test/Transforms/InstCombine/X86/x86-avx512.ll

; NOTE: Assertions have been autogenerated by utils/update_test_checks.py
; RUN: opt < %s -passes=instcombine -mtriple=x86_64-unknown-unknown -S | FileCheck %s
target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128"

declare <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float>, <4 x float>, <4 x float>, i8, i32)

define <4 x float> @test_add_ss(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_add_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fadd float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x float> [[A]], float [[TMP3]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP4]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_add_ss_round(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_add_ss_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 8)
  ret <4 x float> %4
}

define <4 x float> @test_add_ss_mask(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_add_ss_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fadd float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP5]], float [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x float> [[A]], float [[TMP6]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP7]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_add_ss_mask_round(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_add_ss_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 8)
  ret <4 x float> %4
}

define float @test_add_ss_1(float %a, float %b) {
;
; CHECK-LABEL: @test_add_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> undef, float %a, i32 0
  %2 = insertelement <4 x float> %1, float 1.000000e+00, i32 1
  %3 = insertelement <4 x float> %2, float 2.000000e+00, i32 2
  %4 = insertelement <4 x float> %3, float 3.000000e+00, i32 3
  %5 = insertelement <4 x float> undef, float %b, i32 0
  %6 = insertelement <4 x float> %5, float 4.000000e+00, i32 1
  %7 = insertelement <4 x float> %6, float 5.000000e+00, i32 2
  %8 = insertelement <4 x float> %7, float 6.000000e+00, i32 3
  %9 = tail call <4 x float> @llvm.x86.avx512.mask.add.ss.round(<4 x float> %4, <4 x float> %8, <4 x float> undef, i8 -1, i32 8)
  %10 = extractelement <4 x float> %9, i32 1
  ret float %10
}

declare <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double>, <2 x double>, <2 x double>, i8, i32)

define <2 x double> @test_add_sd(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_add_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fadd double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x double> [[A]], double [[TMP3]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP4]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_add_sd_round(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_add_sd_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 8)
  ret <2 x double> %2
}

define <2 x double> @test_add_sd_mask(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_add_sd_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fadd double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP5]], double [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> [[A]], double [[TMP6]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP7]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_add_sd_mask_round(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_add_sd_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 8)
  ret <2 x double> %2
}

define double @test_add_sd_1(double %a, double %b) {
;
; CHECK-LABEL: @test_add_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> undef, double %a, i32 0
  %2 = insertelement <2 x double> %1, double 1.000000e+00, i32 1
  %3 = insertelement <2 x double> undef, double %b, i32 0
  %4 = insertelement <2 x double> %3, double 2.000000e+00, i32 1
  %5 = tail call <2 x double> @llvm.x86.avx512.mask.add.sd.round(<2 x double> %2, <2 x double> %4, <2 x double> undef, i8 -1, i32 8)
  %6 = extractelement <2 x double> %5, i32 1
  ret double %6
}

declare <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float>, <4 x float>, <4 x float>, i8, i32)

define <4 x float> @test_sub_ss(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_sub_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fsub float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x float> [[A]], float [[TMP3]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP4]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_sub_ss_round(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_sub_ss_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 8)
  ret <4 x float> %4
}

define <4 x float> @test_sub_ss_mask(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_sub_ss_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fsub float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP5]], float [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x float> [[A]], float [[TMP6]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP7]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_sub_ss_mask_round(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_sub_ss_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 8)
  ret <4 x float> %4
}

define float @test_sub_ss_1(float %a, float %b) {
;
; CHECK-LABEL: @test_sub_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> undef, float %a, i32 0
  %2 = insertelement <4 x float> %1, float 1.000000e+00, i32 1
  %3 = insertelement <4 x float> %2, float 2.000000e+00, i32 2
  %4 = insertelement <4 x float> %3, float 3.000000e+00, i32 3
  %5 = insertelement <4 x float> undef, float %b, i32 0
  %6 = insertelement <4 x float> %5, float 4.000000e+00, i32 1
  %7 = insertelement <4 x float> %6, float 5.000000e+00, i32 2
  %8 = insertelement <4 x float> %7, float 6.000000e+00, i32 3
  %9 = tail call <4 x float> @llvm.x86.avx512.mask.sub.ss.round(<4 x float> %4, <4 x float> %8, <4 x float> undef, i8 -1, i32 8)
  %10 = extractelement <4 x float> %9, i32 1
  ret float %10
}

declare <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double>, <2 x double>, <2 x double>, i8, i32)

define <2 x double> @test_sub_sd(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_sub_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fsub double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x double> [[A]], double [[TMP3]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP4]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_sub_sd_round(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_sub_sd_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 8)
  ret <2 x double> %2
}

define <2 x double> @test_sub_sd_mask(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_sub_sd_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fsub double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP5]], double [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> [[A]], double [[TMP6]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP7]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_sub_sd_mask_round(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_sub_sd_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 8)
  ret <2 x double> %2
}

define double @test_sub_sd_1(double %a, double %b) {
;
; CHECK-LABEL: @test_sub_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> undef, double %a, i32 0
  %2 = insertelement <2 x double> %1, double 1.000000e+00, i32 1
  %3 = insertelement <2 x double> undef, double %b, i32 0
  %4 = insertelement <2 x double> %3, double 2.000000e+00, i32 1
  %5 = tail call <2 x double> @llvm.x86.avx512.mask.sub.sd.round(<2 x double> %2, <2 x double> %4, <2 x double> undef, i8 -1, i32 8)
  %6 = extractelement <2 x double> %5, i32 1
  ret double %6
}

declare <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float>, <4 x float>, <4 x float>, i8, i32)

define <4 x float> @test_mul_ss(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_mul_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fmul float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x float> [[A]], float [[TMP3]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP4]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_mul_ss_round(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_mul_ss_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 8)
  ret <4 x float> %4
}

define <4 x float> @test_mul_ss_mask(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mul_ss_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fmul float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP5]], float [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x float> [[A]], float [[TMP6]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP7]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_mul_ss_mask_round(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mul_ss_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 8)
  ret <4 x float> %4
}

define float @test_mul_ss_1(float %a, float %b) {
;
; CHECK-LABEL: @test_mul_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> undef, float %a, i32 0
  %2 = insertelement <4 x float> %1, float 1.000000e+00, i32 1
  %3 = insertelement <4 x float> %2, float 2.000000e+00, i32 2
  %4 = insertelement <4 x float> %3, float 3.000000e+00, i32 3
  %5 = insertelement <4 x float> undef, float %b, i32 0
  %6 = insertelement <4 x float> %5, float 4.000000e+00, i32 1
  %7 = insertelement <4 x float> %6, float 5.000000e+00, i32 2
  %8 = insertelement <4 x float> %7, float 6.000000e+00, i32 3
  %9 = tail call <4 x float> @llvm.x86.avx512.mask.mul.ss.round(<4 x float> %4, <4 x float> %8, <4 x float> undef, i8 -1, i32 8)
  %10 = extractelement <4 x float> %9, i32 1
  ret float %10
}

declare <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double>, <2 x double>, <2 x double>, i8, i32)

define <2 x double> @test_mul_sd(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_mul_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fmul double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x double> [[A]], double [[TMP3]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP4]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_mul_sd_round(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_mul_sd_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 8)
  ret <2 x double> %2
}

define <2 x double> @test_mul_sd_mask(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mul_sd_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fmul double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP5]], double [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> [[A]], double [[TMP6]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP7]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_mul_sd_mask_round(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mul_sd_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 8)
  ret <2 x double> %2
}

define double @test_mul_sd_1(double %a, double %b) {
;
; CHECK-LABEL: @test_mul_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> undef, double %a, i32 0
  %2 = insertelement <2 x double> %1, double 1.000000e+00, i32 1
  %3 = insertelement <2 x double> undef, double %b, i32 0
  %4 = insertelement <2 x double> %3, double 2.000000e+00, i32 1
  %5 = tail call <2 x double> @llvm.x86.avx512.mask.mul.sd.round(<2 x double> %2, <2 x double> %4, <2 x double> undef, i8 -1, i32 8)
  %6 = extractelement <2 x double> %5, i32 1
  ret double %6
}

declare <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float>, <4 x float>, <4 x float>, i8, i32)

define <4 x float> @test_div_ss(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_div_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fdiv float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <4 x float> [[A]], float [[TMP3]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP4]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_div_ss_round(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_div_ss_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 8)
  ret <4 x float> %4
}

define <4 x float> @test_div_ss_mask(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_div_ss_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fdiv float [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP5]], float [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x float> [[A]], float [[TMP6]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP7]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_div_ss_mask_round(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_div_ss_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 8)
  ret <4 x float> %4
}

define float @test_div_ss_1(float %a, float %b) {
;
; CHECK-LABEL: @test_div_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> undef, float %a, i32 0
  %2 = insertelement <4 x float> %1, float 1.000000e+00, i32 1
  %3 = insertelement <4 x float> %2, float 2.000000e+00, i32 2
  %4 = insertelement <4 x float> %3, float 3.000000e+00, i32 3
  %5 = insertelement <4 x float> undef, float %b, i32 0
  %6 = insertelement <4 x float> %5, float 4.000000e+00, i32 1
  %7 = insertelement <4 x float> %6, float 5.000000e+00, i32 2
  %8 = insertelement <4 x float> %7, float 6.000000e+00, i32 3
  %9 = tail call <4 x float> @llvm.x86.avx512.mask.div.ss.round(<4 x float> %4, <4 x float> %8, <4 x float> undef, i8 -1, i32 8)
  %10 = extractelement <4 x float> %9, i32 1
  ret float %10
}

declare <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double>, <2 x double>, <2 x double>, i8, i32)

define <2 x double> @test_div_sd(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_div_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fdiv double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = insertelement <2 x double> [[A]], double [[TMP3]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP4]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_div_sd_round(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_div_sd_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> undef, i8 -1, i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 8)
  ret <2 x double> %2
}

define <2 x double> @test_div_sd_mask(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_div_sd_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = fdiv double [[TMP1]], [[TMP2]]
; CHECK-NEXT:    [[TMP4:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP4]], 0
; CHECK-NEXT:    [[TMP5:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP5]], double [[TMP3]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> [[A]], double [[TMP6]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP7]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_div_sd_mask_round(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_div_sd_mask_round(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], i8 [[MASK:%.*]], i32 8)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 8)
  ret <2 x double> %2
}

define double @test_div_sd_1(double %a, double %b) {
;
; CHECK-LABEL: @test_div_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> undef, double %a, i32 0
  %2 = insertelement <2 x double> %1, double 1.000000e+00, i32 1
  %3 = insertelement <2 x double> undef, double %b, i32 0
  %4 = insertelement <2 x double> %3, double 2.000000e+00, i32 1
  %5 = tail call <2 x double> @llvm.x86.avx512.mask.div.sd.round(<2 x double> %2, <2 x double> %4, <2 x double> undef, i8 -1, i32 8)
  %6 = extractelement <2 x double> %5, i32 1
  ret double %6
}

declare <4 x float> @llvm.x86.avx512.mask.max.ss.round(<4 x float>, <4 x float>, <4 x float>, i8, i32)

define <4 x float> @test_max_ss(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_max_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.max.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> undef, i8 -1, i32 4)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.max.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_max_ss_mask(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_max_ss_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.max.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], i8 [[MASK:%.*]], i32 4)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.max.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 4)
  ret <4 x float> %4
}

define float @test_max_ss_1(float %a, float %b) {
;
; CHECK-LABEL: @test_max_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> undef, float %a, i32 0
  %2 = insertelement <4 x float> %1, float 1.000000e+00, i32 1
  %3 = insertelement <4 x float> %2, float 2.000000e+00, i32 2
  %4 = insertelement <4 x float> %3, float 3.000000e+00, i32 3
  %5 = insertelement <4 x float> undef, float %b, i32 0
  %6 = insertelement <4 x float> %5, float 4.000000e+00, i32 1
  %7 = insertelement <4 x float> %6, float 5.000000e+00, i32 2
  %8 = insertelement <4 x float> %7, float 6.000000e+00, i32 3
  %9 = tail call <4 x float> @llvm.x86.avx512.mask.max.ss.round(<4 x float> %4, <4 x float> %8, <4 x float> undef, i8 -1, i32 8)
  %10 = extractelement <4 x float> %9, i32 1
  ret float %10
}

declare <2 x double> @llvm.x86.avx512.mask.max.sd.round(<2 x double>, <2 x double>, <2 x double>, i8, i32)

define <2 x double> @test_max_sd(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_max_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.max.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> undef, i8 -1, i32 4)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.max.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_max_sd_mask(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_max_sd_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.max.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], i8 [[MASK:%.*]], i32 4)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.max.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 4)
  ret <2 x double> %2
}

define double @test_max_sd_1(double %a, double %b) {
;
; CHECK-LABEL: @test_max_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> undef, double %a, i32 0
  %2 = insertelement <2 x double> %1, double 1.000000e+00, i32 1
  %3 = insertelement <2 x double> undef, double %b, i32 0
  %4 = insertelement <2 x double> %3, double 2.000000e+00, i32 1
  %5 = tail call <2 x double> @llvm.x86.avx512.mask.max.sd.round(<2 x double> %2, <2 x double> %4, <2 x double> undef, i8 -1, i32 8)
  %6 = extractelement <2 x double> %5, i32 1
  ret double %6
}

declare <4 x float> @llvm.x86.avx512.mask.min.ss.round(<4 x float>, <4 x float>, <4 x float>, i8, i32)

define <4 x float> @test_min_ss(<4 x float> %a, <4 x float> %b) {
;
; CHECK-LABEL: @test_min_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.min.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> undef, i8 -1, i32 4)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.min.ss.round(<4 x float> %a, <4 x float> %3, <4 x float> undef, i8 -1, i32 4)
  ret <4 x float> %4
}

define <4 x float> @test_min_ss_mask(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_min_ss_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <4 x float> @llvm.x86.avx512.mask.min.ss.round(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], <4 x float> [[C:%.*]], i8 [[MASK:%.*]], i32 4)
; CHECK-NEXT:    ret <4 x float> [[TMP1]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = tail call <4 x float> @llvm.x86.avx512.mask.min.ss.round(<4 x float> %a, <4 x float> %b, <4 x float> %3, i8 %mask, i32 4)
  ret <4 x float> %4
}

define float @test_min_ss_1(float %a, float %b) {
;
; CHECK-LABEL: @test_min_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> undef, float %a, i32 0
  %2 = insertelement <4 x float> %1, float 1.000000e+00, i32 1
  %3 = insertelement <4 x float> %2, float 2.000000e+00, i32 2
  %4 = insertelement <4 x float> %3, float 3.000000e+00, i32 3
  %5 = insertelement <4 x float> undef, float %b, i32 0
  %6 = insertelement <4 x float> %5, float 4.000000e+00, i32 1
  %7 = insertelement <4 x float> %6, float 5.000000e+00, i32 2
  %8 = insertelement <4 x float> %7, float 6.000000e+00, i32 3
  %9 = tail call <4 x float> @llvm.x86.avx512.mask.min.ss.round(<4 x float> %4, <4 x float> %8, <4 x float> undef, i8 -1, i32 8)
  %10 = extractelement <4 x float> %9, i32 1
  ret float %10
}

declare <2 x double> @llvm.x86.avx512.mask.min.sd.round(<2 x double>, <2 x double>, <2 x double>, i8, i32)

define <2 x double> @test_min_sd(<2 x double> %a, <2 x double> %b) {
;
; CHECK-LABEL: @test_min_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.min.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> undef, i8 -1, i32 4)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.min.sd.round(<2 x double> %a, <2 x double> %1, <2 x double> undef, i8 -1, i32 4)
  ret <2 x double> %2
}

define <2 x double> @test_min_sd_mask(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_min_sd_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call <2 x double> @llvm.x86.avx512.mask.min.sd.round(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], <2 x double> [[C:%.*]], i8 [[MASK:%.*]], i32 4)
; CHECK-NEXT:    ret <2 x double> [[TMP1]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = tail call <2 x double> @llvm.x86.avx512.mask.min.sd.round(<2 x double> %a, <2 x double> %b, <2 x double> %1, i8 %mask, i32 4)
  ret <2 x double> %2
}

define double @test_min_sd_1(double %a, double %b) {
;
; CHECK-LABEL: @test_min_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> undef, double %a, i32 0
  %2 = insertelement <2 x double> %1, double 1.000000e+00, i32 1
  %3 = insertelement <2 x double> undef, double %b, i32 0
  %4 = insertelement <2 x double> %3, double 2.000000e+00, i32 1
  %5 = tail call <2 x double> @llvm.x86.avx512.mask.min.sd.round(<2 x double> %2, <2 x double> %4, <2 x double> undef, i8 -1, i32 8)
  %6 = extractelement <2 x double> %5, i32 1
  ret double %6
}

declare i8 @llvm.x86.avx512.mask.cmp.ss(<4 x float>, <4 x float>, i32, i8, i32)

define i8 @test_cmp_ss(<4 x float> %a, <4 x float> %b, i8 %mask) {
;
; CHECK-LABEL: @test_cmp_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call i8 @llvm.x86.avx512.mask.cmp.ss(<4 x float> [[A:%.*]], <4 x float> [[B:%.*]], i32 3, i8 [[MASK:%.*]], i32 4)
; CHECK-NEXT:    ret i8 [[TMP1]]
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = insertelement <4 x float> %b, float 4.000000e+00, i32 1
  %5 = insertelement <4 x float> %4, float 5.000000e+00, i32 2
  %6 = insertelement <4 x float> %5, float 6.000000e+00, i32 3
  %7 = tail call i8 @llvm.x86.avx512.mask.cmp.ss(<4 x float> %3, <4 x float> %6, i32 3, i8 %mask, i32 4)
  ret i8 %7
}

declare i8 @llvm.x86.avx512.mask.cmp.sd(<2 x double>, <2 x double>, i32, i8, i32)

define i8 @test_cmp_sd(<2 x double> %a, <2 x double> %b, i8 %mask) {
;
; CHECK-LABEL: @test_cmp_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = tail call i8 @llvm.x86.avx512.mask.cmp.sd(<2 x double> [[A:%.*]], <2 x double> [[B:%.*]], i32 3, i8 [[MASK:%.*]], i32 4)
; CHECK-NEXT:    ret i8 [[TMP1]]
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = insertelement <2 x double> %b, double 2.000000e+00, i32 1
  %3 = tail call i8 @llvm.x86.avx512.mask.cmp.sd(<2 x double> %1, <2 x double> %2, i32 3, i8 %mask, i32 4)
  ret i8 %3
}

define i64 @test(float %f, double %d) {
;
; CHECK-LABEL: @test(
; CHECK-NEXT:    [[V03:%.*]] = insertelement <4 x float> poison, float [[F:%.*]], i64 0
; CHECK-NEXT:    [[T0:%.*]] = tail call i32 @llvm.x86.avx512.vcvtss2si32(<4 x float> [[V03]], i32 4)
; CHECK-NEXT:    [[V13:%.*]] = insertelement <4 x float> poison, float [[F]], i64 0
; CHECK-NEXT:    [[T1:%.*]] = tail call i64 @llvm.x86.avx512.vcvtss2si64(<4 x float> [[V13]], i32 4)
; CHECK-NEXT:    [[V23:%.*]] = insertelement <4 x float> poison, float [[F]], i64 0
; CHECK-NEXT:    [[T2:%.*]] = tail call i32 @llvm.x86.avx512.cvttss2si(<4 x float> [[V23]], i32 4)
; CHECK-NEXT:    [[V33:%.*]] = insertelement <4 x float> poison, float [[F]], i64 0
; CHECK-NEXT:    [[T3:%.*]] = tail call i64 @llvm.x86.avx512.cvttss2si64(<4 x float> [[V33]], i32 4)
; CHECK-NEXT:    [[V41:%.*]] = insertelement <2 x double> poison, double [[D:%.*]], i64 0
; CHECK-NEXT:    [[T4:%.*]] = tail call i32 @llvm.x86.avx512.vcvtsd2si32(<2 x double> [[V41]], i32 4)
; CHECK-NEXT:    [[V51:%.*]] = insertelement <2 x double> poison, double [[D]], i64 0
; CHECK-NEXT:    [[T5:%.*]] = tail call i64 @llvm.x86.avx512.vcvtsd2si64(<2 x double> [[V51]], i32 4)
; CHECK-NEXT:    [[V61:%.*]] = insertelement <2 x double> poison, double [[D]], i64 0
; CHECK-NEXT:    [[T6:%.*]] = tail call i32 @llvm.x86.avx512.cvttsd2si(<2 x double> [[V61]], i32 4)
; CHECK-NEXT:    [[V71:%.*]] = insertelement <2 x double> poison, double [[D]], i64 0
; CHECK-NEXT:    [[T7:%.*]] = tail call i64 @llvm.x86.avx512.cvttsd2si64(<2 x double> [[V71]], i32 4)
; CHECK-NEXT:    [[T8:%.*]] = add i32 [[T0]], [[T2]]
; CHECK-NEXT:    [[T9:%.*]] = add i32 [[T4]], [[T6]]
; CHECK-NEXT:    [[T10:%.*]] = add i32 [[T8]], [[T9]]
; CHECK-NEXT:    [[T11:%.*]] = sext i32 [[T10]] to i64
; CHECK-NEXT:    [[T12:%.*]] = add i64 [[T1]], [[T3]]
; CHECK-NEXT:    [[T13:%.*]] = add i64 [[T5]], [[T7]]
; CHECK-NEXT:    [[T14:%.*]] = add i64 [[T12]], [[T13]]
; CHECK-NEXT:    [[T15:%.*]] = add i64 [[T14]], [[T11]]
; CHECK-NEXT:    ret i64 [[T15]]
;
  %v00 = insertelement <4 x float> undef, float %f, i32 0
  %v01 = insertelement <4 x float> %v00, float 0.000000e+00, i32 1
  %v02 = insertelement <4 x float> %v01, float 0.000000e+00, i32 2
  %v03 = insertelement <4 x float> %v02, float 0.000000e+00, i32 3
  %t0 = tail call i32 @llvm.x86.avx512.vcvtss2si32(<4 x float> %v03, i32 4)
  %v10 = insertelement <4 x float> undef, float %f, i32 0
  %v11 = insertelement <4 x float> %v10, float 0.000000e+00, i32 1
  %v12 = insertelement <4 x float> %v11, float 0.000000e+00, i32 2
  %v13 = insertelement <4 x float> %v12, float 0.000000e+00, i32 3
  %t1 = tail call i64 @llvm.x86.avx512.vcvtss2si64(<4 x float> %v13, i32 4)
  %v20 = insertelement <4 x float> undef, float %f, i32 0
  %v21 = insertelement <4 x float> %v20, float 0.000000e+00, i32 1
  %v22 = insertelement <4 x float> %v21, float 0.000000e+00, i32 2
  %v23 = insertelement <4 x float> %v22, float 0.000000e+00, i32 3
  %t2 = tail call i32 @llvm.x86.avx512.cvttss2si(<4 x float> %v23, i32 4)
  %v30 = insertelement <4 x float> undef, float %f, i32 0
  %v31 = insertelement <4 x float> %v30, float 0.000000e+00, i32 1
  %v32 = insertelement <4 x float> %v31, float 0.000000e+00, i32 2
  %v33 = insertelement <4 x float> %v32, float 0.000000e+00, i32 3
  %t3 = tail call i64 @llvm.x86.avx512.cvttss2si64(<4 x float> %v33, i32 4)
  %v40 = insertelement <2 x double> undef, double %d, i32 0
  %v41 = insertelement <2 x double> %v40, double 0.000000e+00, i32 1
  %t4 = tail call i32 @llvm.x86.avx512.vcvtsd2si32(<2 x double> %v41, i32 4)
  %v50 = insertelement <2 x double> undef, double %d, i32 0
  %v51 = insertelement <2 x double> %v50, double 0.000000e+00, i32 1
  %t5 = tail call i64 @llvm.x86.avx512.vcvtsd2si64(<2 x double> %v51, i32 4)
  %v60 = insertelement <2 x double> undef, double %d, i32 0
  %v61 = insertelement <2 x double> %v60, double 0.000000e+00, i32 1
  %t6 = tail call i32 @llvm.x86.avx512.cvttsd2si(<2 x double> %v61, i32 4)
  %v70 = insertelement <2 x double> undef, double %d, i32 0
  %v71 = insertelement <2 x double> %v70, double 0.000000e+00, i32 1
  %t7 = tail call i64 @llvm.x86.avx512.cvttsd2si64(<2 x double> %v71, i32 4)
  %t8 = add i32 %t0, %t2
  %t9 = add i32 %t4, %t6
  %t10 = add i32 %t8, %t9
  %t11 = sext i32 %t10 to i64
  %t12 = add i64 %t1, %t3
  %t13 = add i64 %t5, %t7
  %t14 = add i64 %t12, %t13
  %t15 = add i64 %t11, %t14
  ret i64 %t15
}

declare i32 @llvm.x86.avx512.vcvtss2si32(<4 x float>, i32)
declare i64 @llvm.x86.avx512.vcvtss2si64(<4 x float>, i32)
declare i32 @llvm.x86.avx512.cvttss2si(<4 x float>, i32)
declare i64 @llvm.x86.avx512.cvttss2si64(<4 x float>, i32)
declare i32 @llvm.x86.avx512.vcvtsd2si32(<2 x double>, i32)
declare i64 @llvm.x86.avx512.vcvtsd2si64(<2 x double>, i32)
declare i32 @llvm.x86.avx512.cvttsd2si(<2 x double>, i32)
declare i64 @llvm.x86.avx512.cvttsd2si64(<2 x double>, i32)

define i64 @test2(float %f, double %d) {
;
; CHECK-LABEL: @test2(
; CHECK-NEXT:    [[V03:%.*]] = insertelement <4 x float> poison, float [[F:%.*]], i64 0
; CHECK-NEXT:    [[T0:%.*]] = tail call i32 @llvm.x86.avx512.vcvtss2usi32(<4 x float> [[V03]], i32 4)
; CHECK-NEXT:    [[V13:%.*]] = insertelement <4 x float> poison, float [[F]], i64 0
; CHECK-NEXT:    [[T1:%.*]] = tail call i64 @llvm.x86.avx512.vcvtss2usi64(<4 x float> [[V13]], i32 4)
; CHECK-NEXT:    [[V23:%.*]] = insertelement <4 x float> poison, float [[F]], i64 0
; CHECK-NEXT:    [[T2:%.*]] = tail call i32 @llvm.x86.avx512.cvttss2usi(<4 x float> [[V23]], i32 4)
; CHECK-NEXT:    [[V33:%.*]] = insertelement <4 x float> poison, float [[F]], i64 0
; CHECK-NEXT:    [[T3:%.*]] = tail call i64 @llvm.x86.avx512.cvttss2usi64(<4 x float> [[V33]], i32 4)
; CHECK-NEXT:    [[V41:%.*]] = insertelement <2 x double> poison, double [[D:%.*]], i64 0
; CHECK-NEXT:    [[T4:%.*]] = tail call i32 @llvm.x86.avx512.vcvtsd2usi32(<2 x double> [[V41]], i32 4)
; CHECK-NEXT:    [[V51:%.*]] = insertelement <2 x double> poison, double [[D]], i64 0
; CHECK-NEXT:    [[T5:%.*]] = tail call i64 @llvm.x86.avx512.vcvtsd2usi64(<2 x double> [[V51]], i32 4)
; CHECK-NEXT:    [[V61:%.*]] = insertelement <2 x double> poison, double [[D]], i64 0
; CHECK-NEXT:    [[T6:%.*]] = tail call i32 @llvm.x86.avx512.cvttsd2usi(<2 x double> [[V61]], i32 4)
; CHECK-NEXT:    [[V71:%.*]] = insertelement <2 x double> poison, double [[D]], i64 0
; CHECK-NEXT:    [[T7:%.*]] = tail call i64 @llvm.x86.avx512.cvttsd2usi64(<2 x double> [[V71]], i32 4)
; CHECK-NEXT:    [[T8:%.*]] = add i32 [[T0]], [[T2]]
; CHECK-NEXT:    [[T9:%.*]] = add i32 [[T4]], [[T6]]
; CHECK-NEXT:    [[T10:%.*]] = add i32 [[T8]], [[T9]]
; CHECK-NEXT:    [[T11:%.*]] = sext i32 [[T10]] to i64
; CHECK-NEXT:    [[T12:%.*]] = add i64 [[T1]], [[T3]]
; CHECK-NEXT:    [[T13:%.*]] = add i64 [[T5]], [[T7]]
; CHECK-NEXT:    [[T14:%.*]] = add i64 [[T12]], [[T13]]
; CHECK-NEXT:    [[T15:%.*]] = add i64 [[T14]], [[T11]]
; CHECK-NEXT:    ret i64 [[T15]]
;
  %v00 = insertelement <4 x float> undef, float %f, i32 0
  %v01 = insertelement <4 x float> %v00, float 0.000000e+00, i32 1
  %v02 = insertelement <4 x float> %v01, float 0.000000e+00, i32 2
  %v03 = insertelement <4 x float> %v02, float 0.000000e+00, i32 3
  %t0 = tail call i32 @llvm.x86.avx512.vcvtss2usi32(<4 x float> %v03, i32 4)
  %v10 = insertelement <4 x float> undef, float %f, i32 0
  %v11 = insertelement <4 x float> %v10, float 0.000000e+00, i32 1
  %v12 = insertelement <4 x float> %v11, float 0.000000e+00, i32 2
  %v13 = insertelement <4 x float> %v12, float 0.000000e+00, i32 3
  %t1 = tail call i64 @llvm.x86.avx512.vcvtss2usi64(<4 x float> %v13, i32 4)
  %v20 = insertelement <4 x float> undef, float %f, i32 0
  %v21 = insertelement <4 x float> %v20, float 0.000000e+00, i32 1
  %v22 = insertelement <4 x float> %v21, float 0.000000e+00, i32 2
  %v23 = insertelement <4 x float> %v22, float 0.000000e+00, i32 3
  %t2 = tail call i32 @llvm.x86.avx512.cvttss2usi(<4 x float> %v23, i32 4)
  %v30 = insertelement <4 x float> undef, float %f, i32 0
  %v31 = insertelement <4 x float> %v30, float 0.000000e+00, i32 1
  %v32 = insertelement <4 x float> %v31, float 0.000000e+00, i32 2
  %v33 = insertelement <4 x float> %v32, float 0.000000e+00, i32 3
  %t3 = tail call i64 @llvm.x86.avx512.cvttss2usi64(<4 x float> %v33, i32 4)
  %v40 = insertelement <2 x double> undef, double %d, i32 0
  %v41 = insertelement <2 x double> %v40, double 0.000000e+00, i32 1
  %t4 = tail call i32 @llvm.x86.avx512.vcvtsd2usi32(<2 x double> %v41, i32 4)
  %v50 = insertelement <2 x double> undef, double %d, i32 0
  %v51 = insertelement <2 x double> %v50, double 0.000000e+00, i32 1
  %t5 = tail call i64 @llvm.x86.avx512.vcvtsd2usi64(<2 x double> %v51, i32 4)
  %v60 = insertelement <2 x double> undef, double %d, i32 0
  %v61 = insertelement <2 x double> %v60, double 0.000000e+00, i32 1
  %t6 = tail call i32 @llvm.x86.avx512.cvttsd2usi(<2 x double> %v61, i32 4)
  %v70 = insertelement <2 x double> undef, double %d, i32 0
  %v71 = insertelement <2 x double> %v70, double 0.000000e+00, i32 1
  %t7 = tail call i64 @llvm.x86.avx512.cvttsd2usi64(<2 x double> %v71, i32 4)
  %t8 = add i32 %t0, %t2
  %t9 = add i32 %t4, %t6
  %t10 = add i32 %t8, %t9
  %t11 = sext i32 %t10 to i64
  %t12 = add i64 %t1, %t3
  %t13 = add i64 %t5, %t7
  %t14 = add i64 %t12, %t13
  %t15 = add i64 %t11, %t14
  ret i64 %t15
}

declare i32 @llvm.x86.avx512.vcvtss2usi32(<4 x float>, i32)
declare i64 @llvm.x86.avx512.vcvtss2usi64(<4 x float>, i32)
declare i32 @llvm.x86.avx512.cvttss2usi(<4 x float>, i32)
declare i64 @llvm.x86.avx512.cvttss2usi64(<4 x float>, i32)
declare i32 @llvm.x86.avx512.vcvtsd2usi32(<2 x double>, i32)
declare i64 @llvm.x86.avx512.vcvtsd2usi64(<2 x double>, i32)
declare i32 @llvm.x86.avx512.cvttsd2usi(<2 x double>, i32)
declare i64 @llvm.x86.avx512.cvttsd2usi64(<2 x double>, i32)

declare float @llvm.fma.f32(float, float, float) #1

define <4 x float> @test_mask_vfmadd_ss(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask_vfmadd_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP1]], float [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x float> [[A]], float [[TMP6]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP7]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = insertelement <4 x float> %c, float 4.000000e+00, i32 1
  %5 = insertelement <4 x float> %4, float 5.000000e+00, i32 2
  %6 = insertelement <4 x float> %5, float 6.000000e+00, i32 3
  %7 = extractelement <4 x float> %a, i64 0
  %8 = extractelement <4 x float> %3, i64 0
  %9 = extractelement <4 x float> %6, i64 0
  %10 = call float @llvm.fma.f32(float %7, float %8, float %9)
  %11 = bitcast i8 %mask to <8 x i1>
  %12 = extractelement <8 x i1> %11, i64 0
  %13 = select i1 %12, float %10, float %7
  %14 = insertelement <4 x float> %a, float %13, i64 0
  ret <4 x float> %14
}

define float @test_mask_vfmadd_ss_0(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask_vfmadd_ss_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP1]], float [[TMP4]]
; CHECK-NEXT:    ret float [[TMP6]]
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = extractelement <4 x float> %3, i64 0
  %5 = extractelement <4 x float> %b, i64 0
  %6 = extractelement <4 x float> %c, i64 0
  %7 = call float @llvm.fma.f32(float %4, float %5, float %6)
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, float %7, float %4
  %11 = insertelement <4 x float> %3, float %10, i64 0
  %12 = extractelement <4 x float> %11, i32 0
  ret float %12
}

define float @test_mask_vfmadd_ss_1(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask_vfmadd_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = extractelement <4 x float> %3, i64 0
  %5 = extractelement <4 x float> %b, i64 0
  %6 = extractelement <4 x float> %c, i64 0
  %7 = call float @llvm.fma.f32(float %4, float %5, float %6)
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, float %7, float %4
  %11 = insertelement <4 x float> %3, float %10, i64 0
  %12 = extractelement <4 x float> %11, i32 1
  ret float %12
}

declare double @llvm.fma.f64(double, double, double) #1

define <2 x double> @test_mask_vfmadd_sd(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask_vfmadd_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP1]], double [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> [[A]], double [[TMP6]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP7]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = insertelement <2 x double> %c, double 2.000000e+00, i32 1
  %3 = extractelement <2 x double> %a, i64 0
  %4 = extractelement <2 x double> %1, i64 0
  %5 = extractelement <2 x double> %2, i64 0
  %6 = call double @llvm.fma.f64(double %3, double %4, double %5)
  %7 = bitcast i8 %mask to <8 x i1>
  %8 = extractelement <8 x i1> %7, i64 0
  %9 = select i1 %8, double %6, double %3
  %10 = insertelement <2 x double> %a, double %9, i64 0
  ret <2 x double> %10
}

define double @test_mask_vfmadd_sd_0(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask_vfmadd_sd_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP1]], double [[TMP4]]
; CHECK-NEXT:    ret double [[TMP6]]
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = extractelement <2 x double> %1, i64 0
  %3 = extractelement <2 x double> %b, i64 0
  %4 = extractelement <2 x double> %c, i64 0
  %5 = call double @llvm.fma.f64(double %2, double %3, double %4)
  %6 = bitcast i8 %mask to <8 x i1>
  %7 = extractelement <8 x i1> %6, i64 0
  %8 = select i1 %7, double %5, double %2
  %9 = insertelement <2 x double> %1, double %8, i64 0
  %10 = extractelement <2 x double> %9, i32 0
  ret double %10
}

define double @test_mask_vfmadd_sd_1(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask_vfmadd_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = extractelement <2 x double> %1, i64 0
  %3 = extractelement <2 x double> %b, i64 0
  %4 = extractelement <2 x double> %c, i64 0
  %5 = call double @llvm.fma.f64(double %2, double %3, double %4)
  %6 = bitcast i8 %mask to <8 x i1>
  %7 = extractelement <8 x i1> %6, i64 0
  %8 = select i1 %7, double %5, double %2
  %9 = insertelement <2 x double> %1, double %8, i64 0
  %10 = extractelement <2 x double> %9, i32 1
  ret double %10
}

define <4 x float> @test_maskz_vfmadd_ss(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_maskz_vfmadd_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float 0.000000e+00, float [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x float> [[A]], float [[TMP6]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP7]]
;
  %1 = insertelement <4 x float> %b, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = insertelement <4 x float> %c, float 4.000000e+00, i32 1
  %5 = insertelement <4 x float> %4, float 5.000000e+00, i32 2
  %6 = insertelement <4 x float> %5, float 6.000000e+00, i32 3
  %7 = extractelement <4 x float> %a, i64 0
  %8 = extractelement <4 x float> %3, i64 0
  %9 = extractelement <4 x float> %6, i64 0
  %10 = call float @llvm.fma.f32(float %7, float %8, float %9)
  %11 = bitcast i8 %mask to <8 x i1>
  %12 = extractelement <8 x i1> %11, i64 0
  %13 = select i1 %12, float %10, float 0.000000e+00
  %14 = insertelement <4 x float> %a, float %13, i64 0
  ret <4 x float> %14
}

define float @test_maskz_vfmadd_ss_0(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_maskz_vfmadd_ss_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float 0.000000e+00, float [[TMP4]]
; CHECK-NEXT:    ret float [[TMP6]]
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = extractelement <4 x float> %3, i64 0
  %5 = extractelement <4 x float> %b, i64 0
  %6 = extractelement <4 x float> %c, i64 0
  %7 = call float @llvm.fma.f32(float %4, float %5, float %6)
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, float %7, float 0.000000e+00
  %11 = insertelement <4 x float> %3, float %10, i64 0
  %12 = extractelement <4 x float> %11, i32 0
  ret float %12
}

define float @test_maskz_vfmadd_ss_1(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_maskz_vfmadd_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = extractelement <4 x float> %3, i64 0
  %5 = extractelement <4 x float> %b, i64 0
  %6 = extractelement <4 x float> %c, i64 0
  %7 = call float @llvm.fma.f32(float %4, float %5, float %6)
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, float %7, float 0.000000e+00
  %11 = insertelement <4 x float> %3, float %10, i64 0
  %12 = extractelement <4 x float> %11, i32 1
  ret float %12
}

define <2 x double> @test_maskz_vfmadd_sd(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_maskz_vfmadd_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double 0.000000e+00, double [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> [[A]], double [[TMP6]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP7]]
;
  %1 = insertelement <2 x double> %b, double 1.000000e+00, i32 1
  %2 = insertelement <2 x double> %c, double 2.000000e+00, i32 1
  %3 = extractelement <2 x double> %a, i64 0
  %4 = extractelement <2 x double> %1, i64 0
  %5 = extractelement <2 x double> %2, i64 0
  %6 = call double @llvm.fma.f64(double %3, double %4, double %5)
  %7 = bitcast i8 %mask to <8 x i1>
  %8 = extractelement <8 x i1> %7, i64 0
  %9 = select i1 %8, double %6, double 0.000000e+00
  %10 = insertelement <2 x double> %a, double %9, i64 0
  ret <2 x double> %10
}

define double @test_maskz_vfmadd_sd_0(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_maskz_vfmadd_sd_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double 0.000000e+00, double [[TMP4]]
; CHECK-NEXT:    ret double [[TMP6]]
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = extractelement <2 x double> %1, i64 0
  %3 = extractelement <2 x double> %b, i64 0
  %4 = extractelement <2 x double> %c, i64 0
  %5 = call double @llvm.fma.f64(double %2, double %3, double %4)
  %6 = bitcast i8 %mask to <8 x i1>
  %7 = extractelement <8 x i1> %6, i64 0
  %8 = select i1 %7, double %5, double 0.000000e+00
  %9 = insertelement <2 x double> %1, double %8, i64 0
  %10 = extractelement <2 x double> %9, i32 0
  ret double %10
}

define double @test_maskz_vfmadd_sd_1(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_maskz_vfmadd_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = extractelement <2 x double> %1, i64 0
  %3 = extractelement <2 x double> %b, i64 0
  %4 = extractelement <2 x double> %c, i64 0
  %5 = call double @llvm.fma.f64(double %2, double %3, double %4)
  %6 = bitcast i8 %mask to <8 x i1>
  %7 = extractelement <8 x i1> %6, i64 0
  %8 = select i1 %7, double %5, double 0.000000e+00
  %9 = insertelement <2 x double> %1, double %8, i64 0
  %10 = extractelement <2 x double> %9, i32 1
  ret double %10
}

define <4 x float> @test_mask3_vfmadd_ss(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmadd_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP3]], float [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <4 x float> [[C]], float [[TMP6]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP7]]
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = insertelement <4 x float> %b, float 4.000000e+00, i32 1
  %5 = insertelement <4 x float> %4, float 5.000000e+00, i32 2
  %6 = insertelement <4 x float> %5, float 6.000000e+00, i32 3
  %7 = extractelement <4 x float> %3, i64 0
  %8 = extractelement <4 x float> %6, i64 0
  %9 = extractelement <4 x float> %c, i64 0
  %10 = call float @llvm.fma.f32(float %7, float %8, float %9)
  %11 = bitcast i8 %mask to <8 x i1>
  %12 = extractelement <8 x i1> %11, i64 0
  %13 = select i1 %12, float %10, float %9
  %14 = insertelement <4 x float> %c, float %13, i64 0
  ret <4 x float> %14
}

define float @test_mask3_vfmadd_ss_0(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmadd_ss_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], float [[TMP3]], float [[TMP4]]
; CHECK-NEXT:    ret float [[TMP6]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = extractelement <4 x float> %a, i64 0
  %5 = extractelement <4 x float> %b, i64 0
  %6 = extractelement <4 x float> %3, i64 0
  %7 = call float @llvm.fma.f32(float %4, float %5, float %6)
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, float %7, float %6
  %11 = insertelement <4 x float> %3, float %10, i64 0
  %12 = extractelement <4 x float> %11, i32 0
  ret float %12
}

define float @test_mask3_vfmadd_ss_1(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmadd_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = extractelement <4 x float> %a, i64 0
  %5 = extractelement <4 x float> %b, i64 0
  %6 = extractelement <4 x float> %3, i64 0
  %7 = call float @llvm.fma.f32(float %4, float %5, float %6)
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, float %7, float %6
  %11 = insertelement <4 x float> %3, float %10, i64 0
  %12 = extractelement <4 x float> %11, i32 1
  ret float %12
}

define <2 x double> @test_mask3_vfmadd_sd(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmadd_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP3]], double [[TMP4]]
; CHECK-NEXT:    [[TMP7:%.*]] = insertelement <2 x double> [[C]], double [[TMP6]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP7]]
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = insertelement <2 x double> %b, double 2.000000e+00, i32 1
  %3 = extractelement <2 x double> %1, i64 0
  %4 = extractelement <2 x double> %2, i64 0
  %5 = extractelement <2 x double> %c, i64 0
  %6 = call double @llvm.fma.f64(double %3, double %4, double %5)
  %7 = bitcast i8 %mask to <8 x i1>
  %8 = extractelement <8 x i1> %7, i64 0
  %9 = select i1 %8, double %6, double %5
  %10 = insertelement <2 x double> %c, double %9, i64 0
  ret <2 x double> %10
}

define double @test_mask3_vfmadd_sd_0(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmadd_sd_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP3]])
; CHECK-NEXT:    [[TMP5:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP5]], 0
; CHECK-NEXT:    [[TMP6:%.*]] = select i1 [[DOTNOT]], double [[TMP3]], double [[TMP4]]
; CHECK-NEXT:    ret double [[TMP6]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = extractelement <2 x double> %a, i64 0
  %3 = extractelement <2 x double> %b, i64 0
  %4 = extractelement <2 x double> %1, i64 0
  %5 = call double @llvm.fma.f64(double %2, double %3, double %4)
  %6 = bitcast i8 %mask to <8 x i1>
  %7 = extractelement <8 x i1> %6, i64 0
  %8 = select i1 %7, double %5, double %4
  %9 = insertelement <2 x double> %1, double %8, i64 0
  %10 = extractelement <2 x double> %9, i32 0
  ret double %10
}

define double @test_mask3_vfmadd_sd_1(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmadd_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = extractelement <2 x double> %a, i64 0
  %3 = extractelement <2 x double> %b, i64 0
  %4 = extractelement <2 x double> %1, i64 0
  %5 = call double @llvm.fma.f64(double %2, double %3, double %4)
  %6 = bitcast i8 %mask to <8 x i1>
  %7 = extractelement <8 x i1> %6, i64 0
  %8 = select i1 %7, double %5, double %4
  %9 = insertelement <2 x double> %1, double %8, i64 0
  %10 = extractelement <2 x double> %9, i32 1
  ret double %10
}

define <4 x float> @test_mask3_vfmsub_ss(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = fneg float [[TMP3]]
; CHECK-NEXT:    [[TMP5:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP4]])
; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <4 x float> [[C]], i64 0
; CHECK-NEXT:    [[TMP7:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP7]], 0
; CHECK-NEXT:    [[TMP8:%.*]] = select i1 [[DOTNOT]], float [[TMP6]], float [[TMP5]]
; CHECK-NEXT:    [[TMP9:%.*]] = insertelement <4 x float> [[C]], float [[TMP8]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP9]]
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = insertelement <4 x float> %b, float 4.000000e+00, i32 1
  %5 = insertelement <4 x float> %4, float 5.000000e+00, i32 2
  %6 = insertelement <4 x float> %5, float 6.000000e+00, i32 3
  %7 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %c
  %8 = extractelement <4 x float> %3, i64 0
  %9 = extractelement <4 x float> %6, i64 0
  %10 = extractelement <4 x float> %7, i64 0
  %11 = call float @llvm.fma.f32(float %8, float %9, float %10)
  %12 = extractelement <4 x float> %c, i64 0
  %13 = bitcast i8 %mask to <8 x i1>
  %14 = extractelement <8 x i1> %13, i64 0
  %15 = select i1 %14, float %11, float %12
  %16 = insertelement <4 x float> %c, float %15, i64 0
  ret <4 x float> %16
}

define float @test_mask3_vfmsub_ss_0(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_ss_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = fneg float [[TMP3]]
; CHECK-NEXT:    [[TMP5:%.*]] = call float @llvm.fma.f32(float [[TMP1]], float [[TMP2]], float [[TMP4]])
; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <4 x float> [[C]], i64 0
; CHECK-NEXT:    [[TMP7:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP7]], 0
; CHECK-NEXT:    [[TMP8:%.*]] = select i1 [[DOTNOT]], float [[TMP6]], float [[TMP5]]
; CHECK-NEXT:    ret float [[TMP8]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %3
  %5 = extractelement <4 x float> %a, i64 0
  %6 = extractelement <4 x float> %b, i64 0
  %7 = extractelement <4 x float> %4, i64 0
  %8 = call float @llvm.fma.f32(float %5, float %6, float %7)
  %9 = extractelement <4 x float> %3, i64 0
  %10 = bitcast i8 %mask to <8 x i1>
  %11 = extractelement <8 x i1> %10, i64 0
  %12 = select i1 %11, float %8, float %9
  %13 = insertelement <4 x float> %3, float %12, i64 0
  %14 = extractelement <4 x float> %13, i32 0
  ret float %14
}

define float @test_mask3_vfmsub_ss_1(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %3
  %5 = extractelement <4 x float> %a, i64 0
  %6 = extractelement <4 x float> %b, i64 0
  %7 = extractelement <4 x float> %4, i64 0
  %8 = call float @llvm.fma.f32(float %5, float %6, float %7)
  %9 = extractelement <4 x float> %3, i64 0
  %10 = bitcast i8 %mask to <8 x i1>
  %11 = extractelement <8 x i1> %10, i64 0
  %12 = select i1 %11, float %8, float %9
  %13 = insertelement <4 x float> %3, float %12, i64 0
  %14 = extractelement <4 x float> %13, i32 1
  ret float %14
}

define float @test_mask3_vfmsub_ss_1_unary_fneg(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_ss_1_unary_fneg(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = fneg <4 x float> %3
  %5 = extractelement <4 x float> %a, i64 0
  %6 = extractelement <4 x float> %b, i64 0
  %7 = extractelement <4 x float> %4, i64 0
  %8 = call float @llvm.fma.f32(float %5, float %6, float %7)
  %9 = extractelement <4 x float> %3, i64 0
  %10 = bitcast i8 %mask to <8 x i1>
  %11 = extractelement <8 x i1> %10, i64 0
  %12 = select i1 %11, float %8, float %9
  %13 = insertelement <4 x float> %3, float %12, i64 0
  %14 = extractelement <4 x float> %13, i32 1
  ret float %14
}

define <2 x double> @test_mask3_vfmsub_sd(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = fneg double [[TMP3]]
; CHECK-NEXT:    [[TMP5:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP4]])
; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x double> [[C]], i64 0
; CHECK-NEXT:    [[TMP7:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP7]], 0
; CHECK-NEXT:    [[TMP8:%.*]] = select i1 [[DOTNOT]], double [[TMP6]], double [[TMP5]]
; CHECK-NEXT:    [[TMP9:%.*]] = insertelement <2 x double> [[C]], double [[TMP8]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP9]]
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = insertelement <2 x double> %b, double 2.000000e+00, i32 1
  %3 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %c
  %4 = extractelement <2 x double> %1, i64 0
  %5 = extractelement <2 x double> %2, i64 0
  %6 = extractelement <2 x double> %3, i64 0
  %7 = call double @llvm.fma.f64(double %4, double %5, double %6)
  %8 = extractelement <2 x double> %c, i64 0
  %9 = bitcast i8 %mask to <8 x i1>
  %10 = extractelement <8 x i1> %9, i64 0
  %11 = select i1 %10, double %7, double %8
  %12 = insertelement <2 x double> %c, double %11, i64 0
  ret <2 x double> %12
}

define double @test_mask3_vfmsub_sd_0(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_sd_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = fneg double [[TMP3]]
; CHECK-NEXT:    [[TMP5:%.*]] = call double @llvm.fma.f64(double [[TMP1]], double [[TMP2]], double [[TMP4]])
; CHECK-NEXT:    [[TMP6:%.*]] = extractelement <2 x double> [[C]], i64 0
; CHECK-NEXT:    [[TMP7:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP7]], 0
; CHECK-NEXT:    [[TMP8:%.*]] = select i1 [[DOTNOT]], double [[TMP6]], double [[TMP5]]
; CHECK-NEXT:    ret double [[TMP8]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %1
  %3 = extractelement <2 x double> %a, i64 0
  %4 = extractelement <2 x double> %b, i64 0
  %5 = extractelement <2 x double> %2, i64 0
  %6 = call double @llvm.fma.f64(double %3, double %4, double %5)
  %7 = extractelement <2 x double> %1, i64 0
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, double %6, double %7
  %11 = insertelement <2 x double> %1, double %10, i64 0
  %12 = extractelement <2 x double> %11, i32 0
  ret double %12
}

define double @test_mask3_vfmsub_sd_1(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %1
  %3 = extractelement <2 x double> %a, i64 0
  %4 = extractelement <2 x double> %b, i64 0
  %5 = extractelement <2 x double> %2, i64 0
  %6 = call double @llvm.fma.f64(double %3, double %4, double %5)
  %7 = extractelement <2 x double> %1, i64 0
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, double %6, double %7
  %11 = insertelement <2 x double> %1, double %10, i64 0
  %12 = extractelement <2 x double> %11, i32 1
  ret double %12
}

define double @test_mask3_vfmsub_sd_1_unary_fneg(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfmsub_sd_1_unary_fneg(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = fneg <2 x double> %1
  %3 = extractelement <2 x double> %a, i64 0
  %4 = extractelement <2 x double> %b, i64 0
  %5 = extractelement <2 x double> %2, i64 0
  %6 = call double @llvm.fma.f64(double %3, double %4, double %5)
  %7 = extractelement <2 x double> %1, i64 0
  %8 = bitcast i8 %mask to <8 x i1>
  %9 = extractelement <8 x i1> %8, i64 0
  %10 = select i1 %9, double %6, double %7
  %11 = insertelement <2 x double> %1, double %10, i64 0
  %12 = extractelement <2 x double> %11, i32 1
  ret double %12
}

define <4 x float> @test_mask3_vfnmsub_ss(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_ss(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = fneg float [[TMP1]]
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP5:%.*]] = fneg float [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = call float @llvm.fma.f32(float [[TMP2]], float [[TMP3]], float [[TMP5]])
; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <4 x float> [[C]], i64 0
; CHECK-NEXT:    [[TMP8:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP8]], 0
; CHECK-NEXT:    [[TMP9:%.*]] = select i1 [[DOTNOT]], float [[TMP7]], float [[TMP6]]
; CHECK-NEXT:    [[TMP10:%.*]] = insertelement <4 x float> [[C]], float [[TMP9]], i64 0
; CHECK-NEXT:    ret <4 x float> [[TMP10]]
;
  %1 = insertelement <4 x float> %a, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = insertelement <4 x float> %b, float 4.000000e+00, i32 1
  %5 = insertelement <4 x float> %4, float 5.000000e+00, i32 2
  %6 = insertelement <4 x float> %5, float 6.000000e+00, i32 3
  %7 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %3
  %8 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %c
  %9 = extractelement <4 x float> %7, i64 0
  %10 = extractelement <4 x float> %6, i64 0
  %11 = extractelement <4 x float> %8, i64 0
  %12 = call float @llvm.fma.f32(float %9, float %10, float %11)
  %13 = extractelement <4 x float> %c, i64 0
  %14 = bitcast i8 %mask to <8 x i1>
  %15 = extractelement <8 x i1> %14, i64 0
  %16 = select i1 %15, float %12, float %13
  %17 = insertelement <4 x float> %c, float %16, i64 0
  ret <4 x float> %17
}

define float @test_mask3_vfnmsub_ss_0(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_ss_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <4 x float> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = fneg float [[TMP1]]
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <4 x float> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <4 x float> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP5:%.*]] = fneg float [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = call float @llvm.fma.f32(float [[TMP2]], float [[TMP3]], float [[TMP5]])
; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <4 x float> [[C]], i64 0
; CHECK-NEXT:    [[TMP8:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP8]], 0
; CHECK-NEXT:    [[TMP9:%.*]] = select i1 [[DOTNOT]], float [[TMP7]], float [[TMP6]]
; CHECK-NEXT:    ret float [[TMP9]]
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a
  %5 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %3
  %6 = extractelement <4 x float> %4, i64 0
  %7 = extractelement <4 x float> %b, i64 0
  %8 = extractelement <4 x float> %5, i64 0
  %9 = call float @llvm.fma.f32(float %6, float %7, float %8)
  %10 = extractelement <4 x float> %3, i64 0
  %11 = bitcast i8 %mask to <8 x i1>
  %12 = extractelement <8 x i1> %11, i64 0
  %13 = select i1 %12, float %9, float %10
  %14 = insertelement <4 x float> %3, float %13, i64 0
  %15 = extractelement <4 x float> %14, i32 0
  ret float %15
}

define float @test_mask3_vfnmsub_ss_1(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_ss_1(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %a
  %5 = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %3
  %6 = extractelement <4 x float> %4, i64 0
  %7 = extractelement <4 x float> %b, i64 0
  %8 = extractelement <4 x float> %5, i64 0
  %9 = call float @llvm.fma.f32(float %6, float %7, float %8)
  %10 = extractelement <4 x float> %3, i64 0
  %11 = bitcast i8 %mask to <8 x i1>
  %12 = extractelement <8 x i1> %11, i64 0
  %13 = select i1 %12, float %9, float %10
  %14 = insertelement <4 x float> %3, float %13, i64 0
  %15 = extractelement <4 x float> %14, i32 1
  ret float %15
}

define float @test_mask3_vfnmsub_ss_1_unary_fneg(<4 x float> %a, <4 x float> %b, <4 x float> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_ss_1_unary_fneg(
; CHECK-NEXT:    ret float 1.000000e+00
;
  %1 = insertelement <4 x float> %c, float 1.000000e+00, i32 1
  %2 = insertelement <4 x float> %1, float 2.000000e+00, i32 2
  %3 = insertelement <4 x float> %2, float 3.000000e+00, i32 3
  %4 = fneg <4 x float> %a
  %5 = fneg <4 x float> %3
  %6 = extractelement <4 x float> %4, i64 0
  %7 = extractelement <4 x float> %b, i64 0
  %8 = extractelement <4 x float> %5, i64 0
  %9 = call float @llvm.fma.f32(float %6, float %7, float %8)
  %10 = extractelement <4 x float> %3, i64 0
  %11 = bitcast i8 %mask to <8 x i1>
  %12 = extractelement <8 x i1> %11, i64 0
  %13 = select i1 %12, float %9, float %10
  %14 = insertelement <4 x float> %3, float %13, i64 0
  %15 = extractelement <4 x float> %14, i32 1
  ret float %15
}

define <2 x double> @test_mask3_vfnmsub_sd(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_sd(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = fneg double [[TMP1]]
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP5:%.*]] = fneg double [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = call double @llvm.fma.f64(double [[TMP2]], double [[TMP3]], double [[TMP5]])
; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <2 x double> [[C]], i64 0
; CHECK-NEXT:    [[TMP8:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP8]], 0
; CHECK-NEXT:    [[TMP9:%.*]] = select i1 [[DOTNOT]], double [[TMP7]], double [[TMP6]]
; CHECK-NEXT:    [[TMP10:%.*]] = insertelement <2 x double> [[C]], double [[TMP9]], i64 0
; CHECK-NEXT:    ret <2 x double> [[TMP10]]
;
  %1 = insertelement <2 x double> %a, double 1.000000e+00, i32 1
  %2 = insertelement <2 x double> %b, double 2.000000e+00, i32 1
  %3 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %1
  %4 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %c
  %5 = extractelement <2 x double> %3, i64 0
  %6 = extractelement <2 x double> %2, i64 0
  %7 = extractelement <2 x double> %4, i64 0
  %8 = call double @llvm.fma.f64(double %5, double %6, double %7)
  %9 = extractelement <2 x double> %c, i64 0
  %10 = bitcast i8 %mask to <8 x i1>
  %11 = extractelement <8 x i1> %10, i64 0
  %12 = select i1 %11, double %8, double %9
  %13 = insertelement <2 x double> %c, double %12, i64 0
  ret <2 x double> %13
}

define double @test_mask3_vfnmsub_sd_0(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_sd_0(
; CHECK-NEXT:    [[TMP1:%.*]] = extractelement <2 x double> [[A:%.*]], i64 0
; CHECK-NEXT:    [[TMP2:%.*]] = fneg double [[TMP1]]
; CHECK-NEXT:    [[TMP3:%.*]] = extractelement <2 x double> [[B:%.*]], i64 0
; CHECK-NEXT:    [[TMP4:%.*]] = extractelement <2 x double> [[C:%.*]], i64 0
; CHECK-NEXT:    [[TMP5:%.*]] = fneg double [[TMP4]]
; CHECK-NEXT:    [[TMP6:%.*]] = call double @llvm.fma.f64(double [[TMP2]], double [[TMP3]], double [[TMP5]])
; CHECK-NEXT:    [[TMP7:%.*]] = extractelement <2 x double> [[C]], i64 0
; CHECK-NEXT:    [[TMP8:%.*]] = and i8 [[MASK:%.*]], 1
; CHECK-NEXT:    [[DOTNOT:%.*]] = icmp eq i8 [[TMP8]], 0
; CHECK-NEXT:    [[TMP9:%.*]] = select i1 [[DOTNOT]], double [[TMP7]], double [[TMP6]]
; CHECK-NEXT:    ret double [[TMP9]]
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %a
  %3 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %1
  %4 = extractelement <2 x double> %2, i64 0
  %5 = extractelement <2 x double> %b, i64 0
  %6 = extractelement <2 x double> %3, i64 0
  %7 = call double @llvm.fma.f64(double %4, double %5, double %6)
  %8 = extractelement <2 x double> %1, i64 0
  %9 = bitcast i8 %mask to <8 x i1>
  %10 = extractelement <8 x i1> %9, i64 0
  %11 = select i1 %10, double %7, double %8
  %12 = insertelement <2 x double> %1, double %11, i64 0
  %13 = extractelement <2 x double> %12, i32 0
  ret double %13
}

define double @test_mask3_vfnmsub_sd_1(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_sd_1(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %a
  %3 = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %1
  %4 = extractelement <2 x double> %2, i64 0
  %5 = extractelement <2 x double> %b, i64 0
  %6 = extractelement <2 x double> %3, i64 0
  %7 = call double @llvm.fma.f64(double %4, double %5, double %6)
  %8 = extractelement <2 x double> %1, i64 0
  %9 = bitcast i8 %mask to <8 x i1>
  %10 = extractelement <8 x i1> %9, i64 0
  %11 = select i1 %10, double %7, double %8
  %12 = insertelement <2 x double> %1, double %11, i64 0
  %13 = extractelement <2 x double> %12, i32 1
  ret double %13
}

define double @test_mask3_vfnmsub_sd_1_unary_fneg(<2 x double> %a, <2 x double> %b, <2 x double> %c, i8 %mask) {
;
; CHECK-LABEL: @test_mask3_vfnmsub_sd_1_unary_fneg(
; CHECK-NEXT:    ret double 1.000000e+00
;
  %1 = insertelement <2 x double> %c, double 1.000000e+00, i32 1
  %2 = fneg <2 x double> %a
  %3 = fneg <2 x double> %1
  %4 = extractelement <2 x double> %2, i64 0
  %5 = extractelement <2 x double> %b, i64 0
  %6 = extractelement <2 x double> %3, i64 0
  %7 = call double @llvm.fma.f64(double %4, double %5, double %6)
  %8 = extractelement <2 x double> %1, i64 0
  %9 = bitcast i8 %mask to <8 x i1>
  %10 = extractelement <8 x i1> %9, i64 0
  %11 = select i1 %10, double %7, double %8
  %12 = insertelement <2 x double> %1, double %11, i64 0
  %13 = extractelement <2 x double> %12, i32 1
  ret double %13
}

declare <8 x i32> @llvm.x86.avx2.permd(<8 x i32>, <8 x i32>)

define <8 x i32> @identity_test_permvar_si_256(<8 x i32> %a0) {
;
; CHECK-LABEL: @identity_test_permvar_si_256(
; CHECK-NEXT:    ret <8 x i32> [[A0:%.*]]
;
  %1 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>)
  ret <8 x i32> %1
}

define <8 x i32> @identity_test_permvar_si_256_mask(<8 x i32> %a0, <8 x i32> %passthru, i8 %mask) {
;
; CHECK-LABEL: @identity_test_permvar_si_256_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = bitcast i8 [[MASK:%.*]] to <8 x i1>
; CHECK-NEXT:    [[TMP2:%.*]] = select <8 x i1> [[TMP1]], <8 x i32> [[A0:%.*]], <8 x i32> [[PASSTHRU:%.*]]
; CHECK-NEXT:    ret <8 x i32> [[TMP2]]
;
  %1 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> <i32 0, i32 1, i32 2, i32 3, i32 4, i32 5, i32 6, i32 7>)
  %2 = bitcast i8 %mask to <8 x i1>
  %3 = select <8 x i1> %2, <8 x i32> %1, <8 x i32> %passthru
  ret <8 x i32> %3
}

define <8 x i32> @zero_test_permvar_si_256(<8 x i32> %a0) {
;
; CHECK-LABEL: @zero_test_permvar_si_256(
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x i32> [[A0:%.*]], <8 x i32> poison, <8 x i32> zeroinitializer
; CHECK-NEXT:    ret <8 x i32> [[TMP1]]
;
  %1 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> zeroinitializer)
  ret <8 x i32> %1
}

define <8 x i32> @zero_test_permvar_si_256_mask(<8 x i32> %a0, <8 x i32> %passthru, i8 %mask) {
;
; CHECK-LABEL: @zero_test_permvar_si_256_mask(
; CHECK-NEXT:    [[TMP1:%.*]] = shufflevector <8 x i32> [[A0:%.*]], <8 x i32> poison, <8 x i32> zeroinitializer
; CHECK-NEXT:    [[TMP2:%.*]] = bitcast i8 [[MASK:%.*]] to <8 x i1>
; CHECK-NEXT:    [[TMP3:%.*]] = select <8 x i1> [[TMP2]], <8 x i32> [[TMP1]], <8 x i32> [[PASSTHRU:%.*]]
; CHECK-NEXT:    ret <8 x i32> [[TMP3]]
;
  %1 = call <8 x i32> @llvm.x86.avx2.permd(<8 x i32> %a0, <8 x i32> zeroinitializer)
  %2 = bitcast i8 %mask to <8 x i1>
  %3 = select <8