; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc -mtriple=riscv32 -mattr=+d,+zfh,+zvfh,+v,+m -target-abi=ilp32d \
; RUN:     -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,ZVFH
; RUN: llc -mtriple=riscv64 -mattr=+d,+zfh,+zvfh,+v,+m -target-abi=lp64d \
; RUN:     -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,ZVFH
; RUN: llc -mtriple=riscv32 -mattr=+d,+zfh,+zvfhmin,+v,+m -target-abi=ilp32d \
; RUN:     -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,ZVFHMIN
; RUN: llc -mtriple=riscv64 -mattr=+d,+zfh,+zvfhmin,+v,+m -target-abi=lp64d \
; RUN:     -verify-machineinstrs < %s | FileCheck %s --check-prefixes=CHECK,ZVFHMIN

define <vscale x 1 x float> @vfmacc_vv_nxv1f32(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x float> %c, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv1f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv1f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v11, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v11, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> %m, i32 %evl)
  %bext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %b, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %aext, <vscale x 1 x float> %bext, <vscale x 1 x float> %c, <vscale x 1 x i1> %m, i32 %evl)
  ret <vscale x 1 x float> %v
}

define <vscale x 1 x float> @vfmacc_vv_nxv1f32_unmasked(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x float> %c, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv1f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv1f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v11, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v11, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %b, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %aext, <vscale x 1 x float> %bext, <vscale x 1 x float> %c, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 1 x float> %v
}

define <vscale x 1 x float> @vfmacc_vv_nxv1f32_tu(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x float> %c, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv1f32_tu:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli zero, a0, e16, mf4, tu, mu
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9, v0.t
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv1f32_tu:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v11, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, tu, mu
; ZVFHMIN-NEXT:    vfmacc.vv v10, v11, v8, v0.t
; ZVFHMIN-NEXT:    vmv1r.v v8, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %b, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %aext, <vscale x 1 x float> %bext, <vscale x 1 x float> %c, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %u = call <vscale x 1 x float> @llvm.vp.merge.nxv1f32(<vscale x 1 x i1> %m, <vscale x 1 x float> %v, <vscale x 1 x float> %c, i32 %evl)
  ret <vscale x 1 x float> %u
}

; FIXME: Support this case?
define <vscale x 1 x float> @vfmacc_vv_nxv1f32_masked__tu(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x float> %c, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv1f32_masked__tu:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli zero, a0, e16, mf4, tu, mu
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9, v0.t
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv1f32_masked__tu:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v11, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, tu, mu
; ZVFHMIN-NEXT:    vfmacc.vv v10, v11, v8, v0.t
; ZVFHMIN-NEXT:    vmv1r.v v8, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> %m, i32 %evl)
  %bext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %b, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %aext, <vscale x 1 x float> %bext, <vscale x 1 x float> %c, <vscale x 1 x i1> %m, i32 %evl)
  %u = call <vscale x 1 x float> @llvm.vp.merge.nxv1f32(<vscale x 1 x i1> %m, <vscale x 1 x float> %v, <vscale x 1 x float> %c, i32 %evl)
  ret <vscale x 1 x float> %u
}

define <vscale x 1 x float> @vfmacc_vv_nxv1f32_unmasked_tu(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x float> %c, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv1f32_unmasked_tu:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli zero, a0, e16, mf4, tu, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv1f32_unmasked_tu:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v11, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, tu, ma
; ZVFHMIN-NEXT:    vfmacc.vv v10, v11, v8
; ZVFHMIN-NEXT:    vmv1r.v v8, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %b, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %aext, <vscale x 1 x float> %bext, <vscale x 1 x float> %c, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %u = call <vscale x 1 x float> @llvm.vp.merge.nxv1f32(<vscale x 1 x i1> splat (i1 -1), <vscale x 1 x float> %v, <vscale x 1 x float> %c, i32 %evl)
  ret <vscale x 1 x float> %u
}

define <vscale x 1 x float> @vfmacc_vf_nxv1f32(<vscale x 1 x half> %va, half %b, <vscale x 1 x float> %vc, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv1f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv1f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v10, fa5, v9
; ZVFHMIN-NEXT:    vmv1r.v v8, v10
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x i1> %m, i32 %evl)
  %vbext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %vb, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %vaext, <vscale x 1 x float> %vbext, <vscale x 1 x float> %vc, <vscale x 1 x i1> %m, i32 %evl)
  ret <vscale x 1 x float> %v
}

define <vscale x 1 x float> @vfmacc_vf_nxv1f32_commute(<vscale x 1 x half> %va, half %b, <vscale x 1 x float> %vc, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv1f32_commute:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv1f32_commute:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v10, fa5, v9
; ZVFHMIN-NEXT:    vmv1r.v v8, v10
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x i1> %m, i32 %evl)
  %vbext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %vb, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %vbext, <vscale x 1 x float> %vaext, <vscale x 1 x float> %vc, <vscale x 1 x i1> %m, i32 %evl)
  ret <vscale x 1 x float> %v
}

define <vscale x 1 x float> @vfmacc_vf_nxv1f32_unmasked(<vscale x 1 x half> %va, half %b, <vscale x 1 x float> %vc, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv1f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv1f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v10, fa5, v9
; ZVFHMIN-NEXT:    vmv1r.v v8, v10
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %vb, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %vaext, <vscale x 1 x float> %vbext, <vscale x 1 x float> %vc, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 1 x float> %v
}

define <vscale x 1 x float> @vfmacc_vf_nxv1f32_tu(<vscale x 1 x half> %va, half %b, <vscale x 1 x float> %vc, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv1f32_tu:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli zero, a0, e16, mf4, tu, mu
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8, v0.t
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv1f32_tu:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, tu, mu
; ZVFHMIN-NEXT:    vfmacc.vf v9, fa5, v10, v0.t
; ZVFHMIN-NEXT:    vmv1r.v v8, v9
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %vb, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %vaext, <vscale x 1 x float> %vbext, <vscale x 1 x float> %vc, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %u = call <vscale x 1 x float> @llvm.vp.merge.nxv1f32(<vscale x 1 x i1> %m, <vscale x 1 x float> %v, <vscale x 1 x float> %vc, i32 %evl)
  ret <vscale x 1 x float> %u
}

define <vscale x 1 x float> @vfmacc_vf_nxv1f32_commute_tu(<vscale x 1 x half> %va, half %b, <vscale x 1 x float> %vc, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv1f32_commute_tu:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli zero, a0, e16, mf4, tu, mu
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8, v0.t
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv1f32_commute_tu:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, tu, mu
; ZVFHMIN-NEXT:    vfmacc.vf v9, fa5, v10, v0.t
; ZVFHMIN-NEXT:    vmv1r.v v8, v9
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %vb, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %vbext, <vscale x 1 x float> %vaext, <vscale x 1 x float> %vc, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %u = call <vscale x 1 x float> @llvm.vp.merge.nxv1f32(<vscale x 1 x i1> %m, <vscale x 1 x float> %v, <vscale x 1 x float> %vc, i32 %evl)
  ret <vscale x 1 x float> %u
}

define <vscale x 1 x float> @vfmacc_vf_nxv1f32_unmasked_tu(<vscale x 1 x half> %va, half %b, <vscale x 1 x float> %vc, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv1f32_unmasked_tu:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli zero, a0, e16, mf4, tu, ma
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv1f32_unmasked_tu:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli zero, a0, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, tu, ma
; ZVFHMIN-NEXT:    vfmacc.vf v9, fa5, v10
; ZVFHMIN-NEXT:    vmv1r.v v8, v9
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 1 x half> %elt.head, <vscale x 1 x half> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %va, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %vb, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %vaext, <vscale x 1 x float> %vbext, <vscale x 1 x float> %vc, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %u = call <vscale x 1 x float> @llvm.vp.merge.nxv1f32(<vscale x 1 x i1> splat (i1 -1), <vscale x 1 x float> %v, <vscale x 1 x float> %vc, i32 %evl)
  ret <vscale x 1 x float> %u
}

define <vscale x 2 x float> @vfmacc_vv_nxv2f32(<vscale x 2 x half> %a, <vscale x 2 x half> %b, <vscale x 2 x float> %c, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv2f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv2f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v11, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v11, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> %m, i32 %evl)
  %bext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %b, <vscale x 2 x i1> %m, i32 %evl)
  %v = call <vscale x 2 x float> @llvm.vp.fma.nxv2f32(<vscale x 2 x float> %aext, <vscale x 2 x float> %bext, <vscale x 2 x float> %c, <vscale x 2 x i1> %m, i32 %evl)
  ret <vscale x 2 x float> %v
}

define <vscale x 2 x float> @vfmacc_vv_nxv2f32_unmasked(<vscale x 2 x half> %a, <vscale x 2 x half> %b, <vscale x 2 x float> %c, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv2f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv2f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v11, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v9
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v11, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %b, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 2 x float> @llvm.vp.fma.nxv2f32(<vscale x 2 x float> %aext, <vscale x 2 x float> %bext, <vscale x 2 x float> %c, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 2 x float> %v
}

define <vscale x 2 x float> @vfmacc_vf_nxv2f32(<vscale x 2 x half> %va, half %b, <vscale x 2 x float> %vc, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv2f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv2f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v10, fa5, v9
; ZVFHMIN-NEXT:    vmv.v.v v8, v10
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 2 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 2 x half> %elt.head, <vscale x 2 x half> poison, <vscale x 2 x i32> zeroinitializer
  %vaext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %va, <vscale x 2 x i1> %m, i32 %evl)
  %vbext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %vb, <vscale x 2 x i1> %m, i32 %evl)
  %v = call <vscale x 2 x float> @llvm.vp.fma.nxv2f32(<vscale x 2 x float> %vaext, <vscale x 2 x float> %vbext, <vscale x 2 x float> %vc, <vscale x 2 x i1> %m, i32 %evl)
  ret <vscale x 2 x float> %v
}

define <vscale x 2 x float> @vfmacc_vf_nxv2f32_unmasked(<vscale x 2 x half> %va, half %b, <vscale x 2 x float> %vc, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv2f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v9, fa0, v8
; ZVFH-NEXT:    vmv1r.v v8, v9
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv2f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v10, v8
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v10, fa5, v9
; ZVFHMIN-NEXT:    vmv.v.v v8, v10
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 2 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 2 x half> %elt.head, <vscale x 2 x half> poison, <vscale x 2 x i32> zeroinitializer
  %vaext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %va, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 2 x float> @llvm.vp.fpext.nxv2f32.nxv2f16(<vscale x 2 x half> %vb, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 2 x float> @llvm.vp.fma.nxv2f32(<vscale x 2 x float> %vaext, <vscale x 2 x float> %vbext, <vscale x 2 x float> %vc, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 2 x float> %v
}

define <vscale x 4 x float> @vfmacc_vv_nxv4f32(<vscale x 4 x half> %a, <vscale x 4 x half> %b, <vscale x 4 x float> %c, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv4f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9
; ZVFH-NEXT:    vmv2r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv4f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT:    vmv1r.v v12, v9
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v14, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v12
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v14, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %a, <vscale x 4 x i1> %m, i32 %evl)
  %bext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %b, <vscale x 4 x i1> %m, i32 %evl)
  %v = call <vscale x 4 x float> @llvm.vp.fma.nxv4f32(<vscale x 4 x float> %aext, <vscale x 4 x float> %bext, <vscale x 4 x float> %c, <vscale x 4 x i1> %m, i32 %evl)
  ret <vscale x 4 x float> %v
}

define <vscale x 4 x float> @vfmacc_vv_nxv4f32_unmasked(<vscale x 4 x half> %a, <vscale x 4 x half> %b, <vscale x 4 x float> %c, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv4f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v9
; ZVFH-NEXT:    vmv2r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv4f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT:    vmv1r.v v12, v9
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v14, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v12
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v14, v10
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %a, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %b, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 4 x float> @llvm.vp.fma.nxv4f32(<vscale x 4 x float> %aext, <vscale x 4 x float> %bext, <vscale x 4 x float> %c, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 4 x float> %v
}

define <vscale x 4 x float> @vfmacc_vf_nxv4f32(<vscale x 4 x half> %va, half %b, <vscale x 4 x float> %vc, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv4f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v10, fa0, v8
; ZVFH-NEXT:    vmv2r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv4f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT:    vmv1r.v v12, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v12
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v8, fa5, v10
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 4 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 4 x half> %elt.head, <vscale x 4 x half> poison, <vscale x 4 x i32> zeroinitializer
  %vaext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %va, <vscale x 4 x i1> %m, i32 %evl)
  %vbext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %vb, <vscale x 4 x i1> %m, i32 %evl)
  %v = call <vscale x 4 x float> @llvm.vp.fma.nxv4f32(<vscale x 4 x float> %vaext, <vscale x 4 x float> %vbext, <vscale x 4 x float> %vc, <vscale x 4 x i1> %m, i32 %evl)
  ret <vscale x 4 x float> %v
}

define <vscale x 4 x float> @vfmacc_vf_nxv4f32_unmasked(<vscale x 4 x half> %va, half %b, <vscale x 4 x float> %vc, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv4f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v10, fa0, v8
; ZVFH-NEXT:    vmv2r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv4f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; ZVFHMIN-NEXT:    vmv1r.v v12, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v12
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v8, fa5, v10
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 4 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 4 x half> %elt.head, <vscale x 4 x half> poison, <vscale x 4 x i32> zeroinitializer
  %vaext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %va, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 4 x float> @llvm.vp.fpext.nxv4f32.nxv4f16(<vscale x 4 x half> %vb, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 4 x float> @llvm.vp.fma.nxv4f32(<vscale x 4 x float> %vaext, <vscale x 4 x float> %vbext, <vscale x 4 x float> %vc, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 4 x float> %v
}

define <vscale x 8 x float> @vfmacc_vv_nxv8f32(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x float> %c, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv8f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v12, v8, v10
; ZVFH-NEXT:    vmv4r.v v8, v12
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv8f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT:    vmv2r.v v16, v10
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v20, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v16
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v20, v12
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %a, <vscale x 8 x i1> %m, i32 %evl)
  %bext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %b, <vscale x 8 x i1> %m, i32 %evl)
  %v = call <vscale x 8 x float> @llvm.vp.fma.nxv8f32(<vscale x 8 x float> %aext, <vscale x 8 x float> %bext, <vscale x 8 x float> %c, <vscale x 8 x i1> %m, i32 %evl)
  ret <vscale x 8 x float> %v
}

define <vscale x 8 x float> @vfmacc_vv_nxv8f32_unmasked(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x float> %c, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv8f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v12, v8, v10
; ZVFH-NEXT:    vmv4r.v v8, v12
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv8f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT:    vmv2r.v v16, v10
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v20, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v16
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v20, v12
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %a, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %b, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 8 x float> @llvm.vp.fma.nxv8f32(<vscale x 8 x float> %aext, <vscale x 8 x float> %bext, <vscale x 8 x float> %c, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 8 x float> %v
}

define <vscale x 8 x float> @vfmacc_vf_nxv8f32(<vscale x 8 x half> %va, half %b, <vscale x 8 x float> %vc, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv8f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v12, fa0, v8
; ZVFH-NEXT:    vmv4r.v v8, v12
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv8f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT:    vmv2r.v v16, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v16
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v8, fa5, v12
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 8 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 8 x half> %elt.head, <vscale x 8 x half> poison, <vscale x 8 x i32> zeroinitializer
  %vaext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %va, <vscale x 8 x i1> %m, i32 %evl)
  %vbext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %vb, <vscale x 8 x i1> %m, i32 %evl)
  %v = call <vscale x 8 x float> @llvm.vp.fma.nxv8f32(<vscale x 8 x float> %vaext, <vscale x 8 x float> %vbext, <vscale x 8 x float> %vc, <vscale x 8 x i1> %m, i32 %evl)
  ret <vscale x 8 x float> %v
}

define <vscale x 8 x float> @vfmacc_vf_nxv8f32_unmasked(<vscale x 8 x half> %va, half %b, <vscale x 8 x float> %vc, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv8f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v12, fa0, v8
; ZVFH-NEXT:    vmv4r.v v8, v12
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv8f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; ZVFHMIN-NEXT:    vmv2r.v v16, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v16
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v8, fa5, v12
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 8 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 8 x half> %elt.head, <vscale x 8 x half> poison, <vscale x 8 x i32> zeroinitializer
  %vaext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %va, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 8 x float> @llvm.vp.fpext.nxv8f32.nxv8f16(<vscale x 8 x half> %vb, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 8 x float> @llvm.vp.fma.nxv8f32(<vscale x 8 x float> %vaext, <vscale x 8 x float> %vbext, <vscale x 8 x float> %vc, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 8 x float> %v
}

define <vscale x 16 x float> @vfmacc_vv_nxv16f32(<vscale x 16 x half> %a, <vscale x 16 x half> %b, <vscale x 16 x float> %c, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv16f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v16, v8, v12
; ZVFH-NEXT:    vmv8r.v v8, v16
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv16f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT:    vmv4r.v v24, v12
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v0, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v24
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v0, v16
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %a, <vscale x 16 x i1> %m, i32 %evl)
  %bext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %b, <vscale x 16 x i1> %m, i32 %evl)
  %v = call <vscale x 16 x float> @llvm.vp.fma.nxv16f32(<vscale x 16 x float> %aext, <vscale x 16 x float> %bext, <vscale x 16 x float> %c, <vscale x 16 x i1> %m, i32 %evl)
  ret <vscale x 16 x float> %v
}

define <vscale x 16 x float> @vfmacc_vv_nxv16f32_unmasked(<vscale x 16 x half> %a, <vscale x 16 x half> %b, <vscale x 16 x float> %c, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vv_nxv16f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v16, v8, v12
; ZVFH-NEXT:    vmv8r.v v8, v16
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vv_nxv16f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT:    vmv4r.v v24, v12
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v0, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v24
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v8, v0, v16
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %a, <vscale x 16 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %b, <vscale x 16 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 16 x float> @llvm.vp.fma.nxv16f32(<vscale x 16 x float> %aext, <vscale x 16 x float> %bext, <vscale x 16 x float> %c, <vscale x 16 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 16 x float> %v
}

define <vscale x 16 x float> @vfmacc_vf_nxv16f32(<vscale x 16 x half> %va, half %b, <vscale x 16 x float> %vc, <vscale x 16 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv16f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v16, fa0, v8
; ZVFH-NEXT:    vmv8r.v v8, v16
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv16f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT:    vmv4r.v v24, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v24
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v8, fa5, v16
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 16 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 16 x half> %elt.head, <vscale x 16 x half> poison, <vscale x 16 x i32> zeroinitializer
  %vaext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %va, <vscale x 16 x i1> %m, i32 %evl)
  %vbext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %vb, <vscale x 16 x i1> %m, i32 %evl)
  %v = call <vscale x 16 x float> @llvm.vp.fma.nxv16f32(<vscale x 16 x float> %vaext, <vscale x 16 x float> %vbext, <vscale x 16 x float> %vc, <vscale x 16 x i1> %m, i32 %evl)
  ret <vscale x 16 x float> %v
}

define <vscale x 16 x float> @vfmacc_vf_nxv16f32_unmasked(<vscale x 16 x half> %va, half %b, <vscale x 16 x float> %vc, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_vf_nxv16f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFH-NEXT:    vfwmacc.vf v16, fa0, v8
; ZVFH-NEXT:    vmv8r.v v8, v16
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_vf_nxv16f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, m4, ta, ma
; ZVFHMIN-NEXT:    vmv4r.v v24, v8
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v8, v24
; ZVFHMIN-NEXT:    fcvt.s.h fa5, fa0
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, m8, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vf v8, fa5, v16
; ZVFHMIN-NEXT:    ret
  %elt.head = insertelement <vscale x 16 x half> poison, half %b, i32 0
  %vb = shufflevector <vscale x 16 x half> %elt.head, <vscale x 16 x half> poison, <vscale x 16 x i32> zeroinitializer
  %vaext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %va, <vscale x 16 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 16 x float> @llvm.vp.fpext.nxv16f32.nxv16f16(<vscale x 16 x half> %vb, <vscale x 16 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 16 x float> @llvm.vp.fma.nxv16f32(<vscale x 16 x float> %vaext, <vscale x 16 x float> %vbext, <vscale x 16 x float> %vc, <vscale x 16 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 16 x float> %v
}

define <vscale x 1 x double> @vfmacc_vv_nxv1f64(<vscale x 1 x float> %a, <vscale x 1 x float> %b, <vscale x 1 x double> %c, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv1f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v8, v9
; CHECK-NEXT:    vmv1r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %a, <vscale x 1 x i1> %m, i32 %evl)
  %bext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %b, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x double> @llvm.vp.fma.nxv1f64(<vscale x 1 x double> %aext, <vscale x 1 x double> %bext, <vscale x 1 x double> %c, <vscale x 1 x i1> %m, i32 %evl)
  ret <vscale x 1 x double> %v
}

define <vscale x 1 x double> @vfmacc_vv_nxv1f64_unmasked(<vscale x 1 x float> %a, <vscale x 1 x float> %b, <vscale x 1 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv1f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v8, v9
; CHECK-NEXT:    vmv1r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %a, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %b, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x double> @llvm.vp.fma.nxv1f64(<vscale x 1 x double> %aext, <vscale x 1 x double> %bext, <vscale x 1 x double> %c, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 1 x double> %v
}

define <vscale x 1 x double> @vfmacc_vf_nxv1f64(<vscale x 1 x float> %va, float %b, <vscale x 1 x double> %vc, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv1f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vfwmacc.vf v9, fa0, v8
; CHECK-NEXT:    vmv1r.v v8, v9
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 1 x float> %elt.head, <vscale x 1 x float> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x i1> %m, i32 %evl)
  %vbext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %vb, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x double> @llvm.vp.fma.nxv1f64(<vscale x 1 x double> %vaext, <vscale x 1 x double> %vbext, <vscale x 1 x double> %vc, <vscale x 1 x i1> %m, i32 %evl)
  ret <vscale x 1 x double> %v
}

define <vscale x 1 x double> @vfmacc_vf_nxv1f64_unmasked(<vscale x 1 x float> %va, float %b, <vscale x 1 x double> %vc, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv1f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vfwmacc.vf v9, fa0, v8
; CHECK-NEXT:    vmv1r.v v8, v9
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 1 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 1 x float> %elt.head, <vscale x 1 x float> poison, <vscale x 1 x i32> zeroinitializer
  %vaext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %va, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f32(<vscale x 1 x float> %vb, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x double> @llvm.vp.fma.nxv1f64(<vscale x 1 x double> %vaext, <vscale x 1 x double> %vbext, <vscale x 1 x double> %vc, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 1 x double> %v
}

define <vscale x 2 x double> @vfmacc_vv_nxv2f64(<vscale x 2 x float> %a, <vscale x 2 x float> %b, <vscale x 2 x double> %c, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv2f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m1, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v8, v9
; CHECK-NEXT:    vmv2r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %a, <vscale x 2 x i1> %m, i32 %evl)
  %bext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %b, <vscale x 2 x i1> %m, i32 %evl)
  %v = call <vscale x 2 x double> @llvm.vp.fma.nxv2f64(<vscale x 2 x double> %aext, <vscale x 2 x double> %bext, <vscale x 2 x double> %c, <vscale x 2 x i1> %m, i32 %evl)
  ret <vscale x 2 x double> %v
}

define <vscale x 2 x double> @vfmacc_vv_nxv2f64_unmasked(<vscale x 2 x float> %a, <vscale x 2 x float> %b, <vscale x 2 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv2f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m1, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v8, v9
; CHECK-NEXT:    vmv2r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %a, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %b, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 2 x double> @llvm.vp.fma.nxv2f64(<vscale x 2 x double> %aext, <vscale x 2 x double> %bext, <vscale x 2 x double> %c, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 2 x double> %v
}

define <vscale x 2 x double> @vfmacc_vf_nxv2f64(<vscale x 2 x float> %va, float %b, <vscale x 2 x double> %vc, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv2f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m1, ta, ma
; CHECK-NEXT:    vfwmacc.vf v10, fa0, v8
; CHECK-NEXT:    vmv2r.v v8, v10
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 2 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 2 x float> %elt.head, <vscale x 2 x float> poison, <vscale x 2 x i32> zeroinitializer
  %vaext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x i1> %m, i32 %evl)
  %vbext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %vb, <vscale x 2 x i1> %m, i32 %evl)
  %v = call <vscale x 2 x double> @llvm.vp.fma.nxv2f64(<vscale x 2 x double> %vaext, <vscale x 2 x double> %vbext, <vscale x 2 x double> %vc, <vscale x 2 x i1> %m, i32 %evl)
  ret <vscale x 2 x double> %v
}

define <vscale x 2 x double> @vfmacc_vf_nxv2f64_unmasked(<vscale x 2 x float> %va, float %b, <vscale x 2 x double> %vc, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv2f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m1, ta, ma
; CHECK-NEXT:    vfwmacc.vf v10, fa0, v8
; CHECK-NEXT:    vmv2r.v v8, v10
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 2 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 2 x float> %elt.head, <vscale x 2 x float> poison, <vscale x 2 x i32> zeroinitializer
  %vaext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %va, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f32(<vscale x 2 x float> %vb, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 2 x double> @llvm.vp.fma.nxv2f64(<vscale x 2 x double> %vaext, <vscale x 2 x double> %vbext, <vscale x 2 x double> %vc, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 2 x double> %v
}

define <vscale x 4 x double> @vfmacc_vv_nxv4f64(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x double> %c, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv4f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v12, v8, v10
; CHECK-NEXT:    vmv4r.v v8, v12
; CHECK-NEXT:    ret
  %aext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %a, <vscale x 4 x i1> %m, i32 %evl)
  %bext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %b, <vscale x 4 x i1> %m, i32 %evl)
  %v = call <vscale x 4 x double> @llvm.vp.fma.nxv4f64(<vscale x 4 x double> %aext, <vscale x 4 x double> %bext, <vscale x 4 x double> %c, <vscale x 4 x i1> %m, i32 %evl)
  ret <vscale x 4 x double> %v
}

define <vscale x 4 x double> @vfmacc_vv_nxv4f64_unmasked(<vscale x 4 x float> %a, <vscale x 4 x float> %b, <vscale x 4 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv4f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v12, v8, v10
; CHECK-NEXT:    vmv4r.v v8, v12
; CHECK-NEXT:    ret
  %aext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %a, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %b, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 4 x double> @llvm.vp.fma.nxv4f64(<vscale x 4 x double> %aext, <vscale x 4 x double> %bext, <vscale x 4 x double> %c, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 4 x double> %v
}

define <vscale x 4 x double> @vfmacc_vf_nxv4f64(<vscale x 4 x float> %va, float %b, <vscale x 4 x double> %vc, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv4f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m2, ta, ma
; CHECK-NEXT:    vfwmacc.vf v12, fa0, v8
; CHECK-NEXT:    vmv4r.v v8, v12
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 4 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 4 x float> %elt.head, <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
  %vaext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x i1> %m, i32 %evl)
  %vbext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %vb, <vscale x 4 x i1> %m, i32 %evl)
  %v = call <vscale x 4 x double> @llvm.vp.fma.nxv4f64(<vscale x 4 x double> %vaext, <vscale x 4 x double> %vbext, <vscale x 4 x double> %vc, <vscale x 4 x i1> %m, i32 %evl)
  ret <vscale x 4 x double> %v
}

define <vscale x 4 x double> @vfmacc_vf_nxv4f64_unmasked(<vscale x 4 x float> %va, float %b, <vscale x 4 x double> %vc, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv4f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m2, ta, ma
; CHECK-NEXT:    vfwmacc.vf v12, fa0, v8
; CHECK-NEXT:    vmv4r.v v8, v12
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 4 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 4 x float> %elt.head, <vscale x 4 x float> poison, <vscale x 4 x i32> zeroinitializer
  %vaext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %va, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f32(<vscale x 4 x float> %vb, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 4 x double> @llvm.vp.fma.nxv4f64(<vscale x 4 x double> %vaext, <vscale x 4 x double> %vbext, <vscale x 4 x double> %vc, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 4 x double> %v
}

define <vscale x 8 x double> @vfmacc_vv_nxv8f64(<vscale x 8 x float> %a, <vscale x 8 x float> %b, <vscale x 8 x double> %c, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv8f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m4, ta, ma
; CHECK-NEXT:    vfwmacc.vv v16, v8, v12
; CHECK-NEXT:    vmv8r.v v8, v16
; CHECK-NEXT:    ret
  %aext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %a, <vscale x 8 x i1> %m, i32 %evl)
  %bext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %b, <vscale x 8 x i1> %m, i32 %evl)
  %v = call <vscale x 8 x double> @llvm.vp.fma.nxv8f64(<vscale x 8 x double> %aext, <vscale x 8 x double> %bext, <vscale x 8 x double> %c, <vscale x 8 x i1> %m, i32 %evl)
  ret <vscale x 8 x double> %v
}

define <vscale x 8 x double> @vfmacc_vv_nxv8f64_unmasked(<vscale x 8 x float> %a, <vscale x 8 x float> %b, <vscale x 8 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv8f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m4, ta, ma
; CHECK-NEXT:    vfwmacc.vv v16, v8, v12
; CHECK-NEXT:    vmv8r.v v8, v16
; CHECK-NEXT:    ret
  %aext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %a, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %b, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 8 x double> @llvm.vp.fma.nxv8f64(<vscale x 8 x double> %aext, <vscale x 8 x double> %bext, <vscale x 8 x double> %c, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 8 x double> %v
}

define <vscale x 8 x double> @vfmacc_vf_nxv8f64(<vscale x 8 x float> %va, float %b, <vscale x 8 x double> %vc, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv8f64:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m4, ta, ma
; CHECK-NEXT:    vfwmacc.vf v16, fa0, v8
; CHECK-NEXT:    vmv8r.v v8, v16
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 8 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 8 x float> %elt.head, <vscale x 8 x float> poison, <vscale x 8 x i32> zeroinitializer
  %vaext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x i1> %m, i32 %evl)
  %vbext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %vb, <vscale x 8 x i1> %m, i32 %evl)
  %v = call <vscale x 8 x double> @llvm.vp.fma.nxv8f64(<vscale x 8 x double> %vaext, <vscale x 8 x double> %vbext, <vscale x 8 x double> %vc, <vscale x 8 x i1> %m, i32 %evl)
  ret <vscale x 8 x double> %v
}

define <vscale x 8 x double> @vfmacc_vf_nxv8f64_unmasked(<vscale x 8 x float> %va, float %b, <vscale x 8 x double> %vc, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vf_nxv8f64_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e32, m4, ta, ma
; CHECK-NEXT:    vfwmacc.vf v16, fa0, v8
; CHECK-NEXT:    vmv8r.v v8, v16
; CHECK-NEXT:    ret
  %elt.head = insertelement <vscale x 8 x float> poison, float %b, i32 0
  %vb = shufflevector <vscale x 8 x float> %elt.head, <vscale x 8 x float> poison, <vscale x 8 x i32> zeroinitializer
  %vaext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %va, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %vbext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f32(<vscale x 8 x float> %vb, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 8 x double> @llvm.vp.fma.nxv8f64(<vscale x 8 x double> %vaext, <vscale x 8 x double> %vbext, <vscale x 8 x double> %vc, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 8 x double> %v
}

define <vscale x 1 x double> @vfmacc_vv_nxv1f64_nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x double> %c, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv1f64_nxv1f16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v11, v8
; CHECK-NEXT:    vfwcvt.f.f.v v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v11, v8
; CHECK-NEXT:    vmv1r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> %m, i32 %evl)
  %bext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f16(<vscale x 1 x half> %b, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x double> @llvm.vp.fma.nxv1f64(<vscale x 1 x double> %aext, <vscale x 1 x double> %bext, <vscale x 1 x double> %c, <vscale x 1 x i1> %m, i32 %evl)
  ret <vscale x 1 x double> %v
}

define <vscale x 1 x double> @vfmacc_vv_nxv1f64_nxv1f16_unmasked(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv1f64_nxv1f16_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v11, v8
; CHECK-NEXT:    vfwcvt.f.f.v v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v11, v8
; CHECK-NEXT:    vmv1r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 1 x double> @llvm.vp.fpext.nxv1f64.nxv1f16(<vscale x 1 x half> %b, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x double> @llvm.vp.fma.nxv1f64(<vscale x 1 x double> %aext, <vscale x 1 x double> %bext, <vscale x 1 x double> %c, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 1 x double> %v
}

define <vscale x 2 x double> @vfmacc_vv_nxv2f64_nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x half> %b, <vscale x 2 x double> %c, <vscale x 2 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv2f64_nxv2f16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v12, v8
; CHECK-NEXT:    vfwcvt.f.f.v v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v12, v8
; CHECK-NEXT:    vmv2r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> %m, i32 %evl)
  %bext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f16(<vscale x 2 x half> %b, <vscale x 2 x i1> %m, i32 %evl)
  %v = call <vscale x 2 x double> @llvm.vp.fma.nxv2f64(<vscale x 2 x double> %aext, <vscale x 2 x double> %bext, <vscale x 2 x double> %c, <vscale x 2 x i1> %m, i32 %evl)
  ret <vscale x 2 x double> %v
}

define <vscale x 2 x double> @vfmacc_vv_nxv2f64_nxv2f16_unmasked(<vscale x 2 x half> %a, <vscale x 2 x half> %b, <vscale x 2 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv2f64_nxv2f16_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, mf2, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v12, v8
; CHECK-NEXT:    vfwcvt.f.f.v v8, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m1, ta, ma
; CHECK-NEXT:    vfwmacc.vv v10, v12, v8
; CHECK-NEXT:    vmv2r.v v8, v10
; CHECK-NEXT:    ret
  %aext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f16(<vscale x 2 x half> %a, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 2 x double> @llvm.vp.fpext.nxv2f64.nxv2f16(<vscale x 2 x half> %b, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 2 x double> @llvm.vp.fma.nxv2f64(<vscale x 2 x double> %aext, <vscale x 2 x double> %bext, <vscale x 2 x double> %c, <vscale x 2 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 2 x double> %v
}

define <vscale x 4 x double> @vfmacc_vv_nxv4f64_nxv4f16(<vscale x 4 x half> %a, <vscale x 4 x half> %b, <vscale x 4 x double> %c, <vscale x 4 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv4f64_nxv4f16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v10, v8
; CHECK-NEXT:    vfwcvt.f.f.v v16, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v12, v10, v16
; CHECK-NEXT:    vmv4r.v v8, v12
; CHECK-NEXT:    ret
  %aext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f16(<vscale x 4 x half> %a, <vscale x 4 x i1> %m, i32 %evl)
  %bext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f16(<vscale x 4 x half> %b, <vscale x 4 x i1> %m, i32 %evl)
  %v = call <vscale x 4 x double> @llvm.vp.fma.nxv4f64(<vscale x 4 x double> %aext, <vscale x 4 x double> %bext, <vscale x 4 x double> %c, <vscale x 4 x i1> %m, i32 %evl)
  ret <vscale x 4 x double> %v
}

define <vscale x 4 x double> @vfmacc_vv_nxv4f64_nxv4f16_unmasked(<vscale x 4 x half> %a, <vscale x 4 x half> %b, <vscale x 4 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv4f64_nxv4f16_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, m1, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v10, v8
; CHECK-NEXT:    vfwcvt.f.f.v v16, v9
; CHECK-NEXT:    vsetvli zero, zero, e32, m2, ta, ma
; CHECK-NEXT:    vfwmacc.vv v12, v10, v16
; CHECK-NEXT:    vmv4r.v v8, v12
; CHECK-NEXT:    ret
  %aext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f16(<vscale x 4 x half> %a, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 4 x double> @llvm.vp.fpext.nxv4f64.nxv4f16(<vscale x 4 x half> %b, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 4 x double> @llvm.vp.fma.nxv4f64(<vscale x 4 x double> %aext, <vscale x 4 x double> %bext, <vscale x 4 x double> %c, <vscale x 4 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 4 x double> %v
}

define <vscale x 8 x double> @vfmacc_vv_nxv8f64_nxv8f16(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x double> %c, <vscale x 8 x i1> %m, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv8f64_nxv8f16:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v12, v8
; CHECK-NEXT:    vfwcvt.f.f.v v24, v10
; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; CHECK-NEXT:    vfwmacc.vv v16, v12, v24
; CHECK-NEXT:    vmv8r.v v8, v16
; CHECK-NEXT:    ret
  %aext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f16(<vscale x 8 x half> %a, <vscale x 8 x i1> %m, i32 %evl)
  %bext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f16(<vscale x 8 x half> %b, <vscale x 8 x i1> %m, i32 %evl)
  %v = call <vscale x 8 x double> @llvm.vp.fma.nxv8f64(<vscale x 8 x double> %aext, <vscale x 8 x double> %bext, <vscale x 8 x double> %c, <vscale x 8 x i1> %m, i32 %evl)
  ret <vscale x 8 x double> %v
}

define <vscale x 8 x double> @vfmacc_vv_nxv8f64_nxv8f16_unmasked(<vscale x 8 x half> %a, <vscale x 8 x half> %b, <vscale x 8 x double> %c, i32 zeroext %evl) {
; CHECK-LABEL: vfmacc_vv_nxv8f64_nxv8f16_unmasked:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vsetvli a0, zero, e16, m2, ta, ma
; CHECK-NEXT:    vfwcvt.f.f.v v12, v8
; CHECK-NEXT:    vfwcvt.f.f.v v24, v10
; CHECK-NEXT:    vsetvli zero, zero, e32, m4, ta, ma
; CHECK-NEXT:    vfwmacc.vv v16, v12, v24
; CHECK-NEXT:    vmv8r.v v8, v16
; CHECK-NEXT:    ret
  %aext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f16(<vscale x 8 x half> %a, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %bext = call <vscale x 8 x double> @llvm.vp.fpext.nxv8f64.nxv8f16(<vscale x 8 x half> %b, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 8 x double> @llvm.vp.fma.nxv8f64(<vscale x 8 x double> %aext, <vscale x 8 x double> %bext, <vscale x 8 x double> %c, <vscale x 8 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 8 x double> %v
}

define <vscale x 1 x float> @vfmacc_squared_nxv1f32(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x float> %c, <vscale x 1 x i1> %m, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_squared_nxv1f32:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v8
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_squared_nxv1f32:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v9, v9, v10
; ZVFHMIN-NEXT:    vmv1r.v v8, v9
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> %m, i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %aext, <vscale x 1 x float> %aext, <vscale x 1 x float> %c, <vscale x 1 x i1> %m, i32 %evl)
  ret <vscale x 1 x float> %v
}

define <vscale x 1 x float> @vfmacc_squared_nxv1f32_unmasked(<vscale x 1 x half> %a, <vscale x 1 x half> %b, <vscale x 1 x float> %c, i32 zeroext %evl) {
; ZVFH-LABEL: vfmacc_squared_nxv1f32_unmasked:
; ZVFH:       # %bb.0:
; ZVFH-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFH-NEXT:    vfwmacc.vv v10, v8, v8
; ZVFH-NEXT:    vmv1r.v v8, v10
; ZVFH-NEXT:    ret
;
; ZVFHMIN-LABEL: vfmacc_squared_nxv1f32_unmasked:
; ZVFHMIN:       # %bb.0:
; ZVFHMIN-NEXT:    vsetvli a0, zero, e16, mf4, ta, ma
; ZVFHMIN-NEXT:    vfwcvt.f.f.v v9, v8
; ZVFHMIN-NEXT:    vsetvli zero, zero, e32, mf2, ta, ma
; ZVFHMIN-NEXT:    vfmadd.vv v9, v9, v10
; ZVFHMIN-NEXT:    vmv1r.v v8, v9
; ZVFHMIN-NEXT:    ret
  %aext = call <vscale x 1 x float> @llvm.vp.fpext.nxv1f32.nxv1f16(<vscale x 1 x half> %a, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  %v = call <vscale x 1 x float> @llvm.vp.fma.nxv1f32(<vscale x 1 x float> %aext, <vscale x 1 x float> %aext, <vscale x 1 x float> %c, <vscale x 1 x i1> splat (i1 -1), i32 %evl)
  ret <vscale x 1 x float> %v
}