; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+fma4,-fma -show-mc-encoding | FileCheck %s --check-prefix=CHECK
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+fma4,+fma -show-mc-encoding | FileCheck %s --check-prefix=CHECK

define <4 x float> @test_x86_fma_vfmadd_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmadd_ps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddps %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x68,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) + xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.fma.vfmadd.ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.fma.vfmadd.ps(<4 x float>, <4 x float>, <4 x float>)

define <2 x double> @test_x86_fma_vfmadd_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmadd_pd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddpd %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x69,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) + xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.fma.vfmadd.pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
  ret <2 x double> %res
}
declare <2 x double> @llvm.x86.fma.vfmadd.pd(<2 x double>, <2 x double>, <2 x double>)

define <8 x float> @test_x86_fma_vfmadd_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmadd_ps_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddps %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x68,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) + ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.fma.vfmadd.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfmadd.ps.256(<8 x float>, <8 x float>, <8 x float>)

define <4 x double> @test_x86_fma_vfmadd_pd_256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmadd_pd_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddpd %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x69,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) + ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.fma.vfmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
  ret <4 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfmadd.pd.256(<4 x double>, <4 x double>, <4 x double>)

; VFMSUB
define <4 x float> @test_x86_fma_vfmsub_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsub_ps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubps %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x6c,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) - xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.fma.vfmsub.ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.fma.vfmsub.ps(<4 x float>, <4 x float>, <4 x float>)

define <2 x double> @test_x86_fma_vfmsub_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsub_pd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubpd %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x6d,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) - xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.fma.vfmsub.pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
  ret <2 x double> %res
}
declare <2 x double> @llvm.x86.fma.vfmsub.pd(<2 x double>, <2 x double>, <2 x double>)

define <8 x float> @test_x86_fma_vfmsub_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsub_ps_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubps %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x6c,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) - ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.fma.vfmsub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfmsub.ps.256(<8 x float>, <8 x float>, <8 x float>)

define <4 x double> @test_x86_fma_vfmsub_pd_256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsub_pd_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubpd %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x6d,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) - ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.fma.vfmsub.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
  ret <4 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfmsub.pd.256(<4 x double>, <4 x double>, <4 x double>)

; VFNMADD
define <4 x float> @test_x86_fma_vfnmadd_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmadd_ps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmaddps %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x78,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = -(xmm0 * xmm1) + xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.fma.vfnmadd.ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.fma.vfnmadd.ps(<4 x float>, <4 x float>, <4 x float>)

define <2 x double> @test_x86_fma_vfnmadd_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmadd_pd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmaddpd %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x79,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = -(xmm0 * xmm1) + xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.fma.vfnmadd.pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
  ret <2 x double> %res
}
declare <2 x double> @llvm.x86.fma.vfnmadd.pd(<2 x double>, <2 x double>, <2 x double>)

define <8 x float> @test_x86_fma_vfnmadd_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmadd_ps_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmaddps %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x78,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = -(ymm0 * ymm1) + ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.fma.vfnmadd.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfnmadd.ps.256(<8 x float>, <8 x float>, <8 x float>)

define <4 x double> @test_x86_fma_vfnmadd_pd_256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmadd_pd_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmaddpd %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x79,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = -(ymm0 * ymm1) + ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.fma.vfnmadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
  ret <4 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfnmadd.pd.256(<4 x double>, <4 x double>, <4 x double>)

; VFNMSUB
define <4 x float> @test_x86_fma_vfnmsub_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmsub_ps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmsubps %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x7c,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = -(xmm0 * xmm1) - xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.fma.vfnmsub.ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.fma.vfnmsub.ps(<4 x float>, <4 x float>, <4 x float>)

define <2 x double> @test_x86_fma_vfnmsub_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmsub_pd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmsubpd %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x7d,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = -(xmm0 * xmm1) - xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.fma.vfnmsub.pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
  ret <2 x double> %res
}
declare <2 x double> @llvm.x86.fma.vfnmsub.pd(<2 x double>, <2 x double>, <2 x double>)

define <8 x float> @test_x86_fma_vfnmsub_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmsub_ps_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmsubps %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x7c,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = -(ymm0 * ymm1) - ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.fma.vfnmsub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfnmsub.ps.256(<8 x float>, <8 x float>, <8 x float>)

define <4 x double> @test_x86_fma_vfnmsub_pd_256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfnmsub_pd_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfnmsubpd %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x7d,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = -(ymm0 * ymm1) - ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.fma.vfnmsub.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
  ret <4 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfnmsub.pd.256(<4 x double>, <4 x double>, <4 x double>)

; VFMADDSUB
define <4 x float> @test_x86_fma_vfmaddsub_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmaddsub_ps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddsubps %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x5c,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) +/- xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.fma.vfmaddsub.ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.fma.vfmaddsub.ps(<4 x float>, <4 x float>, <4 x float>)

define <2 x double> @test_x86_fma_vfmaddsub_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmaddsub_pd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddsubpd %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x5d,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) +/- xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.fma.vfmaddsub.pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
  ret <2 x double> %res
}
declare <2 x double> @llvm.x86.fma.vfmaddsub.pd(<2 x double>, <2 x double>, <2 x double>)

define <8 x float> @test_x86_fma_vfmaddsub_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmaddsub_ps_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddsubps %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x5c,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) +/- ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.fma.vfmaddsub.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfmaddsub.ps.256(<8 x float>, <8 x float>, <8 x float>)

define <4 x double> @test_x86_fma_vfmaddsub_pd_256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmaddsub_pd_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmaddsubpd %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x5d,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) +/- ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.fma.vfmaddsub.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
  ret <4 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfmaddsub.pd.256(<4 x double>, <4 x double>, <4 x double>)

; VFMSUBADD
define <4 x float> @test_x86_fma_vfmsubadd_ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsubadd_ps:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubaddps %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x5e,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) -/+ xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x float> @llvm.x86.fma.vfmsubadd.ps(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2)
  ret <4 x float> %res
}
declare <4 x float> @llvm.x86.fma.vfmsubadd.ps(<4 x float>, <4 x float>, <4 x float>)

define <2 x double> @test_x86_fma_vfmsubadd_pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsubadd_pd:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubaddpd %xmm2, %xmm1, %xmm0, %xmm0 # encoding: [0xc4,0xe3,0xf9,0x5f,0xc2,0x10]
; CHECK-NEXT:    # xmm0 = (xmm0 * xmm1) -/+ xmm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <2 x double> @llvm.x86.fma.vfmsubadd.pd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2)
  ret <2 x double> %res
}
declare <2 x double> @llvm.x86.fma.vfmsubadd.pd(<2 x double>, <2 x double>, <2 x double>)

define <8 x float> @test_x86_fma_vfmsubadd_ps_256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsubadd_ps_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubaddps %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x5e,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) -/+ ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <8 x float> @llvm.x86.fma.vfmsubadd.ps.256(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2)
  ret <8 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfmsubadd.ps.256(<8 x float>, <8 x float>, <8 x float>)

define <4 x double> @test_x86_fma_vfmsubadd_pd_256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; CHECK-LABEL: test_x86_fma_vfmsubadd_pd_256:
; CHECK:       # %bb.0:
; CHECK-NEXT:    vfmsubaddpd %ymm2, %ymm1, %ymm0, %ymm0 # encoding: [0xc4,0xe3,0xfd,0x5f,0xc2,0x10]
; CHECK-NEXT:    # ymm0 = (ymm0 * ymm1) -/+ ymm2
; CHECK-NEXT:    retq # encoding: [0xc3]
  %res = call <4 x double> @llvm.x86.fma.vfmsubadd.pd.256(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2)
  ret <4 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfmsubadd.pd.256(<4 x double>, <4 x double>, <4 x double>)

attributes #0 = { nounwind }