OSchip
/
llvm-project
mirror of https://github.com/THU-DSP-LAB/llvm-project.git
13
0
Fork 6
Code Issues Packages Projects Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/fma_patterns.ll

2025 lines
72 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fma -fp-contract=fast | FileCheck %s --check-prefixes=FMA,FMA-INFS
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fma4,+fma -fp-contract=fast | FileCheck %s --check-prefixes=FMA4,FMA4-INFS
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fma4 -fp-contract=fast | FileCheck %s --check-prefixes=FMA4,FMA4-INFS
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512dq,+avx512vl -fp-contract=fast | FileCheck %s --check-prefixes=AVX512,AVX512-INFS
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fma -fp-contract=fast -enable-no-infs-fp-math | FileCheck %s --check-prefixes=FMA,FMA-NOINFS
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fma4,+fma -fp-contract=fast -enable-no-infs-fp-math | FileCheck %s --check-prefixes=FMA4,FMA4-NOINFS
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx,+fma4 -fp-contract=fast -enable-no-infs-fp-math | FileCheck %s --check-prefixes=FMA4,FMA4-NOINFS
; RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+avx512dq,+avx512vl -fp-contract=fast -enable-no-infs-fp-math | FileCheck %s --check-prefixes=AVX512,AVX512-NOINFS
;
; Pattern: (fadd (fmul x, y), z) -> (fmadd x,y,z)
;
define float @test_f32_fmadd(float %a0, float %a1, float %a2) {
; FMA-LABEL: test_f32_fmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f32_fmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f32_fmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul float %a0, %a1
%res = fadd float %x, %a2
ret float %res
}
define <4 x float> @test_4f32_fmadd(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
; FMA-LABEL: test_4f32_fmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f32_fmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f32_fmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul <4 x float> %a0, %a1
%res = fadd <4 x float> %x, %a2
ret <4 x float> %res
}
define <8 x float> @test_8f32_fmadd(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) {
; FMA-LABEL: test_8f32_fmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213ps {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_8f32_fmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddps {{.*#+}} ymm0 = (ymm0 * ymm1) + ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_8f32_fmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213ps {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
; AVX512-NEXT: retq
%x = fmul <8 x float> %a0, %a1
%res = fadd <8 x float> %x, %a2
ret <8 x float> %res
}
define double @test_f64_fmadd(double %a0, double %a1, double %a2) {
; FMA-LABEL: test_f64_fmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f64_fmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddsd {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f64_fmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul double %a0, %a1
%res = fadd double %x, %a2
ret double %res
}
define <2 x double> @test_2f64_fmadd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) {
; FMA-LABEL: test_2f64_fmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213pd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_2f64_fmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_2f64_fmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213pd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul <2 x double> %a0, %a1
%res = fadd <2 x double> %x, %a2
ret <2 x double> %res
}
define <4 x double> @test_4f64_fmadd(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) {
; FMA-LABEL: test_4f64_fmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f64_fmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddpd {{.*#+}} ymm0 = (ymm0 * ymm1) + ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f64_fmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
; AVX512-NEXT: retq
%x = fmul <4 x double> %a0, %a1
%res = fadd <4 x double> %x, %a2
ret <4 x double> %res
}
;
; Pattern: (fsub (fmul x, y), z) -> (fmsub x, y, z)
;
define float @test_f32_fmsub(float %a0, float %a1, float %a2) {
; FMA-LABEL: test_f32_fmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f32_fmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubss {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f32_fmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul float %a0, %a1
%res = fsub float %x, %a2
ret float %res
}
define <4 x float> @test_4f32_fmsub(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
; FMA-LABEL: test_4f32_fmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f32_fmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f32_fmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul <4 x float> %a0, %a1
%res = fsub <4 x float> %x, %a2
ret <4 x float> %res
}
define <8 x float> @test_8f32_fmsub(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) {
; FMA-LABEL: test_8f32_fmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub213ps {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_8f32_fmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubps {{.*#+}} ymm0 = (ymm0 * ymm1) - ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_8f32_fmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub213ps {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
; AVX512-NEXT: retq
%x = fmul <8 x float> %a0, %a1
%res = fsub <8 x float> %x, %a2
ret <8 x float> %res
}
define double @test_f64_fmsub(double %a0, double %a1, double %a2) {
; FMA-LABEL: test_f64_fmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f64_fmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubsd {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f64_fmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul double %a0, %a1
%res = fsub double %x, %a2
ret double %res
}
define <2 x double> @test_2f64_fmsub(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) {
; FMA-LABEL: test_2f64_fmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub213pd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_2f64_fmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubpd {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_2f64_fmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub213pd {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul <2 x double> %a0, %a1
%res = fsub <2 x double> %x, %a2
ret <2 x double> %res
}
define <4 x double> @test_4f64_fmsub(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) {
; FMA-LABEL: test_4f64_fmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub213pd {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f64_fmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubpd {{.*#+}} ymm0 = (ymm0 * ymm1) - ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f64_fmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub213pd {{.*#+}} ymm0 = (ymm1 * ymm0) - ymm2
; AVX512-NEXT: retq
%x = fmul <4 x double> %a0, %a1
%res = fsub <4 x double> %x, %a2
ret <4 x double> %res
}
;
; Pattern: (fsub z, (fmul x, y)) -> (fnmadd x, y, z)
;
define float @test_f32_fnmadd(float %a0, float %a1, float %a2) {
; FMA-LABEL: test_f32_fnmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f32_fnmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmaddss {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f32_fnmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmadd213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul float %a0, %a1
%res = fsub float %a2, %x
ret float %res
}
define <4 x float> @test_4f32_fnmadd(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
; FMA-LABEL: test_4f32_fnmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f32_fnmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmaddps {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f32_fnmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul <4 x float> %a0, %a1
%res = fsub <4 x float> %a2, %x
ret <4 x float> %res
}
define <8 x float> @test_8f32_fnmadd(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) {
; FMA-LABEL: test_8f32_fnmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfnmadd213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_8f32_fnmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmaddps {{.*#+}} ymm0 = -(ymm0 * ymm1) + ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_8f32_fnmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmadd213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
; AVX512-NEXT: retq
%x = fmul <8 x float> %a0, %a1
%res = fsub <8 x float> %a2, %x
ret <8 x float> %res
}
define double @test_f64_fnmadd(double %a0, double %a1, double %a2) {
; FMA-LABEL: test_f64_fnmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f64_fnmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmaddsd {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f64_fnmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmadd213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul double %a0, %a1
%res = fsub double %a2, %x
ret double %res
}
define <2 x double> @test_2f64_fnmadd(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) {
; FMA-LABEL: test_2f64_fnmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfnmadd213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_2f64_fnmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmaddpd {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_2f64_fnmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmadd213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%x = fmul <2 x double> %a0, %a1
%res = fsub <2 x double> %a2, %x
ret <2 x double> %res
}
define <4 x double> @test_4f64_fnmadd(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) {
; FMA-LABEL: test_4f64_fnmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfnmadd213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f64_fnmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmaddpd {{.*#+}} ymm0 = -(ymm0 * ymm1) + ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f64_fnmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmadd213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
; AVX512-NEXT: retq
%x = fmul <4 x double> %a0, %a1
%res = fsub <4 x double> %a2, %x
ret <4 x double> %res
}
;
; Pattern: (fsub (fneg (fmul x, y)), z) -> (fnmsub x, y, z)
;
define float @test_f32_fnmsub(float %a0, float %a1, float %a2) {
; FMA-LABEL: test_f32_fnmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f32_fnmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmsubss {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f32_fnmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmsub213ss {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul float %a0, %a1
%y = fsub float -0.000000e+00, %x
%res = fsub float %y, %a2
ret float %res
}
define <4 x float> @test_4f32_fnmsub(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) {
; FMA-LABEL: test_4f32_fnmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f32_fnmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f32_fnmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul <4 x float> %a0, %a1
%y = fsub <4 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x
%res = fsub <4 x float> %y, %a2
ret <4 x float> %res
}
define <8 x float> @test_8f32_fnmsub(<8 x float> %a0, <8 x float> %a1, <8 x float> %a2) {
; FMA-LABEL: test_8f32_fnmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfnmsub213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_8f32_fnmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmsubps {{.*#+}} ymm0 = -(ymm0 * ymm1) - ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_8f32_fnmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmsub213ps {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
; AVX512-NEXT: retq
%x = fmul <8 x float> %a0, %a1
%y = fsub <8 x float> <float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00, float -0.000000e+00>, %x
%res = fsub <8 x float> %y, %a2
ret <8 x float> %res
}
define double @test_f64_fnmsub(double %a0, double %a1, double %a2) {
; FMA-LABEL: test_f64_fnmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f64_fnmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmsubsd {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f64_fnmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul double %a0, %a1
%y = fsub double -0.000000e+00, %x
%res = fsub double %y, %a2
ret double %res
}
define <2 x double> @test_2f64_fnmsub(<2 x double> %a0, <2 x double> %a1, <2 x double> %a2) {
; FMA-LABEL: test_2f64_fnmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfnmsub213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_2f64_fnmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmsubpd {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_2f64_fnmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmsub213pd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%x = fmul <2 x double> %a0, %a1
%y = fsub <2 x double> <double -0.000000e+00, double -0.000000e+00>, %x
%res = fsub <2 x double> %y, %a2
ret <2 x double> %res
}
define <4 x double> @test_4f64_fnmsub(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) {
; FMA-LABEL: test_4f64_fnmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfnmsub213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f64_fnmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmsubpd {{.*#+}} ymm0 = -(ymm0 * ymm1) - ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f64_fnmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmsub213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
; AVX512-NEXT: retq
%x = fmul <4 x double> %a0, %a1
%y = fsub <4 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %x
%res = fsub <4 x double> %y, %a2
ret <4 x double> %res
}
;
; Load Folding Patterns
;
define <4 x float> @test_4f32_fmadd_load(<4 x float>* %a0, <4 x float> %a1, <4 x float> %a2) {
; FMA-LABEL: test_4f32_fmadd_load:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd132ps {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
; FMA-NEXT: retq
;
; FMA4-LABEL: test_4f32_fmadd_load:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_4f32_fmadd_load:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd132ps {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
; AVX512-NEXT: retq
%x = load <4 x float>, <4 x float>* %a0
%y = fmul <4 x float> %x, %a1
%res = fadd <4 x float> %y, %a2
ret <4 x float> %res
}
define <2 x double> @test_2f64_fmsub_load(<2 x double>* %a0, <2 x double> %a1, <2 x double> %a2) {
; FMA-LABEL: test_2f64_fmsub_load:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub132pd {{.*#+}} xmm0 = (xmm0 * mem) - xmm1
; FMA-NEXT: retq
;
; FMA4-LABEL: test_2f64_fmsub_load:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubpd {{.*#+}} xmm0 = (xmm0 * mem) - xmm1
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_2f64_fmsub_load:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub132pd {{.*#+}} xmm0 = (xmm0 * mem) - xmm1
; AVX512-NEXT: retq
%x = load <2 x double>, <2 x double>* %a0
%y = fmul <2 x double> %x, %a1
%res = fsub <2 x double> %y, %a2
ret <2 x double> %res
}
;
; Patterns (+ fneg variants): mul(add(1.0,x),y), mul(sub(1.0,x),y), mul(sub(x,1.0),y)
;
define <4 x float> @test_v4f32_mul_add_x_one_y(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_add_x_one_y:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_add_x_one_y:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_add_x_one_y:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_add_x_one_y:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_add_x_one_y:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_add_x_one_y:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%a = fadd <4 x float> %x, <float 1.0, float 1.0, float 1.0, float 1.0>
%m = fmul <4 x float> %a, %y
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_add_x_one(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_add_x_one:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_add_x_one:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_add_x_one:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_add_x_one:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_add_x_one:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_add_x_one:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%a = fadd <4 x float> %x, <float 1.0, float 1.0, float 1.0, float 1.0>
%m = fmul <4 x float> %y, %a
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_add_x_one_undefs(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_add_x_one_undefs:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_add_x_one_undefs:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_add_x_one_undefs:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_add_x_one_undefs:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_add_x_one_undefs:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_add_x_one_undefs:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%a = fadd <4 x float> %x, <float 1.0, float undef, float 1.0, float undef>
%m = fmul <4 x float> %y, %a
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_add_x_negone_y(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_add_x_negone_y:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_add_x_negone_y:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_add_x_negone_y:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_add_x_negone_y:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_add_x_negone_y:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_add_x_negone_y:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%a = fadd <4 x float> %x, <float -1.0, float -1.0, float -1.0, float -1.0>
%m = fmul <4 x float> %a, %y
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_add_x_negone(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_add_x_negone:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_add_x_negone:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_add_x_negone:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_add_x_negone:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_add_x_negone:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_add_x_negone:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%a = fadd <4 x float> %x, <float -1.0, float -1.0, float -1.0, float -1.0>
%m = fmul <4 x float> %y, %a
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_add_x_negone_undefs(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_add_x_negone_undefs:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_add_x_negone_undefs:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_add_x_negone_undefs:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_add_x_negone_undefs:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_add_x_negone_undefs:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_add_x_negone_undefs:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%a = fadd <4 x float> %x, <float undef, float -1.0, float undef, float -1.0>
%m = fmul <4 x float> %y, %a
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_sub_one_x_y(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_sub_one_x_y:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_sub_one_x_y:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA4-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_sub_one_x_y:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} xmm2 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; AVX512-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_sub_one_x_y:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_sub_one_x_y:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfnmaddps {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_sub_one_x_y:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> <float 1.0, float 1.0, float 1.0, float 1.0>, %x
%m = fmul <4 x float> %s, %y
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_one_x(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_one_x:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_one_x:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA4-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_one_x:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} xmm2 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; AVX512-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_one_x:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_one_x:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfnmaddps {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_one_x:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> <float 1.0, float 1.0, float 1.0, float 1.0>, %x
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_one_x_undefs(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_one_x_undefs:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} xmm2 = <1.0E+0,u,1.0E+0,1.0E+0>
; FMA-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_one_x_undefs:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} xmm2 = <1.0E+0,u,1.0E+0,1.0E+0>
; FMA4-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_one_x_undefs:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} xmm2 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; AVX512-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_one_x_undefs:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_one_x_undefs:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfnmaddps {{.*#+}} xmm0 = -(xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_one_x_undefs:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfnmadd213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> <float 1.0, float undef, float 1.0, float 1.0>, %x
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_sub_negone_x_y(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_sub_negone_x_y:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [-1.0E+0,-1.0E+0,-1.0E+0,-1.0E+0]
; FMA-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_sub_negone_x_y:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [-1.0E+0,-1.0E+0,-1.0E+0,-1.0E+0]
; FMA4-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_sub_negone_x_y:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} xmm2 = [-1.0E+0,-1.0E+0,-1.0E+0,-1.0E+0]
; AVX512-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_sub_negone_x_y:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_sub_negone_x_y:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_sub_negone_x_y:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> <float -1.0, float -1.0, float -1.0, float -1.0>, %x
%m = fmul <4 x float> %s, %y
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_negone_x(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_negone_x:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [-1.0E+0,-1.0E+0,-1.0E+0,-1.0E+0]
; FMA-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_negone_x:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} xmm2 = [-1.0E+0,-1.0E+0,-1.0E+0,-1.0E+0]
; FMA4-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_negone_x:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} xmm2 = [-1.0E+0,-1.0E+0,-1.0E+0,-1.0E+0]
; AVX512-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_negone_x:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_negone_x:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_negone_x:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> <float -1.0, float -1.0, float -1.0, float -1.0>, %x
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_negone_x_undefs(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_negone_x_undefs:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} xmm2 = <-1.0E+0,-1.0E+0,u,-1.0E+0>
; FMA-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_negone_x_undefs:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} xmm2 = <-1.0E+0,-1.0E+0,u,-1.0E+0>
; FMA4-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_negone_x_undefs:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} xmm2 = [-1.0E+0,-1.0E+0,-1.0E+0,-1.0E+0]
; AVX512-INFS-NEXT: vsubps %xmm0, %xmm2, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_negone_x_undefs:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_negone_x_undefs:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_negone_x_undefs:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> <float -1.0, float -1.0, float undef, float -1.0>, %x
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_sub_x_one_y(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_sub_x_one_y:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_sub_x_one_y:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_sub_x_one_y:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_sub_x_one_y:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_sub_x_one_y:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_sub_x_one_y:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> %x, <float 1.0, float 1.0, float 1.0, float 1.0>
%m = fmul <4 x float> %s, %y
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_x_one(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_x_one:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_x_one:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_x_one:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_x_one:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_x_one:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_x_one:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> %x, <float 1.0, float 1.0, float 1.0, float 1.0>
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_x_one_undefs(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_x_one_undefs:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_x_one_undefs:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_x_one_undefs:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_x_one_undefs:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_x_one_undefs:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_x_one_undefs:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> %x, <float 1.0, float 1.0, float 1.0, float undef>
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_sub_x_negone_y(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_sub_x_negone_y:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_sub_x_negone_y:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_sub_x_negone_y:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm1, %xmm0, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_sub_x_negone_y:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_sub_x_negone_y:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_sub_x_negone_y:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> %x, <float -1.0, float -1.0, float -1.0, float -1.0>
%m = fmul <4 x float> %s, %y
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_x_negone(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_x_negone:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_x_negone:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_x_negone:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_x_negone:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_x_negone:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_x_negone:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> %x, <float -1.0, float -1.0, float -1.0, float -1.0>
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
define <4 x float> @test_v4f32_mul_y_sub_x_negone_undefs(<4 x float> %x, <4 x float> %y) {
; FMA-INFS-LABEL: test_v4f32_mul_y_sub_x_negone_undefs:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_mul_y_sub_x_negone_undefs:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_mul_y_sub_x_negone_undefs:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vaddps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-INFS-NEXT: vmulps %xmm0, %xmm1, %xmm0
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_mul_y_sub_x_negone_undefs:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_mul_y_sub_x_negone_undefs:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_mul_y_sub_x_negone_undefs:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm1
; AVX512-NOINFS-NEXT: retq
%s = fsub <4 x float> %x, <float undef, float -1.0, float -1.0, float -1.0>
%m = fmul <4 x float> %y, %s
ret <4 x float> %m
}
;
; Interpolation Patterns: add(mul(x,t),mul(sub(1.0,t),y))
;
define float @test_f32_interp(float %x, float %y, float %t) {
; FMA-INFS-LABEL: test_f32_interp:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovss {{.*#+}} xmm3 = mem[0],zero,zero,zero
; FMA-INFS-NEXT: vsubss %xmm2, %xmm3, %xmm3
; FMA-INFS-NEXT: vmulss %xmm3, %xmm1, %xmm1
; FMA-INFS-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_f32_interp:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovss {{.*#+}} xmm3 = mem[0],zero,zero,zero
; FMA4-INFS-NEXT: vsubss %xmm2, %xmm3, %xmm3
; FMA4-INFS-NEXT: vmulss %xmm3, %xmm1, %xmm1
; FMA4-INFS-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * xmm2) + xmm1
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_f32_interp:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vmovss {{.*#+}} xmm3 = mem[0],zero,zero,zero
; AVX512-INFS-NEXT: vsubss %xmm2, %xmm3, %xmm3
; AVX512-INFS-NEXT: vmulss %xmm3, %xmm1, %xmm1
; AVX512-INFS-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_f32_interp:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ss {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA-NOINFS-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_f32_interp:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubss {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: vfmsubss {{.*#+}} xmm0 = (xmm0 * xmm2) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_f32_interp:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ss {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; AVX512-NOINFS-NEXT: vfmsub213ss {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%t1 = fsub nsz float 1.0, %t
%tx = fmul nsz float %x, %t
%ty = fmul nsz float %y, %t1
%r = fadd nsz float %tx, %ty
ret float %r
}
define <4 x float> @test_v4f32_interp(<4 x float> %x, <4 x float> %y, <4 x float> %t) {
; FMA-INFS-LABEL: test_v4f32_interp:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} xmm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA-INFS-NEXT: vsubps %xmm2, %xmm3, %xmm3
; FMA-INFS-NEXT: vmulps %xmm3, %xmm1, %xmm1
; FMA-INFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f32_interp:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} xmm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA4-INFS-NEXT: vsubps %xmm2, %xmm3, %xmm3
; FMA4-INFS-NEXT: vmulps %xmm3, %xmm1, %xmm1
; FMA4-INFS-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm2) + xmm1
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f32_interp:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} xmm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; AVX512-INFS-NEXT: vsubps %xmm2, %xmm3, %xmm3
; AVX512-INFS-NEXT: vmulps %xmm3, %xmm1, %xmm1
; AVX512-INFS-NEXT: vfmadd213ps {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f32_interp:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f32_interp:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm2) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f32_interp:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%t1 = fsub nsz <4 x float> <float 1.0, float 1.0, float 1.0, float 1.0>, %t
%tx = fmul nsz <4 x float> %x, %t
%ty = fmul nsz <4 x float> %y, %t1
%r = fadd nsz <4 x float> %tx, %ty
ret <4 x float> %r
}
define <8 x float> @test_v8f32_interp(<8 x float> %x, <8 x float> %y, <8 x float> %t) {
; FMA-INFS-LABEL: test_v8f32_interp:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovaps {{.*#+}} ymm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA-INFS-NEXT: vsubps %ymm2, %ymm3, %ymm3
; FMA-INFS-NEXT: vmulps %ymm3, %ymm1, %ymm1
; FMA-INFS-NEXT: vfmadd213ps {{.*#+}} ymm0 = (ymm2 * ymm0) + ymm1
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v8f32_interp:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovaps {{.*#+}} ymm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA4-INFS-NEXT: vsubps %ymm2, %ymm3, %ymm3
; FMA4-INFS-NEXT: vmulps %ymm3, %ymm1, %ymm1
; FMA4-INFS-NEXT: vfmaddps {{.*#+}} ymm0 = (ymm0 * ymm2) + ymm1
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v8f32_interp:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastss {{.*#+}} ymm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; AVX512-INFS-NEXT: vsubps %ymm2, %ymm3, %ymm3
; AVX512-INFS-NEXT: vmulps %ymm3, %ymm1, %ymm1
; AVX512-INFS-NEXT: vfmadd213ps {{.*#+}} ymm0 = (ymm2 * ymm0) + ymm1
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v8f32_interp:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} ymm1 = (ymm2 * ymm1) - ymm1
; FMA-NOINFS-NEXT: vfmsub213ps {{.*#+}} ymm0 = (ymm2 * ymm0) - ymm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v8f32_interp:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} ymm1 = (ymm2 * ymm1) - ymm1
; FMA4-NOINFS-NEXT: vfmsubps {{.*#+}} ymm0 = (ymm0 * ymm2) - ymm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v8f32_interp:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} ymm1 = (ymm2 * ymm1) - ymm1
; AVX512-NOINFS-NEXT: vfmsub213ps {{.*#+}} ymm0 = (ymm2 * ymm0) - ymm1
; AVX512-NOINFS-NEXT: retq
%t1 = fsub nsz <8 x float> <float 1.0, float 1.0, float 1.0, float 1.0, float 1.0, float 1.0, float 1.0, float 1.0>, %t
%tx = fmul nsz <8 x float> %x, %t
%ty = fmul nsz <8 x float> %y, %t1
%r = fadd nsz <8 x float> %tx, %ty
ret <8 x float> %r
}
define double @test_f64_interp(double %x, double %y, double %t) {
; FMA-INFS-LABEL: test_f64_interp:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovsd {{.*#+}} xmm3 = mem[0],zero
; FMA-INFS-NEXT: vsubsd %xmm2, %xmm3, %xmm3
; FMA-INFS-NEXT: vmulsd %xmm3, %xmm1, %xmm1
; FMA-INFS-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_f64_interp:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovsd {{.*#+}} xmm3 = mem[0],zero
; FMA4-INFS-NEXT: vsubsd %xmm2, %xmm3, %xmm3
; FMA4-INFS-NEXT: vmulsd %xmm3, %xmm1, %xmm1
; FMA4-INFS-NEXT: vfmaddsd {{.*#+}} xmm0 = (xmm0 * xmm2) + xmm1
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_f64_interp:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vmovsd {{.*#+}} xmm3 = mem[0],zero
; AVX512-INFS-NEXT: vsubsd %xmm2, %xmm3, %xmm3
; AVX512-INFS-NEXT: vmulsd %xmm3, %xmm1, %xmm1
; AVX512-INFS-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_f64_interp:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213sd {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA-NOINFS-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_f64_interp:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubsd {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: vfmsubsd {{.*#+}} xmm0 = (xmm0 * xmm2) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_f64_interp:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213sd {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; AVX512-NOINFS-NEXT: vfmsub213sd {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%t1 = fsub nsz double 1.0, %t
%tx = fmul nsz double %x, %t
%ty = fmul nsz double %y, %t1
%r = fadd nsz double %tx, %ty
ret double %r
}
define <2 x double> @test_v2f64_interp(<2 x double> %x, <2 x double> %y, <2 x double> %t) {
; FMA-INFS-LABEL: test_v2f64_interp:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovapd {{.*#+}} xmm3 = [1.0E+0,1.0E+0]
; FMA-INFS-NEXT: vsubpd %xmm2, %xmm3, %xmm3
; FMA-INFS-NEXT: vmulpd %xmm3, %xmm1, %xmm1
; FMA-INFS-NEXT: vfmadd213pd {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v2f64_interp:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovapd {{.*#+}} xmm3 = [1.0E+0,1.0E+0]
; FMA4-INFS-NEXT: vsubpd %xmm2, %xmm3, %xmm3
; FMA4-INFS-NEXT: vmulpd %xmm3, %xmm1, %xmm1
; FMA4-INFS-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm0 * xmm2) + xmm1
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v2f64_interp:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vmovddup {{.*#+}} xmm3 = [1.0E+0,1.0E+0]
; AVX512-INFS-NEXT: # xmm3 = mem[0,0]
; AVX512-INFS-NEXT: vsubpd %xmm2, %xmm3, %xmm3
; AVX512-INFS-NEXT: vmulpd %xmm3, %xmm1, %xmm1
; AVX512-INFS-NEXT: vfmadd213pd {{.*#+}} xmm0 = (xmm2 * xmm0) + xmm1
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v2f64_interp:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213pd {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA-NOINFS-NEXT: vfmsub213pd {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v2f64_interp:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubpd {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; FMA4-NOINFS-NEXT: vfmsubpd {{.*#+}} xmm0 = (xmm0 * xmm2) - xmm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v2f64_interp:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213pd {{.*#+}} xmm1 = (xmm2 * xmm1) - xmm1
; AVX512-NOINFS-NEXT: vfmsub213pd {{.*#+}} xmm0 = (xmm2 * xmm0) - xmm1
; AVX512-NOINFS-NEXT: retq
%t1 = fsub nsz <2 x double> <double 1.0, double 1.0>, %t
%tx = fmul nsz <2 x double> %x, %t
%ty = fmul nsz <2 x double> %y, %t1
%r = fadd nsz <2 x double> %tx, %ty
ret <2 x double> %r
}
define <4 x double> @test_v4f64_interp(<4 x double> %x, <4 x double> %y, <4 x double> %t) {
; FMA-INFS-LABEL: test_v4f64_interp:
; FMA-INFS: # %bb.0:
; FMA-INFS-NEXT: vmovapd {{.*#+}} ymm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA-INFS-NEXT: vsubpd %ymm2, %ymm3, %ymm3
; FMA-INFS-NEXT: vmulpd %ymm3, %ymm1, %ymm1
; FMA-INFS-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm2 * ymm0) + ymm1
; FMA-INFS-NEXT: retq
;
; FMA4-INFS-LABEL: test_v4f64_interp:
; FMA4-INFS: # %bb.0:
; FMA4-INFS-NEXT: vmovapd {{.*#+}} ymm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; FMA4-INFS-NEXT: vsubpd %ymm2, %ymm3, %ymm3
; FMA4-INFS-NEXT: vmulpd %ymm3, %ymm1, %ymm1
; FMA4-INFS-NEXT: vfmaddpd {{.*#+}} ymm0 = (ymm0 * ymm2) + ymm1
; FMA4-INFS-NEXT: retq
;
; AVX512-INFS-LABEL: test_v4f64_interp:
; AVX512-INFS: # %bb.0:
; AVX512-INFS-NEXT: vbroadcastsd {{.*#+}} ymm3 = [1.0E+0,1.0E+0,1.0E+0,1.0E+0]
; AVX512-INFS-NEXT: vsubpd %ymm2, %ymm3, %ymm3
; AVX512-INFS-NEXT: vmulpd %ymm3, %ymm1, %ymm1
; AVX512-INFS-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm2 * ymm0) + ymm1
; AVX512-INFS-NEXT: retq
;
; FMA-NOINFS-LABEL: test_v4f64_interp:
; FMA-NOINFS: # %bb.0:
; FMA-NOINFS-NEXT: vfmsub213pd {{.*#+}} ymm1 = (ymm2 * ymm1) - ymm1
; FMA-NOINFS-NEXT: vfmsub213pd {{.*#+}} ymm0 = (ymm2 * ymm0) - ymm1
; FMA-NOINFS-NEXT: retq
;
; FMA4-NOINFS-LABEL: test_v4f64_interp:
; FMA4-NOINFS: # %bb.0:
; FMA4-NOINFS-NEXT: vfmsubpd {{.*#+}} ymm1 = (ymm2 * ymm1) - ymm1
; FMA4-NOINFS-NEXT: vfmsubpd {{.*#+}} ymm0 = (ymm0 * ymm2) - ymm1
; FMA4-NOINFS-NEXT: retq
;
; AVX512-NOINFS-LABEL: test_v4f64_interp:
; AVX512-NOINFS: # %bb.0:
; AVX512-NOINFS-NEXT: vfmsub213pd {{.*#+}} ymm1 = (ymm2 * ymm1) - ymm1
; AVX512-NOINFS-NEXT: vfmsub213pd {{.*#+}} ymm0 = (ymm2 * ymm0) - ymm1
; AVX512-NOINFS-NEXT: retq
%t1 = fsub nsz <4 x double> <double 1.0, double 1.0, double 1.0, double 1.0>, %t
%tx = fmul nsz <4 x double> %x, %t
%ty = fmul nsz <4 x double> %y, %t1
%r = fadd nsz <4 x double> %tx, %ty
ret <4 x double> %r
}
;
; Pattern: (fneg (fma x, y, z)) -> (fma x, -y, -z)
;
define <4 x float> @test_v4f32_fneg_fmadd(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; FMA-LABEL: test_v4f32_fneg_fmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f32_fneg_fmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f32_fneg_fmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%mul = fmul nsz <4 x float> %a0, %a1
%add = fadd nsz <4 x float> %mul, %a2
%neg = fsub nsz <4 x float> <float -0.0, float -0.0, float -0.0, float -0.0>, %add
ret <4 x float> %neg
}
define <4 x double> @test_v4f64_fneg_fmsub(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; FMA-LABEL: test_v4f64_fneg_fmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfnmadd213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f64_fneg_fmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfnmaddpd {{.*#+}} ymm0 = -(ymm0 * ymm1) + ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f64_fneg_fmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfnmadd213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) + ymm2
; AVX512-NEXT: retq
%mul = fmul nsz <4 x double> %a0, %a1
%sub = fsub nsz <4 x double> %mul, %a2
%neg = fsub nsz <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %sub
ret <4 x double> %neg
}
define <4 x float> @test_v4f32_fneg_fnmadd(<4 x float> %a0, <4 x float> %a1, <4 x float> %a2) #0 {
; FMA-LABEL: test_v4f32_fneg_fnmadd:
; FMA: # %bb.0:
; FMA-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f32_fneg_fnmadd:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f32_fneg_fnmadd:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmsub213ps {{.*#+}} xmm0 = (xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%mul = fmul nsz <4 x float> %a0, %a1
%neg0 = fsub nsz <4 x float> <float -0.0, float -0.0, float -0.0, float -0.0>, %mul
%add = fadd nsz <4 x float> %neg0, %a2
%neg1 = fsub nsz <4 x float> <float -0.0, float -0.0, float -0.0, float -0.0>, %add
ret <4 x float> %neg1
}
define <4 x double> @test_v4f64_fneg_fnmsub(<4 x double> %a0, <4 x double> %a1, <4 x double> %a2) #0 {
; FMA-LABEL: test_v4f64_fneg_fnmsub:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f64_fneg_fnmsub:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddpd {{.*#+}} ymm0 = (ymm0 * ymm1) + ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f64_fneg_fnmsub:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213pd {{.*#+}} ymm0 = (ymm1 * ymm0) + ymm2
; AVX512-NEXT: retq
%mul = fmul nsz <4 x double> %a0, %a1
%neg0 = fsub nsz <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %mul
%sub = fsub nsz <4 x double> %neg0, %a2
%neg1 = fsub nsz <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %sub
ret <4 x double> %neg1
}
;
; Pattern: (fma x, c1, (fmul x, c2)) -> (fmul x, c1+c2)
;
define <4 x float> @test_v4f32_fma_x_c1_fmul_x_c2(<4 x float> %x) #0 {
; FMA-LABEL: test_v4f32_fma_x_c1_fmul_x_c2:
; FMA: # %bb.0:
; FMA-NEXT: vmulps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f32_fma_x_c1_fmul_x_c2:
; FMA4: # %bb.0:
; FMA4-NEXT: vmulps {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %xmm0, %xmm0
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f32_fma_x_c1_fmul_x_c2:
; AVX512: # %bb.0:
; AVX512-NEXT: vmulps {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %xmm0, %xmm0
; AVX512-NEXT: retq
%m0 = fmul <4 x float> %x, <float 1.0, float 2.0, float 3.0, float 4.0>
%m1 = fmul <4 x float> %x, <float 4.0, float 3.0, float 2.0, float 1.0>
%a = fadd <4 x float> %m0, %m1
ret <4 x float> %a
}
;
; Pattern: (fma (fmul x, c1), c2, y) -> (fma x, c1*c2, y)
;
define <4 x float> @test_v4f32_fma_fmul_x_c1_c2_y(<4 x float> %x, <4 x float> %y) #0 {
; FMA-LABEL: test_v4f32_fma_fmul_x_c1_c2_y:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd132ps {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f32_fma_fmul_x_c1_c2_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f32_fma_fmul_x_c1_c2_y:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd132ps {{.*#+}} xmm0 = (xmm0 * mem) + xmm1
; AVX512-NEXT: retq
%m0 = fmul <4 x float> %x, <float 1.0, float 2.0, float 3.0, float 4.0>
%m1 = fmul <4 x float> %m0, <float 4.0, float 3.0, float 2.0, float 1.0>
%a = fadd <4 x float> %m1, %y
ret <4 x float> %a
}
; Pattern: (fneg (fmul x, y)) -> (fnmsub x, y, 0)
define double @test_f64_fneg_fmul(double %x, double %y) #0 {
; FMA-LABEL: test_f64_fneg_fmul:
; FMA: # %bb.0:
; FMA-NEXT: vxorpd %xmm2, %xmm2, %xmm2
; FMA-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_f64_fneg_fmul:
; FMA4: # %bb.0:
; FMA4-NEXT: vxorpd %xmm2, %xmm2, %xmm2
; FMA4-NEXT: vfnmsubsd {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_f64_fneg_fmul:
; AVX512: # %bb.0:
; AVX512-NEXT: vxorpd %xmm2, %xmm2, %xmm2
; AVX512-NEXT: vfnmsub213sd {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%m = fmul nsz double %x, %y
%n = fsub double -0.0, %m
ret double %n
}
define <4 x float> @test_v4f32_fneg_fmul(<4 x float> %x, <4 x float> %y) #0 {
; FMA-LABEL: test_v4f32_fneg_fmul:
; FMA: # %bb.0:
; FMA-NEXT: vxorps %xmm2, %xmm2, %xmm2
; FMA-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f32_fneg_fmul:
; FMA4: # %bb.0:
; FMA4-NEXT: vxorps %xmm2, %xmm2, %xmm2
; FMA4-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * xmm1) - xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f32_fneg_fmul:
; AVX512: # %bb.0:
; AVX512-NEXT: vxorps %xmm2, %xmm2, %xmm2
; AVX512-NEXT: vfnmsub213ps {{.*#+}} xmm0 = -(xmm1 * xmm0) - xmm2
; AVX512-NEXT: retq
%m = fmul nsz <4 x float> %x, %y
%n = fsub <4 x float> <float -0.0, float -0.0, float -0.0, float -0.0>, %m
ret <4 x float> %n
}
define <4 x double> @test_v4f64_fneg_fmul(<4 x double> %x, <4 x double> %y) #0 {
; FMA-LABEL: test_v4f64_fneg_fmul:
; FMA: # %bb.0:
; FMA-NEXT: vxorpd %xmm2, %xmm2, %xmm2
; FMA-NEXT: vfnmsub213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f64_fneg_fmul:
; FMA4: # %bb.0:
; FMA4-NEXT: vxorpd %xmm2, %xmm2, %xmm2
; FMA4-NEXT: vfnmsubpd {{.*#+}} ymm0 = -(ymm0 * ymm1) - ymm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f64_fneg_fmul:
; AVX512: # %bb.0:
; AVX512-NEXT: vxorpd %xmm2, %xmm2, %xmm2
; AVX512-NEXT: vfnmsub213pd {{.*#+}} ymm0 = -(ymm1 * ymm0) - ymm2
; AVX512-NEXT: retq
%m = fmul nsz <4 x double> %x, %y
%n = fsub <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %m
ret <4 x double> %n
}
define <4 x double> @test_v4f64_fneg_fmul_no_nsz(<4 x double> %x, <4 x double> %y) #0 {
; FMA-LABEL: test_v4f64_fneg_fmul_no_nsz:
; FMA: # %bb.0:
; FMA-NEXT: vmulpd %ymm1, %ymm0, %ymm0
; FMA-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; FMA-NEXT: retq
;
; FMA4-LABEL: test_v4f64_fneg_fmul_no_nsz:
; FMA4: # %bb.0:
; FMA4-NEXT: vmulpd %ymm1, %ymm0, %ymm0
; FMA4-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip), %ymm0, %ymm0
; FMA4-NEXT: retq
;
; AVX512-LABEL: test_v4f64_fneg_fmul_no_nsz:
; AVX512: # %bb.0:
; AVX512-NEXT: vmulpd %ymm1, %ymm0, %ymm0
; AVX512-NEXT: vxorpd {{\.?LCPI[0-9]+_[0-9]+}}(%rip){1to4}, %ymm0, %ymm0
; AVX512-NEXT: retq
%m = fmul <4 x double> %x, %y
%n = fsub <4 x double> <double -0.0, double -0.0, double -0.0, double -0.0>, %m
ret <4 x double> %n
}
; ((a*b) + (c*d)) + n1 --> (a*b) + ((c*d) + n1)
define double @fadd_fma_fmul_1(double %a, double %b, double %c, double %d, double %n1) nounwind {
; FMA-LABEL: fadd_fma_fmul_1:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213sd {{.*#+}} xmm2 = (xmm3 * xmm2) + xmm4
; FMA-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: fadd_fma_fmul_1:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddsd {{.*#+}} xmm2 = (xmm2 * xmm3) + xmm4
; FMA4-NEXT: vfmaddsd {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: fadd_fma_fmul_1:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213sd {{.*#+}} xmm2 = (xmm3 * xmm2) + xmm4
; AVX512-NEXT: vfmadd213sd {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%m1 = fmul fast double %a, %b
%m2 = fmul fast double %c, %d
%a1 = fadd fast double %m1, %m2
%a2 = fadd fast double %a1, %n1
ret double %a2
}
; Minimum FMF - the 1st fadd is contracted because that combines
; fmul+fadd as specified by the order of operations; the 2nd fadd
; requires reassociation to fuse with c*d.
define float @fadd_fma_fmul_fmf(float %a, float %b, float %c, float %d, float %n0) nounwind {
; FMA-LABEL: fadd_fma_fmul_fmf:
; FMA: # %bb.0:
; FMA-NEXT: vfmadd213ss {{.*#+}} xmm2 = (xmm3 * xmm2) + xmm4
; FMA-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: retq
;
; FMA4-LABEL: fadd_fma_fmul_fmf:
; FMA4: # %bb.0:
; FMA4-NEXT: vfmaddss {{.*#+}} xmm2 = (xmm2 * xmm3) + xmm4
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: retq
;
; AVX512-LABEL: fadd_fma_fmul_fmf:
; AVX512: # %bb.0:
; AVX512-NEXT: vfmadd213ss {{.*#+}} xmm2 = (xmm3 * xmm2) + xmm4
; AVX512-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: retq
%m1 = fmul float %a, %b
%m2 = fmul float %c, %d
%a1 = fadd contract float %m1, %m2
%a2 = fadd reassoc float %n0, %a1
ret float %a2
}
; Not minimum FMF.
define float @fadd_fma_fmul_2(float %a, float %b, float %c, float %d, float %n0) nounwind {
; FMA-LABEL: fadd_fma_fmul_2:
; FMA: # %bb.0:
; FMA-NEXT: vmulss %xmm3, %xmm2, %xmm2
; FMA-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: vaddss %xmm2, %xmm4, %xmm0
; FMA-NEXT: retq
;
; FMA4-LABEL: fadd_fma_fmul_2:
; FMA4: # %bb.0:
; FMA4-NEXT: vmulss %xmm3, %xmm2, %xmm2
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: vaddss %xmm0, %xmm4, %xmm0
; FMA4-NEXT: retq
;
; AVX512-LABEL: fadd_fma_fmul_2:
; AVX512: # %bb.0:
; AVX512-NEXT: vmulss %xmm3, %xmm2, %xmm2
; AVX512-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: vaddss %xmm2, %xmm4, %xmm0
; AVX512-NEXT: retq
%m1 = fmul float %a, %b
%m2 = fmul float %c, %d
%a1 = fadd contract float %m1, %m2
%a2 = fadd contract float %n0, %a1
ret float %a2
}
; The final fadd can be folded with either 1 of the leading fmuls.
define <2 x double> @fadd_fma_fmul_3(<2 x double> %x1, <2 x double> %x2, <2 x double> %x3, <2 x double> %x4, <2 x double> %x5, <2 x double> %x6, <2 x double> %x7, <2 x double> %x8) nounwind {
; FMA-LABEL: fadd_fma_fmul_3:
; FMA: # %bb.0:
; FMA-NEXT: vmulpd %xmm3, %xmm2, %xmm2
; FMA-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm7 * xmm6) + xmm2
; FMA-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm5 * xmm4) + xmm2
; FMA-NEXT: vmovapd %xmm2, %xmm0
; FMA-NEXT: retq
;
; FMA4-LABEL: fadd_fma_fmul_3:
; FMA4: # %bb.0:
; FMA4-NEXT: vmulpd %xmm3, %xmm2, %xmm2
; FMA4-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm6 * xmm7) + xmm0
; FMA4-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm4 * xmm5) + xmm0
; FMA4-NEXT: retq
;
; AVX512-LABEL: fadd_fma_fmul_3:
; AVX512: # %bb.0:
; AVX512-NEXT: vmulpd %xmm3, %xmm2, %xmm2
; AVX512-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm7 * xmm6) + xmm2
; AVX512-NEXT: vfmadd231pd {{.*#+}} xmm2 = (xmm5 * xmm4) + xmm2
; AVX512-NEXT: vmovapd %xmm2, %xmm0
; AVX512-NEXT: retq
%m1 = fmul fast <2 x double> %x1, %x2
%m2 = fmul fast <2 x double> %x3, %x4
%m3 = fmul fast <2 x double> %x5, %x6
%m4 = fmul fast <2 x double> %x7, %x8
%a1 = fadd fast <2 x double> %m1, %m2
%a2 = fadd fast <2 x double> %m3, %m4
%a3 = fadd fast <2 x double> %a1, %a2
ret <2 x double> %a3
}
; negative test
define float @fadd_fma_fmul_extra_use_1(float %a, float %b, float %c, float %d, float %n0, float* %p) nounwind {
; FMA-LABEL: fadd_fma_fmul_extra_use_1:
; FMA: # %bb.0:
; FMA-NEXT: vmulss %xmm1, %xmm0, %xmm0
; FMA-NEXT: vmovss %xmm0, (%rdi)
; FMA-NEXT: vfmadd213ss {{.*#+}} xmm2 = (xmm3 * xmm2) + xmm0
; FMA-NEXT: vaddss %xmm2, %xmm4, %xmm0
; FMA-NEXT: retq
;
; FMA4-LABEL: fadd_fma_fmul_extra_use_1:
; FMA4: # %bb.0:
; FMA4-NEXT: vmulss %xmm1, %xmm0, %xmm0
; FMA4-NEXT: vmovss %xmm0, (%rdi)
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm2 * xmm3) + xmm0
; FMA4-NEXT: vaddss %xmm0, %xmm4, %xmm0
; FMA4-NEXT: retq
;
; AVX512-LABEL: fadd_fma_fmul_extra_use_1:
; AVX512: # %bb.0:
; AVX512-NEXT: vmulss %xmm1, %xmm0, %xmm0
; AVX512-NEXT: vmovss %xmm0, (%rdi)
; AVX512-NEXT: vfmadd213ss {{.*#+}} xmm2 = (xmm3 * xmm2) + xmm0
; AVX512-NEXT: vaddss %xmm2, %xmm4, %xmm0
; AVX512-NEXT: retq
%m1 = fmul fast float %a, %b
store float %m1, float* %p
%m2 = fmul fast float %c, %d
%a1 = fadd fast float %m1, %m2
%a2 = fadd fast float %n0, %a1
ret float %a2
}
; negative test
define float @fadd_fma_fmul_extra_use_2(float %a, float %b, float %c, float %d, float %n0, float* %p) nounwind {
; FMA-LABEL: fadd_fma_fmul_extra_use_2:
; FMA: # %bb.0:
; FMA-NEXT: vmulss %xmm3, %xmm2, %xmm2
; FMA-NEXT: vmovss %xmm2, (%rdi)
; FMA-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: vaddss %xmm0, %xmm4, %xmm0
; FMA-NEXT: retq
;
; FMA4-LABEL: fadd_fma_fmul_extra_use_2:
; FMA4: # %bb.0:
; FMA4-NEXT: vmulss %xmm3, %xmm2, %xmm2
; FMA4-NEXT: vmovss %xmm2, (%rdi)
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: vaddss %xmm0, %xmm4, %xmm0
; FMA4-NEXT: retq
;
; AVX512-LABEL: fadd_fma_fmul_extra_use_2:
; AVX512: # %bb.0:
; AVX512-NEXT: vmulss %xmm3, %xmm2, %xmm2
; AVX512-NEXT: vmovss %xmm2, (%rdi)
; AVX512-NEXT: vfmadd213ss {{.*#+}} xmm0 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: vaddss %xmm0, %xmm4, %xmm0
; AVX512-NEXT: retq
%m1 = fmul fast float %a, %b
%m2 = fmul fast float %c, %d
store float %m2, float* %p
%a1 = fadd fast float %m1, %m2
%a2 = fadd fast float %n0, %a1
ret float %a2
}
; negative test
define float @fadd_fma_fmul_extra_use_3(float %a, float %b, float %c, float %d, float %n0, float* %p) nounwind {
; FMA-LABEL: fadd_fma_fmul_extra_use_3:
; FMA: # %bb.0:
; FMA-NEXT: vmulss %xmm3, %xmm2, %xmm2
; FMA-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm1 * xmm0) + xmm2
; FMA-NEXT: vmovss %xmm2, (%rdi)
; FMA-NEXT: vaddss %xmm2, %xmm4, %xmm0
; FMA-NEXT: retq
;
; FMA4-LABEL: fadd_fma_fmul_extra_use_3:
; FMA4: # %bb.0:
; FMA4-NEXT: vmulss %xmm3, %xmm2, %xmm2
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * xmm1) + xmm2
; FMA4-NEXT: vmovss %xmm0, (%rdi)
; FMA4-NEXT: vaddss %xmm0, %xmm4, %xmm0
; FMA4-NEXT: retq
;
; AVX512-LABEL: fadd_fma_fmul_extra_use_3:
; AVX512: # %bb.0:
; AVX512-NEXT: vmulss %xmm3, %xmm2, %xmm2
; AVX512-NEXT: vfmadd231ss {{.*#+}} xmm2 = (xmm1 * xmm0) + xmm2
; AVX512-NEXT: vmovss %xmm2, (%rdi)
; AVX512-NEXT: vaddss %xmm2, %xmm4, %xmm0
; AVX512-NEXT: retq
%m1 = fmul fast float %a, %b
%m2 = fmul fast float %c, %d
%a1 = fadd fast float %m1, %m2
store float %a1, float* %p
%a2 = fadd fast float %n0, %a1
ret float %a2
}
attributes #0 = { "unsafe-fp-math"="true" }
Reference in New Issue View Git Blame Copy Permalink