RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-23 11:13:28 +01:00
Code Issues Projects Releases Wiki Activity
5272bea13d
llvm-mirror/test/CodeGen/X86/fma4-commute-x86.ll
Wang, Pengfei c275334e12 [CodeGen][X86] Remove unused trivial check-prefixes from all CodeGen/X86 directory. 
I had manually removed unused prefixes from CodeGen/X86 directory for more than 100 tests.
I checked the change history for each of them at the beginning, and then I mainly focused on the format since I found all of the unused prefixes were result from either insensible copy or residuum after functional update.
I think it's OK to remove the remaining X86 tests by script now. I wrote a rough script which works for me in most tests. I put it in llvm/utils temporarily for review and hope it may help other components owners.
The tests in this patch are all generated by the tool and checked by update tool for the autogenerated tests. I skimmed them and checked about 30 tests and didn't find any unexpected changes.

Reviewed By: mtrofin, MaskRay

Differential Revision: https://reviews.llvm.org/D91496
2020-11-16 09:45:55 +08:00

564 lines
23 KiB
LLVM
Raw Blame History

; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
; RUN: llc < %s -mcpu=bdver2 -mtriple=x86_64-pc-win32 | FileCheck %s --check-prefix=FMA4
attributes #0 = { nounwind }
declare <4 x float> @llvm.x86.fma4.vfmadd.ss(<4 x float>, <4 x float>, <4 x float>) nounwind readnone
define <4 x float> @test_x86_fmadd_baa_ss(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_baa_ss:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma4.vfmadd.ss(<4 x float> %b, <4 x float> %a, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fmadd_aba_ss(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_aba_ss:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma4.vfmadd.ss(<4 x float> %a, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fmadd_bba_ss(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_bba_ss:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovss {{.*#+}} xmm0 = mem[0],zero,zero,zero
; FMA4-NEXT: vfmaddss {{.*#+}} xmm0 = (xmm0 * xmm0) + mem
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma4.vfmadd.ss(<4 x float> %b, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
declare <4 x float> @llvm.x86.fma.vfmadd.ps(<4 x float>, <4 x float>, <4 x float>) nounwind readnone
define <4 x float> @test_x86_fmadd_baa_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_baa_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfmadd.ps(<4 x float> %b, <4 x float> %a, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fmadd_aba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_aba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfmadd.ps(<4 x float> %a, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fmadd_bba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_bba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfmaddps {{.*#+}} xmm0 = (xmm0 * xmm0) + mem
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfmadd.ps(<4 x float> %b, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfmadd.ps.256(<8 x float>, <8 x float>, <8 x float>) nounwind readnone
define <8 x float> @test_x86_fmadd_baa_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_baa_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmaddps {{.*#+}} ymm0 = (ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfmadd.ps.256(<8 x float> %b, <8 x float> %a, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fmadd_aba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_aba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmaddps {{.*#+}} ymm0 = (ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfmadd.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fmadd_bba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_bba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfmaddps {{.*#+}} ymm0 = (ymm0 * ymm0) + mem
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfmadd.ps.256(<8 x float> %b, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
declare <2 x double> @llvm.x86.fma4.vfmadd.sd(<2 x double>, <2 x double>, <2 x double>) nounwind readnone
define <2 x double> @test_x86_fmadd_baa_sd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_baa_sd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero
; FMA4-NEXT: vfmaddsd {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma4.vfmadd.sd(<2 x double> %b, <2 x double> %a, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fmadd_aba_sd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_aba_sd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero
; FMA4-NEXT: vfmaddsd {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma4.vfmadd.sd(<2 x double> %a, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fmadd_bba_sd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_bba_sd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovsd {{.*#+}} xmm0 = mem[0],zero
; FMA4-NEXT: vfmaddsd {{.*#+}} xmm0 = (xmm0 * xmm0) + mem
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma4.vfmadd.sd(<2 x double> %b, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
declare <2 x double> @llvm.x86.fma.vfmadd.pd(<2 x double>, <2 x double>, <2 x double>) nounwind readnone
define <2 x double> @test_x86_fmadd_baa_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_baa_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfmadd.pd(<2 x double> %b, <2 x double> %a, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fmadd_aba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_aba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfmadd.pd(<2 x double> %a, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fmadd_bba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_bba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfmaddpd {{.*#+}} xmm0 = (xmm0 * xmm0) + mem
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfmadd.pd(<2 x double> %b, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfmadd.pd.256(<4 x double>, <4 x double>, <4 x double>) nounwind readnone
define <4 x double> @test_x86_fmadd_baa_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_baa_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmaddpd {{.*#+}} ymm0 = (ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfmadd.pd.256(<4 x double> %b, <4 x double> %a, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fmadd_aba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_aba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmaddpd {{.*#+}} ymm0 = (ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfmadd.pd.256(<4 x double> %a, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fmadd_bba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmadd_bba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfmaddpd {{.*#+}} ymm0 = (ymm0 * ymm0) + mem
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfmadd.pd.256(<4 x double> %b, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
declare <4 x float> @llvm.x86.fma.vfnmadd.ps(<4 x float>, <4 x float>, <4 x float>) nounwind readnone
define <4 x float> @test_x86_fnmadd_baa_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_baa_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmaddps {{.*#+}} xmm0 = -(xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfnmadd.ps(<4 x float> %b, <4 x float> %a, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fnmadd_aba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_aba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmaddps {{.*#+}} xmm0 = -(xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfnmadd.ps(<4 x float> %a, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fnmadd_bba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_bba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfnmaddps {{.*#+}} xmm0 = -(xmm0 * xmm0) + mem
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfnmadd.ps(<4 x float> %b, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfnmadd.ps.256(<8 x float>, <8 x float>, <8 x float>) nounwind readnone
define <8 x float> @test_x86_fnmadd_baa_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_baa_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmaddps {{.*#+}} ymm0 = -(ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfnmadd.ps.256(<8 x float> %b, <8 x float> %a, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fnmadd_aba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_aba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmaddps {{.*#+}} ymm0 = -(ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfnmadd.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fnmadd_bba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_bba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfnmaddps {{.*#+}} ymm0 = -(ymm0 * ymm0) + mem
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfnmadd.ps.256(<8 x float> %b, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
declare <2 x double> @llvm.x86.fma.vfnmadd.pd(<2 x double>, <2 x double>, <2 x double>) nounwind readnone
define <2 x double> @test_x86_fnmadd_baa_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_baa_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmaddpd {{.*#+}} xmm0 = -(xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfnmadd.pd(<2 x double> %b, <2 x double> %a, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fnmadd_aba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_aba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmaddpd {{.*#+}} xmm0 = -(xmm0 * mem) + xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfnmadd.pd(<2 x double> %a, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fnmadd_bba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_bba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfnmaddpd {{.*#+}} xmm0 = -(xmm0 * xmm0) + mem
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfnmadd.pd(<2 x double> %b, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfnmadd.pd.256(<4 x double>, <4 x double>, <4 x double>) nounwind readnone
define <4 x double> @test_x86_fnmadd_baa_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_baa_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmaddpd {{.*#+}} ymm0 = -(ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfnmadd.pd.256(<4 x double> %b, <4 x double> %a, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fnmadd_aba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_aba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmaddpd {{.*#+}} ymm0 = -(ymm0 * mem) + ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfnmadd.pd.256(<4 x double> %a, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fnmadd_bba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmadd_bba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfnmaddpd {{.*#+}} ymm0 = -(ymm0 * ymm0) + mem
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfnmadd.pd.256(<4 x double> %b, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
declare <4 x float> @llvm.x86.fma.vfmsub.ps(<4 x float>, <4 x float>, <4 x float>) nounwind readnone
define <4 x float> @test_x86_fmsub_baa_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_baa_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfmsub.ps(<4 x float> %b, <4 x float> %a, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fmsub_aba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_aba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfmsub.ps(<4 x float> %a, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fmsub_bba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_bba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfmsubps {{.*#+}} xmm0 = (xmm0 * xmm0) - mem
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfmsub.ps(<4 x float> %b, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfmsub.ps.256(<8 x float>, <8 x float>, <8 x float>) nounwind readnone
define <8 x float> @test_x86_fmsub_baa_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_baa_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmsubps {{.*#+}} ymm0 = (ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfmsub.ps.256(<8 x float> %b, <8 x float> %a, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fmsub_aba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_aba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmsubps {{.*#+}} ymm0 = (ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfmsub.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fmsub_bba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_bba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfmsubps {{.*#+}} ymm0 = (ymm0 * ymm0) - mem
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfmsub.ps.256(<8 x float> %b, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
declare <2 x double> @llvm.x86.fma.vfmsub.pd(<2 x double>, <2 x double>, <2 x double>) nounwind readnone
define <2 x double> @test_x86_fmsub_baa_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_baa_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmsubpd {{.*#+}} xmm0 = (xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfmsub.pd(<2 x double> %b, <2 x double> %a, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fmsub_aba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_aba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmsubpd {{.*#+}} xmm0 = (xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfmsub.pd(<2 x double> %a, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fmsub_bba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_bba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfmsubpd {{.*#+}} xmm0 = (xmm0 * xmm0) - mem
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfmsub.pd(<2 x double> %b, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfmsub.pd.256(<4 x double>, <4 x double>, <4 x double>) nounwind readnone
define <4 x double> @test_x86_fmsub_baa_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_baa_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmsubpd {{.*#+}} ymm0 = (ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfmsub.pd.256(<4 x double> %b, <4 x double> %a, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fmsub_aba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_aba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmsubpd {{.*#+}} ymm0 = (ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfmsub.pd.256(<4 x double> %a, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fmsub_bba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fmsub_bba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfmsubpd {{.*#+}} ymm0 = (ymm0 * ymm0) - mem
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfmsub.pd.256(<4 x double> %b, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
declare <4 x float> @llvm.x86.fma.vfnmsub.ps(<4 x float>, <4 x float>, <4 x float>) nounwind readnone
define <4 x float> @test_x86_fnmsub_baa_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_baa_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfnmsub.ps(<4 x float> %b, <4 x float> %a, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fnmsub_aba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_aba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfnmsub.ps(<4 x float> %a, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
define <4 x float> @test_x86_fnmsub_bba_ps(<4 x float> %a, <4 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_bba_ps:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfnmsubps {{.*#+}} xmm0 = -(xmm0 * xmm0) - mem
; FMA4-NEXT: retq
%res = call <4 x float> @llvm.x86.fma.vfnmsub.ps(<4 x float> %b, <4 x float> %b, <4 x float> %a) nounwind
ret <4 x float> %res
}
declare <8 x float> @llvm.x86.fma.vfnmsub.ps.256(<8 x float>, <8 x float>, <8 x float>) nounwind readnone
define <8 x float> @test_x86_fnmsub_baa_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_baa_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmsubps {{.*#+}} ymm0 = -(ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfnmsub.ps.256(<8 x float> %b, <8 x float> %a, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fnmsub_aba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_aba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmsubps {{.*#+}} ymm0 = -(ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfnmsub.ps.256(<8 x float> %a, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
define <8 x float> @test_x86_fnmsub_bba_ps_y(<8 x float> %a, <8 x float> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_bba_ps_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfnmsubps {{.*#+}} ymm0 = -(ymm0 * ymm0) - mem
; FMA4-NEXT: retq
%res = call <8 x float> @llvm.x86.fma.vfnmsub.ps.256(<8 x float> %b, <8 x float> %b, <8 x float> %a) nounwind
ret <8 x float> %res
}
declare <2 x double> @llvm.x86.fma.vfnmsub.pd(<2 x double>, <2 x double>, <2 x double>) nounwind readnone
define <2 x double> @test_x86_fnmsub_baa_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_baa_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmsubpd {{.*#+}} xmm0 = -(xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfnmsub.pd(<2 x double> %b, <2 x double> %a, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fnmsub_aba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_aba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmsubpd {{.*#+}} xmm0 = -(xmm0 * mem) - xmm0
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfnmsub.pd(<2 x double> %a, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
define <2 x double> @test_x86_fnmsub_bba_pd(<2 x double> %a, <2 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_bba_pd:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfnmsubpd {{.*#+}} xmm0 = -(xmm0 * xmm0) - mem
; FMA4-NEXT: retq
%res = call <2 x double> @llvm.x86.fma.vfnmsub.pd(<2 x double> %b, <2 x double> %b, <2 x double> %a) nounwind
ret <2 x double> %res
}
declare <4 x double> @llvm.x86.fma.vfnmsub.pd.256(<4 x double>, <4 x double>, <4 x double>) nounwind readnone
define <4 x double> @test_x86_fnmsub_baa_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_baa_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmsubpd {{.*#+}} ymm0 = -(ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfnmsub.pd.256(<4 x double> %b, <4 x double> %a, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fnmsub_aba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_aba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmsubpd {{.*#+}} ymm0 = -(ymm0 * mem) - ymm0
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfnmsub.pd.256(<4 x double> %a, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
define <4 x double> @test_x86_fnmsub_bba_pd_y(<4 x double> %a, <4 x double> %b) #0 {
; FMA4-LABEL: test_x86_fnmsub_bba_pd_y:
; FMA4: # %bb.0:
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfnmsubpd {{.*#+}} ymm0 = -(ymm0 * ymm0) - mem
; FMA4-NEXT: retq
%res = call <4 x double> @llvm.x86.fma.vfnmsub.pd.256(<4 x double> %b, <4 x double> %b, <4 x double> %a) nounwind
ret <4 x double> %res
}
Reference in New Issue View Git Blame Copy Permalink