv2f16(<vscale x 2 x half> %0)
+  ret half %a
+}
+
+declare half @llvm.riscv.vfmv.f.s.nxv4f16(<vscale x 4 x half>)
+
+define half @intrinsic_vfmv.f.s_s_nxv4f16(<vscale x 4 x half> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv4f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e16,m1,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h killed $f10_f
+; CHECK-NEXT:    ret
+entry:
+  %a = call half @llvm.riscv.vfmv.f.s.nxv4f16(<vscale x 4 x half> %0)
+  ret half %a
+}
+
+declare half @llvm.riscv.vfmv.f.s.nxv8f16(<vscale x 8 x half>)
+
+define half @intrinsic_vfmv.f.s_s_nxv8f16(<vscale x 8 x half> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv8f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e16,m2,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h killed $f10_f
+; CHECK-NEXT:    ret
+entry:
+  %a = call half @llvm.riscv.vfmv.f.s.nxv8f16(<vscale x 8 x half> %0)
+  ret half %a
+}
+
+declare half @llvm.riscv.vfmv.f.s.nxv16f16(<vscale x 16 x half>)
+
+define half @intrinsic_vfmv.f.s_s_nxv16f16(<vscale x 16 x half> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv16f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e16,m4,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h killed $f10_f
+; CHECK-NEXT:    ret
+entry:
+  %a = call half @llvm.riscv.vfmv.f.s.nxv16f16(<vscale x 16 x half> %0)
+  ret half %a
+}
+
+declare half @llvm.riscv.vfmv.f.s.nxv32f16(<vscale x 32 x half>)
+
+define half @intrinsic_vfmv.f.s_s_nxv32f16(<vscale x 32 x half> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv32f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e16,m8,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h killed $f10_f
+; CHECK-NEXT:    ret
+entry:
+  %a = call half @llvm.riscv.vfmv.f.s.nxv32f16(<vscale x 32 x half> %0)
+  ret half %a
+}
+
+declare float @llvm.riscv.vfmv.f.s.nxv1f32(<vscale x 1 x float>)
+
+define float @intrinsic_vfmv.f.s_s_nxv1f32(<vscale x 1 x float> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv1f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e32,mf2,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call float @llvm.riscv.vfmv.f.s.nxv1f32(<vscale x 1 x float> %0)
+  ret float %a
+}
+
+declare float @llvm.riscv.vfmv.f.s.nxv2f32(<vscale x 2 x float>)
+
+define float @intrinsic_vfmv.f.s_s_nxv2f32(<vscale x 2 x float> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv2f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e32,m1,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call float @llvm.riscv.vfmv.f.s.nxv2f32(<vscale x 2 x float> %0)
+  ret float %a
+}
+
+declare float @llvm.riscv.vfmv.f.s.nxv4f32(<vscale x 4 x float>)
+
+define float @intrinsic_vfmv.f.s_s_nxv4f32(<vscale x 4 x float> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv4f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e32,m2,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call float @llvm.riscv.vfmv.f.s.nxv4f32(<vscale x 4 x float> %0)
+  ret float %a
+}
+
+declare float @llvm.riscv.vfmv.f.s.nxv8f32(<vscale x 8 x float>)
+
+define float @intrinsic_vfmv.f.s_s_nxv8f32(<vscale x 8 x float> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv8f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e32,m4,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call float @llvm.riscv.vfmv.f.s.nxv8f32(<vscale x 8 x float> %0)
+  ret float %a
+}
+
+declare float @llvm.riscv.vfmv.f.s.nxv16f32(<vscale x 16 x float>)
+
+define float @intrinsic_vfmv.f.s_s_nxv16f32(<vscale x 16 x float> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv16f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e32,m8,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call float @llvm.riscv.vfmv.f.s.nxv16f32(<vscale x 16 x float> %0)
+  ret float %a
+}
+
+declare double @llvm.riscv.vfmv.f.s.nxv1f64(<vscale x 1 x double>)
+
+define double @intrinsic_vfmv.f.s_s_nxv1f64(<vscale x 1 x double> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv1f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e64,m1,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call double @llvm.riscv.vfmv.f.s.nxv1f64(<vscale x 1 x double> %0)
+  ret double %a
+}
+
+declare double @llvm.riscv.vfmv.f.s.nxv2f64(<vscale x 2 x double>)
+
+define double @intrinsic_vfmv.f.s_s_nxv2f64(<vscale x 2 x double> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv2f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e64,m2,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call double @llvm.riscv.vfmv.f.s.nxv2f64(<vscale x 2 x double> %0)
+  ret double %a
+}
+
+declare double @llvm.riscv.vfmv.f.s.nxv4f64(<vscale x 4 x double>)
+
+define double @intrinsic_vfmv.f.s_s_nxv4f64(<vscale x 4 x double> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv4f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e64,m4,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call double @llvm.riscv.vfmv.f.s.nxv4f64(<vscale x 4 x double> %0)
+  ret double %a
+}
+
+declare double @llvm.riscv.vfmv.f.s.nxv8f64(<vscale x 8 x double>)
+
+define double @intrinsic_vfmv.f.s_s_nxv8f64(<vscale x 8 x double> %0) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.f.s_s_nxv8f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli zero, zero, e64,m8,ta,mu
+; CHECK-NEXT:    vfmv.f.s fa0, v16
+; CHECK-NEXT:    ret
+entry:
+  %a = call double @llvm.riscv.vfmv.f.s.nxv8f64(<vscale x 8 x double> %0)
+  ret double %a
+}

llvm/test/CodeGen/RISCV/rvv/vfmv.s.f-rv32.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv32 -mattr=+d,+experimental-v,+experimental-zfh -target-abi ilp32d -verify-machineinstrs < %s | FileCheck %s
+
+declare <vscale x 1 x half> @llvm.riscv.vfmv.s.f.nxv1f16(<vscale x 1 x half>, half, i32)
+
+define <vscale x 1 x half> @intrinsic_vfmv.s.f_f_nxv1f16(<vscale x 1 x half> %0, half %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x half> @llvm.riscv.vfmv.s.f.nxv1f16(<vscale x 1 x half> %0, half %1, i32 %2)
+  ret <vscale x 1 x half> %a
+}
+
+declare <vscale x 2 x half> @llvm.riscv.vfmv.s.f.nxv2f16(<vscale x 2 x half>, half, i32)
+
+define <vscale x 2 x half> @intrinsic_vfmv.s.f_f_nxv2f16(<vscale x 2 x half> %0, half %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv2f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x half> @llvm.riscv.vfmv.s.f.nxv2f16(<vscale x 2 x half> %0, half %1, i32 %2)
+  ret <vscale x 2 x half> %a
+}
+
+declare <vscale x 4 x half> @llvm.riscv.vfmv.s.f.nxv4f16(<vscale x 4 x half>, half, i32)
+
+define <vscale x 4 x half> @intrinsic_vfmv.s.f_f_nxv4f16(<vscale x 4 x half> %0, half %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv4f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m1,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x half> @llvm.riscv.vfmv.s.f.nxv4f16(<vscale x 4 x half> %0, half %1, i32 %2)
+  ret <vscale x 4 x half> %a
+}
+
+declare <vscale x 8 x half> @llvm.riscv.vfmv.s.f.nxv8f16(<vscale x 8 x half>, half, i32)
+
+define <vscale x 8 x half> @intrinsic_vfmv.s.f_f_nxv8f16(<vscale x 8 x half> %0, half %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv8f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x half> @llvm.riscv.vfmv.s.f.nxv8f16(<vscale x 8 x half> %0, half %1, i32 %2)
+  ret <vscale x 8 x half> %a
+}
+
+declare <vscale x 16 x half> @llvm.riscv.vfmv.s.f.nxv16f16(<vscale x 16 x half>, half, i32)
+
+define <vscale x 16 x half> @intrinsic_vfmv.s.f_f_nxv16f16(<vscale x 16 x half> %0, half %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv16f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x half> @llvm.riscv.vfmv.s.f.nxv16f16(<vscale x 16 x half> %0, half %1, i32 %2)
+  ret <vscale x 16 x half> %a
+}
+
+declare <vscale x 32 x half> @llvm.riscv.vfmv.s.f.nxv32f16(<vscale x 32 x half>, half, i32)
+
+define <vscale x 32 x half> @intrinsic_vfmv.s.f_f_nxv32f16(<vscale x 32 x half> %0, half %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv32f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m8,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x half> @llvm.riscv.vfmv.s.f.nxv32f16(<vscale x 32 x half> %0, half %1, i32 %2)
+  ret <vscale x 32 x half> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfmv.s.f.nxv1f32(<vscale x 1 x float>, float, i32)
+
+define <vscale x 1 x float> @intrinsic_vfmv.s.f_f_nxv1f32(<vscale x 1 x float> %0, float %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,mf2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfmv.s.f.nxv1f32(<vscale x 1 x float> %0, float %1, i32 %2)
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfmv.s.f.nxv2f32(<vscale x 2 x float>, float, i32)
+
+define <vscale x 2 x float> @intrinsic_vfmv.s.f_f_nxv2f32(<vscale x 2 x float> %0, float %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv2f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m1,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfmv.s.f.nxv2f32(<vscale x 2 x float> %0, float %1, i32 %2)
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfmv.s.f.nxv4f32(<vscale x 4 x float>, float, i32)
+
+define <vscale x 4 x float> @intrinsic_vfmv.s.f_f_nxv4f32(<vscale x 4 x float> %0, float %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv4f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfmv.s.f.nxv4f32(<vscale x 4 x float> %0, float %1, i32 %2)
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfmv.s.f.nxv8f32(<vscale x 8 x float>, float, i32)
+
+define <vscale x 8 x float> @intrinsic_vfmv.s.f_f_nxv8f32(<vscale x 8 x float> %0, float %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv8f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfmv.s.f.nxv8f32(<vscale x 8 x float> %0, float %1, i32 %2)
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfmv.s.f.nxv16f32(<vscale x 16 x float>, float, i32)
+
+define <vscale x 16 x float> @intrinsic_vfmv.s.f_f_nxv16f32(<vscale x 16 x float> %0, float %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv16f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m8,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfmv.s.f.nxv16f32(<vscale x 16 x float> %0, float %1, i32 %2)
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x double> @llvm.riscv.vfmv.s.f.nxv1f64(<vscale x 1 x double>, double, i32)
+
+define <vscale x 1 x double> @intrinsic_vfmv.s.f_f_nxv1f64(<vscale x 1 x double> %0, double %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m1,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x double> @llvm.riscv.vfmv.s.f.nxv1f64(<vscale x 1 x double> %0, double %1, i32 %2)
+  ret <vscale x 1 x double> %a
+}
+
+declare <vscale x 2 x double> @llvm.riscv.vfmv.s.f.nxv2f64(<vscale x 2 x double>, double, i32)
+
+define <vscale x 2 x double> @intrinsic_vfmv.s.f_f_nxv2f64(<vscale x 2 x double> %0, double %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv2f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x double> @llvm.riscv.vfmv.s.f.nxv2f64(<vscale x 2 x double> %0, double %1, i32 %2)
+  ret <vscale x 2 x double> %a
+}
+
+declare <vscale x 4 x double> @llvm.riscv.vfmv.s.f.nxv4f64(<vscale x 4 x double>, double, i32)
+
+define <vscale x 4 x double> @intrinsic_vfmv.s.f_f_nxv4f64(<vscale x 4 x double> %0, double %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv4f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x double> @llvm.riscv.vfmv.s.f.nxv4f64(<vscale x 4 x double> %0, double %1, i32 %2)
+  ret <vscale x 4 x double> %a
+}
+
+declare <vscale x 8 x double> @llvm.riscv.vfmv.s.f.nxv8f64(<vscale x 8 x double>, double, i32)
+
+define <vscale x 8 x double> @intrinsic_vfmv.s.f_f_nxv8f64(<vscale x 8 x double> %0, double %1, i32 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv8f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m8,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x double> @llvm.riscv.vfmv.s.f.nxv8f64(<vscale x 8 x double> %0, double %1, i32 %2)
+  ret <vscale x 8 x double> %a
+}

llvm/test/CodeGen/RISCV/rvv/vfmv.s.f-rv64.ll

+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=riscv64 -mattr=+d,+experimental-v,+experimental-zfh -target-abi lp64d -verify-machineinstrs < %s | FileCheck %s
+
+declare <vscale x 1 x half> @llvm.riscv.vfmv.s.f.nxv1f16(<vscale x 1 x half>, half, i64)
+
+define <vscale x 1 x half> @intrinsic_vfmv.s.f_f_nxv1f16(<vscale x 1 x half> %0, half %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x half> @llvm.riscv.vfmv.s.f.nxv1f16(<vscale x 1 x half> %0, half %1, i64 %2)
+  ret <vscale x 1 x half> %a
+}
+
+declare <vscale x 2 x half> @llvm.riscv.vfmv.s.f.nxv2f16(<vscale x 2 x half>, half, i64)
+
+define <vscale x 2 x half> @intrinsic_vfmv.s.f_f_nxv2f16(<vscale x 2 x half> %0, half %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv2f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,mf2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x half> @llvm.riscv.vfmv.s.f.nxv2f16(<vscale x 2 x half> %0, half %1, i64 %2)
+  ret <vscale x 2 x half> %a
+}
+
+declare <vscale x 4 x half> @llvm.riscv.vfmv.s.f.nxv4f16(<vscale x 4 x half>, half, i64)
+
+define <vscale x 4 x half> @intrinsic_vfmv.s.f_f_nxv4f16(<vscale x 4 x half> %0, half %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv4f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m1,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x half> @llvm.riscv.vfmv.s.f.nxv4f16(<vscale x 4 x half> %0, half %1, i64 %2)
+  ret <vscale x 4 x half> %a
+}
+
+declare <vscale x 8 x half> @llvm.riscv.vfmv.s.f.nxv8f16(<vscale x 8 x half>, half, i64)
+
+define <vscale x 8 x half> @intrinsic_vfmv.s.f_f_nxv8f16(<vscale x 8 x half> %0, half %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv8f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x half> @llvm.riscv.vfmv.s.f.nxv8f16(<vscale x 8 x half> %0, half %1, i64 %2)
+  ret <vscale x 8 x half> %a
+}
+
+declare <vscale x 16 x half> @llvm.riscv.vfmv.s.f.nxv16f16(<vscale x 16 x half>, half, i64)
+
+define <vscale x 16 x half> @intrinsic_vfmv.s.f_f_nxv16f16(<vscale x 16 x half> %0, half %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv16f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x half> @llvm.riscv.vfmv.s.f.nxv16f16(<vscale x 16 x half> %0, half %1, i64 %2)
+  ret <vscale x 16 x half> %a
+}
+
+declare <vscale x 32 x half> @llvm.riscv.vfmv.s.f.nxv32f16(<vscale x 32 x half>, half, i64)
+
+define <vscale x 32 x half> @intrinsic_vfmv.s.f_f_nxv32f16(<vscale x 32 x half> %0, half %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv32f16:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    # kill: def $f10_h killed $f10_h def $f10_f
+; CHECK-NEXT:    vsetvli a0, a0, e16,m8,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 32 x half> @llvm.riscv.vfmv.s.f.nxv32f16(<vscale x 32 x half> %0, half %1, i64 %2)
+  ret <vscale x 32 x half> %a
+}
+
+declare <vscale x 1 x float> @llvm.riscv.vfmv.s.f.nxv1f32(<vscale x 1 x float>, float, i64)
+
+define <vscale x 1 x float> @intrinsic_vfmv.s.f_f_nxv1f32(<vscale x 1 x float> %0, float %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,mf2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x float> @llvm.riscv.vfmv.s.f.nxv1f32(<vscale x 1 x float> %0, float %1, i64 %2)
+  ret <vscale x 1 x float> %a
+}
+
+declare <vscale x 2 x float> @llvm.riscv.vfmv.s.f.nxv2f32(<vscale x 2 x float>, float, i64)
+
+define <vscale x 2 x float> @intrinsic_vfmv.s.f_f_nxv2f32(<vscale x 2 x float> %0, float %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv2f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m1,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x float> @llvm.riscv.vfmv.s.f.nxv2f32(<vscale x 2 x float> %0, float %1, i64 %2)
+  ret <vscale x 2 x float> %a
+}
+
+declare <vscale x 4 x float> @llvm.riscv.vfmv.s.f.nxv4f32(<vscale x 4 x float>, float, i64)
+
+define <vscale x 4 x float> @intrinsic_vfmv.s.f_f_nxv4f32(<vscale x 4 x float> %0, float %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv4f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x float> @llvm.riscv.vfmv.s.f.nxv4f32(<vscale x 4 x float> %0, float %1, i64 %2)
+  ret <vscale x 4 x float> %a
+}
+
+declare <vscale x 8 x float> @llvm.riscv.vfmv.s.f.nxv8f32(<vscale x 8 x float>, float, i64)
+
+define <vscale x 8 x float> @intrinsic_vfmv.s.f_f_nxv8f32(<vscale x 8 x float> %0, float %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv8f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x float> @llvm.riscv.vfmv.s.f.nxv8f32(<vscale x 8 x float> %0, float %1, i64 %2)
+  ret <vscale x 8 x float> %a
+}
+
+declare <vscale x 16 x float> @llvm.riscv.vfmv.s.f.nxv16f32(<vscale x 16 x float>, float, i64)
+
+define <vscale x 16 x float> @intrinsic_vfmv.s.f_f_nxv16f32(<vscale x 16 x float> %0, float %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv16f32:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e32,m8,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 16 x float> @llvm.riscv.vfmv.s.f.nxv16f32(<vscale x 16 x float> %0, float %1, i64 %2)
+  ret <vscale x 16 x float> %a
+}
+
+declare <vscale x 1 x double> @llvm.riscv.vfmv.s.f.nxv1f64(<vscale x 1 x double>, double, i64)
+
+define <vscale x 1 x double> @intrinsic_vfmv.s.f_f_nxv1f64(<vscale x 1 x double> %0, double %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv1f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m1,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 1 x double> @llvm.riscv.vfmv.s.f.nxv1f64(<vscale x 1 x double> %0, double %1, i64 %2)
+  ret <vscale x 1 x double> %a
+}
+
+declare <vscale x 2 x double> @llvm.riscv.vfmv.s.f.nxv2f64(<vscale x 2 x double>, double, i64)
+
+define <vscale x 2 x double> @intrinsic_vfmv.s.f_f_nxv2f64(<vscale x 2 x double> %0, double %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv2f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m2,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 2 x double> @llvm.riscv.vfmv.s.f.nxv2f64(<vscale x 2 x double> %0, double %1, i64 %2)
+  ret <vscale x 2 x double> %a
+}
+
+declare <vscale x 4 x double> @llvm.riscv.vfmv.s.f.nxv4f64(<vscale x 4 x double>, double, i64)
+
+define <vscale x 4 x double> @intrinsic_vfmv.s.f_f_nxv4f64(<vscale x 4 x double> %0, double %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv4f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m4,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 4 x double> @llvm.riscv.vfmv.s.f.nxv4f64(<vscale x 4 x double> %0, double %1, i64 %2)
+  ret <vscale x 4 x double> %a
+}
+
+declare <vscale x 8 x double> @llvm.riscv.vfmv.s.f.nxv8f64(<vscale x 8 x double>, double, i64)
+
+define <vscale x 8 x double> @intrinsic_vfmv.s.f_f_nxv8f64(<vscale x 8 x double> %0, double %1, i64 %2) nounwind {
+; CHECK-LABEL: intrinsic_vfmv.s.f_f_nxv8f64:
+; CHECK:       # %bb.0: # %entry
+; CHECK-NEXT:    vsetvli a0, a0, e64,m8,ta,mu
+; CHECK-NEXT:    vfmv.s.f v16, fa0
+; CHECK-NEXT:    ret
+entry:
+  %a = call <vscale x 8 x double> @llvm.riscv.vfmv.s.f.nxv8f64(<vscale x 8 x double> %0, double %1, i64 %2)
+  ret <vscale x 8 x double> %a
+}