maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/llvm/test/CodeGen/X86/fma4-commute-x86.ll

564 lines
23 KiB
LLVM
Raw Normal View History

[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; 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 --check-prefix=FMA
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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmaddss %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmaddss %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfmaddss (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmaddps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmaddps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfmaddps (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmaddps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmaddps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfmaddps (%rcx), %ymm0, %ymm0, %ymm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmaddsd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmaddsd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfmaddsd (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmaddpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmaddpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfmaddpd (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmaddpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmaddpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfmaddpd (%rcx), %ymm0, %ymm0, %ymm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmaddps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmaddps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfnmaddps (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmaddps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmaddps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfnmaddps (%rcx), %ymm0, %ymm0, %ymm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmaddpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmaddpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfnmaddpd (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmaddpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmaddpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfnmaddpd (%rcx), %ymm0, %ymm0, %ymm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmsubps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfmsubps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfmsubps (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmsubps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfmsubps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfmsubps (%rcx), %ymm0, %ymm0, %ymm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmsubpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfmsubpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfmsubpd (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmsubpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfmsubpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfmsubpd (%rcx), %ymm0, %ymm0, %ymm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmsubps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %xmm0
; FMA4-NEXT: vfnmsubps %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %xmm0
; FMA4-NEXT: vfnmsubps (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmsubps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rcx), %ymm0
; FMA4-NEXT: vfnmsubps %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovaps (%rdx), %ymm0
; FMA4-NEXT: vfnmsubps (%rcx), %ymm0, %ymm0, %ymm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmsubpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %xmm0
; FMA4-NEXT: vfnmsubpd %xmm0, (%rdx), %xmm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %xmm0
; FMA4-NEXT: vfnmsubpd (%rcx), %xmm0, %xmm0, %xmm0
; 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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmsubpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rcx), %ymm0
; FMA4-NEXT: vfnmsubpd %ymm0, (%rdx), %ymm0, %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:
[CodeGen] Unify MBB reference format in both MIR and debug output As part of the unification of the debug format and the MIR format, print MBB references as '%bb.5'. The MIR printer prints the IR name of a MBB only for block definitions. * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)->getNumber\(\)/" << printMBBReference(*\1)/g' * find . \( -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#" << ([a-zA-Z0-9_]+)\.getNumber\(\)/" << printMBBReference(\1)/g' * find . \( -name "*.txt" -o -name "*.s" -o -name "*.mir" -o -name "*.cpp" -o -name "*.h" -o -name "*.ll" \) -type f -print0 | xargs -0 sed -i '' -E 's/BB#([0-9]+)/%bb.\1/g' * grep -nr 'BB#' and fix Differential Revision: https://reviews.llvm.org/D40422 llvm-svn: 319665
2017-12-05 01:18:51 +08:00
; FMA4: # %bb.0:
[X86] Add separate intrinsics for scalar FMA4 instructions. Summary: These instructions zero the non-scalar part of the lower 128-bits which makes them different than the FMA3 instructions which pass through the non-scalar part of the lower 128-bits. I've only added fmadd because we should be able to derive all other variants using operand negation in the intrinsic header like we do for AVX512. I think there are still some missed negate folding opportunities with the FMA4 instructions in light of this behavior difference that I hadn't noticed before. I've split the tests so that we can use different intrinsics for scalar testing between the two. I just copied the tests split the RUN lines and changed out the scalar intrinsics. fma4-fneg-combine.ll is a new test to make sure we negate the fma4 intrinsics correctly though there are a couple TODOs in it. Reviewers: RKSimon, spatel Reviewed By: RKSimon Subscribers: llvm-commits Differential Revision: https://reviews.llvm.org/D39851 llvm-svn: 318984
2017-11-26 02:32:43 +08:00
; FMA4-NEXT: vmovapd (%rdx), %ymm0
; FMA4-NEXT: vfnmsubpd (%rcx), %ymm0, %ymm0, %ymm0
; 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
}