This is an archive of the discontinued LLVM Phabricator instance.

Differential D111129

[AMDGPU][GlobalISel] Fix legalization of G_UMULH
ClosedPublic

Authored by foad on Oct 5 2021, 2:42 AM.

Details

Reviewers
aemerson
Petar.Avramovic
arsenm
rampitec
Commits
rG9ce4f372067c: [AMDGPU][GlobalISel] Fix legalization of G_UMULH
Summary

Scalarize before narrowing because the narrowing implementation does not
work on vectors. This matches what we do for regular G_MUL.

Diff Detail

Event Timeline

foad created this revision.Oct 5 2021, 2:42 AM
Herald added subscribers: kerbowa, hiraditya, t-tye and 8 others. · View Herald TranscriptOct 5 2021, 2:42 AM
foad requested review of this revision.Oct 5 2021, 2:42 AM
Herald added a project: Restricted Project. · View Herald TranscriptOct 5 2021, 2:42 AM
Herald added subscribers: llvm-commits, wdng. · View Herald Transcript
foad added reviewers: aemerson, Petar.Avramovic, arsenm, rampitec.Oct 5 2021, 2:43 AM
Petar.Avramovic accepted this revision.Oct 5 2021, 2:48 AM
This revision is now accepted and ready to land.Oct 5 2021, 2:48 AM
This revision was landed with ongoing or failed builds.Oct 5 2021, 2:56 AM
Closed by commit rG9ce4f372067c: [AMDGPU][GlobalISel] Fix legalization of G_UMULH (authored by foad). · Explain Why
This revision was automatically updated to reflect the committed changes.
foad added a commit: rG9ce4f372067c: [AMDGPU][GlobalISel] Fix legalization of G_UMULH.
Harbormaster completed remote builds in B127004: Diff 377128.Oct 5 2021, 3:13 AM
foad mentioned this in D110890: [GlobalISel] Port the udiv -> mul by constant combine..Oct 5 2021, 5:36 AM

Revision Contents

PathSize
llvm/
lib/
Target/
AMDGPU/
2 lines
test/
CodeGen/
AMDGPU/
GlobalISel/
126 lines
180 lines

Diff 377136

llvm/lib/Target/AMDGPU/AMDGPULegalizerInfo.cpp

Show First 20 Lines • Show All 600 Lines • ▼ Show 20 Lines getActionDefinitionsBuilder(
{G_SDIV, G_UDIV, G_SREM, G_UREM, G_SDIVREM, G_UDIVREM}) {G_SDIV, G_UDIV, G_SREM, G_UREM, G_SDIVREM, G_UDIVREM})
.customFor({S32, S64}) .customFor({S32, S64})
.clampScalar(0, S32, S64) .clampScalar(0, S32, S64)
.widenScalarToNextPow2(0, 32) .widenScalarToNextPow2(0, 32)
.scalarize(0); .scalarize(0);
auto &Mulh = getActionDefinitionsBuilder({G_UMULH, G_SMULH}) auto &Mulh = getActionDefinitionsBuilder({G_UMULH, G_SMULH})
.legalFor({S32}) .legalFor({S32})
.maxScalarOrElt(0, S32); .maxScalar(0, S32);
if (ST.hasVOP3PInsts()) { if (ST.hasVOP3PInsts()) {
Mulh Mulh
.clampMaxNumElements(0, S8, 2) .clampMaxNumElements(0, S8, 2)
.lowerFor({V2S8}); .lowerFor({V2S8});
} }
Mulh Mulh
▲ Show 20 Lines • Show All 4,529 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-umulh.mir

# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py # NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -global-isel-abort=0 -O0 -run-pass=legalizer %s -o - | FileCheck %s --check-prefix=GFX8 # RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=fiji -O0 -run-pass=legalizer %s -o - | FileCheck %s --check-prefix=GFX8
# RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx906 -global-isel-abort=0 -O0 -run-pass=legalizer %s -o - | FileCheck %s --check-prefix=GFX9 # RUN: llc -mtriple=amdgcn-mesa-mesa3d -mcpu=gfx906 -O0 -run-pass=legalizer %s -o - | FileCheck %s --check-prefix=GFX9
--- ---
name: test_umulh_s32 name: test_umulh_s32
body: | body: |
bb.0: bb.0:
liveins: $vgpr0, $vgpr1 liveins: $vgpr0, $vgpr1
; GFX8-LABEL: name: test_umulh_s32 ; GFX8-LABEL: name: test_umulh_s32
▲ Show 20 Lines • Show All 106 Lines • ▼ Show 20 Lines bb.0:
; GFX9: $vgpr0_vgpr1 = COPY [[MV]](s64) ; GFX9: $vgpr0_vgpr1 = COPY [[MV]](s64)
%0:_(s64) = COPY $vgpr0_vgpr1 %0:_(s64) = COPY $vgpr0_vgpr1
%1:_(s64) = COPY $vgpr2_vgpr3 %1:_(s64) = COPY $vgpr2_vgpr3
%2:_(s64) = G_UMULH %0, %1 %2:_(s64) = G_UMULH %0, %1
$vgpr0_vgpr1 = COPY %2 $vgpr0_vgpr1 = COPY %2
... ...
--- ---
name: test_umulh_v2s64
body: |
bb.0:
liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7
; GFX8-LABEL: name: test_umulh_v2s64
; GFX8: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; GFX8: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
; GFX8: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
; GFX8: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
; GFX8: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
; GFX8: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
; GFX8: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV6]]
; GFX8: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV4]], [[UV7]]
; GFX8: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV6]]
; GFX8: [[UADDO:%[0-9]+]]:_(s32), [[UADDO1:%[0-9]+]]:_(s1) = G_UADDO [[MUL]], [[MUL1]]
; GFX8: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO1]](s1)
; GFX8: [[UADDO2:%[0-9]+]]:_(s32), [[UADDO3:%[0-9]+]]:_(s1) = G_UADDO [[UADDO]], [[UMULH]]
; GFX8: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO3]](s1)
; GFX8: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[ZEXT]], [[ZEXT1]]
; GFX8: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV7]]
; GFX8: [[UMULH1:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV6]]
; GFX8: [[UMULH2:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV7]]
; GFX8: [[UADDO4:%[0-9]+]]:_(s32), [[UADDO5:%[0-9]+]]:_(s1) = G_UADDO [[MUL2]], [[UMULH1]]
; GFX8: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO5]](s1)
; GFX8: [[UADDO6:%[0-9]+]]:_(s32), [[UADDO7:%[0-9]+]]:_(s1) = G_UADDO [[UADDO4]], [[UMULH2]]
; GFX8: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO7]](s1)
; GFX8: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ZEXT2]], [[ZEXT3]]
; GFX8: [[UADDO8:%[0-9]+]]:_(s32), [[UADDO9:%[0-9]+]]:_(s1) = G_UADDO [[UADDO6]], [[ADD]]
; GFX8: [[ZEXT4:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO9]](s1)
; GFX8: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[ADD1]], [[ZEXT4]]
; GFX8: [[UMULH3:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV7]]
; GFX8: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[UMULH3]], [[ADD2]]
; GFX8: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO8]](s32), [[ADD3]](s32)
; GFX8: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
; GFX8: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
; GFX8: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[UV9]], [[UV10]]
; GFX8: [[MUL4:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV11]]
; GFX8: [[UMULH4:%[0-9]+]]:_(s32) = G_UMULH [[UV8]], [[UV10]]
; GFX8: [[UADDO10:%[0-9]+]]:_(s32), [[UADDO11:%[0-9]+]]:_(s1) = G_UADDO [[MUL3]], [[MUL4]]
; GFX8: [[ZEXT5:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO11]](s1)
; GFX8: [[UADDO12:%[0-9]+]]:_(s32), [[UADDO13:%[0-9]+]]:_(s1) = G_UADDO [[UADDO10]], [[UMULH4]]
; GFX8: [[ZEXT6:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO13]](s1)
; GFX8: [[ADD4:%[0-9]+]]:_(s32) = G_ADD [[ZEXT5]], [[ZEXT6]]
; GFX8: [[MUL5:%[0-9]+]]:_(s32) = G_MUL [[UV9]], [[UV11]]
; GFX8: [[UMULH5:%[0-9]+]]:_(s32) = G_UMULH [[UV9]], [[UV10]]
; GFX8: [[UMULH6:%[0-9]+]]:_(s32) = G_UMULH [[UV8]], [[UV11]]
; GFX8: [[UADDO14:%[0-9]+]]:_(s32), [[UADDO15:%[0-9]+]]:_(s1) = G_UADDO [[MUL5]], [[UMULH5]]
; GFX8: [[ZEXT7:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO15]](s1)
; GFX8: [[UADDO16:%[0-9]+]]:_(s32), [[UADDO17:%[0-9]+]]:_(s1) = G_UADDO [[UADDO14]], [[UMULH6]]
; GFX8: [[ZEXT8:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO17]](s1)
; GFX8: [[ADD5:%[0-9]+]]:_(s32) = G_ADD [[ZEXT7]], [[ZEXT8]]
; GFX8: [[UADDO18:%[0-9]+]]:_(s32), [[UADDO19:%[0-9]+]]:_(s1) = G_UADDO [[UADDO16]], [[ADD4]]
; GFX8: [[ZEXT9:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO19]](s1)
; GFX8: [[ADD6:%[0-9]+]]:_(s32) = G_ADD [[ADD5]], [[ZEXT9]]
; GFX8: [[UMULH7:%[0-9]+]]:_(s32) = G_UMULH [[UV9]], [[UV11]]
; GFX8: [[ADD7:%[0-9]+]]:_(s32) = G_ADD [[UMULH7]], [[ADD6]]
; GFX8: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO18]](s32), [[ADD7]](s32)
; GFX8: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV]](s64), [[MV1]](s64)
; GFX8: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
; GFX9-LABEL: name: test_umulh_v2s64
; GFX9: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; GFX9: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
; GFX9: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
; GFX9: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
; GFX9: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
; GFX9: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
; GFX9: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV6]]
; GFX9: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV4]], [[UV7]]
; GFX9: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV6]]
; GFX9: [[UADDO:%[0-9]+]]:_(s32), [[UADDO1:%[0-9]+]]:_(s1) = G_UADDO [[MUL]], [[MUL1]]
; GFX9: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO1]](s1)
; GFX9: [[UADDO2:%[0-9]+]]:_(s32), [[UADDO3:%[0-9]+]]:_(s1) = G_UADDO [[UADDO]], [[UMULH]]
; GFX9: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO3]](s1)
; GFX9: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[ZEXT]], [[ZEXT1]]
; GFX9: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV7]]
; GFX9: [[UMULH1:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV6]]
; GFX9: [[UMULH2:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV7]]
; GFX9: [[UADDO4:%[0-9]+]]:_(s32), [[UADDO5:%[0-9]+]]:_(s1) = G_UADDO [[MUL2]], [[UMULH1]]
; GFX9: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO5]](s1)
; GFX9: [[UADDO6:%[0-9]+]]:_(s32), [[UADDO7:%[0-9]+]]:_(s1) = G_UADDO [[UADDO4]], [[UMULH2]]
; GFX9: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO7]](s1)
; GFX9: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ZEXT2]], [[ZEXT3]]
; GFX9: [[UADDO8:%[0-9]+]]:_(s32), [[UADDO9:%[0-9]+]]:_(s1) = G_UADDO [[UADDO6]], [[ADD]]
; GFX9: [[ZEXT4:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO9]](s1)
; GFX9: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[ADD1]], [[ZEXT4]]
; GFX9: [[UMULH3:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV7]]
; GFX9: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[UMULH3]], [[ADD2]]
; GFX9: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO8]](s32), [[ADD3]](s32)
; GFX9: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
; GFX9: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
; GFX9: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[UV9]], [[UV10]]
; GFX9: [[MUL4:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV11]]
; GFX9: [[UMULH4:%[0-9]+]]:_(s32) = G_UMULH [[UV8]], [[UV10]]
; GFX9: [[UADDO10:%[0-9]+]]:_(s32), [[UADDO11:%[0-9]+]]:_(s1) = G_UADDO [[MUL3]], [[MUL4]]
; GFX9: [[ZEXT5:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO11]](s1)
; GFX9: [[UADDO12:%[0-9]+]]:_(s32), [[UADDO13:%[0-9]+]]:_(s1) = G_UADDO [[UADDO10]], [[UMULH4]]
; GFX9: [[ZEXT6:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO13]](s1)
; GFX9: [[ADD4:%[0-9]+]]:_(s32) = G_ADD [[ZEXT5]], [[ZEXT6]]
; GFX9: [[MUL5:%[0-9]+]]:_(s32) = G_MUL [[UV9]], [[UV11]]
; GFX9: [[UMULH5:%[0-9]+]]:_(s32) = G_UMULH [[UV9]], [[UV10]]
; GFX9: [[UMULH6:%[0-9]+]]:_(s32) = G_UMULH [[UV8]], [[UV11]]
; GFX9: [[UADDO14:%[0-9]+]]:_(s32), [[UADDO15:%[0-9]+]]:_(s1) = G_UADDO [[MUL5]], [[UMULH5]]
; GFX9: [[ZEXT7:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO15]](s1)
; GFX9: [[UADDO16:%[0-9]+]]:_(s32), [[UADDO17:%[0-9]+]]:_(s1) = G_UADDO [[UADDO14]], [[UMULH6]]
; GFX9: [[ZEXT8:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO17]](s1)
; GFX9: [[ADD5:%[0-9]+]]:_(s32) = G_ADD [[ZEXT7]], [[ZEXT8]]
; GFX9: [[UADDO18:%[0-9]+]]:_(s32), [[UADDO19:%[0-9]+]]:_(s1) = G_UADDO [[UADDO16]], [[ADD4]]
; GFX9: [[ZEXT9:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO19]](s1)
; GFX9: [[ADD6:%[0-9]+]]:_(s32) = G_ADD [[ADD5]], [[ZEXT9]]
; GFX9: [[UMULH7:%[0-9]+]]:_(s32) = G_UMULH [[UV9]], [[UV11]]
; GFX9: [[ADD7:%[0-9]+]]:_(s32) = G_ADD [[UMULH7]], [[ADD6]]
; GFX9: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO18]](s32), [[ADD7]](s32)
; GFX9: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV]](s64), [[MV1]](s64)
; GFX9: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
%0:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
%1:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
%2:_(<2 x s64>) = G_UMULH %0, %1
$vgpr0_vgpr1_vgpr2_vgpr3 = COPY %2
...
---
name: test_umulh_s16 name: test_umulh_s16
body: | body: |
bb.0: bb.0:
liveins: $vgpr0, $vgpr1 liveins: $vgpr0, $vgpr1
; GFX8-LABEL: name: test_umulh_s16 ; GFX8-LABEL: name: test_umulh_s16
; GFX8: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX8: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX8: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; GFX8: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
▲ Show 20 Lines • Show All 495 Lines • Show Last 20 Lines

llvm/test/CodeGen/AMDGPU/GlobalISel/legalize-umulo.mir

Show First 20 Lines • Show All 183 Lines • ▼ Show 20 Lines bb.0:
%1:_(s64) = COPY $vgpr2_vgpr3 %1:_(s64) = COPY $vgpr2_vgpr3
%2:_(s64), %3:_(s1) = G_UMULO %0, %1 %2:_(s64), %3:_(s1) = G_UMULO %0, %1
%4:_(s64) = G_ZEXT %3 %4:_(s64) = G_ZEXT %3
$vgpr0_vgpr1 = COPY %2 $vgpr0_vgpr1 = COPY %2
$vgpr2_vgpr3 = COPY %4 $vgpr2_vgpr3 = COPY %4
... ...
--- ---
name: test_umulo_v2s64
body: |
bb.0:
liveins: $vgpr0_vgpr1_vgpr2_vgpr3, $vgpr4_vgpr5_vgpr6_vgpr7
; GFX8-LABEL: name: test_umulo_v2s64
; GFX8: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; GFX8: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
; GFX8: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
; GFX8: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
; GFX8: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
; GFX8: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
; GFX8: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV6]]
; GFX8: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV4]], [[UV7]]
; GFX8: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV6]]
; GFX8: [[UADDO:%[0-9]+]]:_(s32), [[UADDO1:%[0-9]+]]:_(s1) = G_UADDO [[MUL]], [[MUL1]]
; GFX8: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO1]](s1)
; GFX8: [[UADDO2:%[0-9]+]]:_(s32), [[UADDO3:%[0-9]+]]:_(s1) = G_UADDO [[UADDO]], [[UMULH]]
; GFX8: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO3]](s1)
; GFX8: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[ZEXT]], [[ZEXT1]]
; GFX8: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV7]]
; GFX8: [[UMULH1:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV6]]
; GFX8: [[UMULH2:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV7]]
; GFX8: [[UADDO4:%[0-9]+]]:_(s32), [[UADDO5:%[0-9]+]]:_(s1) = G_UADDO [[MUL2]], [[UMULH1]]
; GFX8: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO5]](s1)
; GFX8: [[UADDO6:%[0-9]+]]:_(s32), [[UADDO7:%[0-9]+]]:_(s1) = G_UADDO [[UADDO4]], [[UMULH2]]
; GFX8: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO7]](s1)
; GFX8: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ZEXT2]], [[ZEXT3]]
; GFX8: [[UADDO8:%[0-9]+]]:_(s32), [[UADDO9:%[0-9]+]]:_(s1) = G_UADDO [[UADDO6]], [[ADD]]
; GFX8: [[ZEXT4:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO9]](s1)
; GFX8: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[ADD1]], [[ZEXT4]]
; GFX8: [[UMULH3:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV7]]
; GFX8: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[UMULH3]], [[ADD2]]
; GFX8: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO8]](s32), [[ADD3]](s32)
; GFX8: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
; GFX8: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
; GFX8: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
; GFX8: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV10]]
; GFX8: [[MUL4:%[0-9]+]]:_(s32) = G_MUL [[UV9]], [[UV10]]
; GFX8: [[MUL5:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV11]]
; GFX8: [[UMULH4:%[0-9]+]]:_(s32) = G_UMULH [[UV8]], [[UV10]]
; GFX8: [[ADD4:%[0-9]+]]:_(s32) = G_ADD [[MUL4]], [[MUL5]]
; GFX8: [[ADD5:%[0-9]+]]:_(s32) = G_ADD [[ADD4]], [[UMULH4]]
; GFX8: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL3]](s32), [[ADD5]](s32)
; GFX8: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ne), [[MV]](s64), [[C]]
; GFX8: [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
; GFX8: [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
; GFX8: [[MUL6:%[0-9]+]]:_(s32) = G_MUL [[UV13]], [[UV14]]
; GFX8: [[MUL7:%[0-9]+]]:_(s32) = G_MUL [[UV12]], [[UV15]]
; GFX8: [[UMULH5:%[0-9]+]]:_(s32) = G_UMULH [[UV12]], [[UV14]]
; GFX8: [[UADDO10:%[0-9]+]]:_(s32), [[UADDO11:%[0-9]+]]:_(s1) = G_UADDO [[MUL6]], [[MUL7]]
; GFX8: [[ZEXT5:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO11]](s1)
; GFX8: [[UADDO12:%[0-9]+]]:_(s32), [[UADDO13:%[0-9]+]]:_(s1) = G_UADDO [[UADDO10]], [[UMULH5]]
; GFX8: [[ZEXT6:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO13]](s1)
; GFX8: [[ADD6:%[0-9]+]]:_(s32) = G_ADD [[ZEXT5]], [[ZEXT6]]
; GFX8: [[MUL8:%[0-9]+]]:_(s32) = G_MUL [[UV13]], [[UV15]]
; GFX8: [[UMULH6:%[0-9]+]]:_(s32) = G_UMULH [[UV13]], [[UV14]]
; GFX8: [[UMULH7:%[0-9]+]]:_(s32) = G_UMULH [[UV12]], [[UV15]]
; GFX8: [[UADDO14:%[0-9]+]]:_(s32), [[UADDO15:%[0-9]+]]:_(s1) = G_UADDO [[MUL8]], [[UMULH6]]
; GFX8: [[ZEXT7:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO15]](s1)
; GFX8: [[UADDO16:%[0-9]+]]:_(s32), [[UADDO17:%[0-9]+]]:_(s1) = G_UADDO [[UADDO14]], [[UMULH7]]
; GFX8: [[ZEXT8:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO17]](s1)
; GFX8: [[ADD7:%[0-9]+]]:_(s32) = G_ADD [[ZEXT7]], [[ZEXT8]]
; GFX8: [[UADDO18:%[0-9]+]]:_(s32), [[UADDO19:%[0-9]+]]:_(s1) = G_UADDO [[UADDO16]], [[ADD6]]
; GFX8: [[ZEXT9:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO19]](s1)
; GFX8: [[ADD8:%[0-9]+]]:_(s32) = G_ADD [[ADD7]], [[ZEXT9]]
; GFX8: [[UMULH8:%[0-9]+]]:_(s32) = G_UMULH [[UV13]], [[UV15]]
; GFX8: [[ADD9:%[0-9]+]]:_(s32) = G_ADD [[UMULH8]], [[ADD8]]
; GFX8: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO18]](s32), [[ADD9]](s32)
; GFX8: [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
; GFX8: [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
; GFX8: [[MUL9:%[0-9]+]]:_(s32) = G_MUL [[UV16]], [[UV18]]
; GFX8: [[MUL10:%[0-9]+]]:_(s32) = G_MUL [[UV17]], [[UV18]]
; GFX8: [[MUL11:%[0-9]+]]:_(s32) = G_MUL [[UV16]], [[UV19]]
; GFX8: [[UMULH9:%[0-9]+]]:_(s32) = G_UMULH [[UV16]], [[UV18]]
; GFX8: [[ADD10:%[0-9]+]]:_(s32) = G_ADD [[MUL10]], [[MUL11]]
; GFX8: [[ADD11:%[0-9]+]]:_(s32) = G_ADD [[ADD10]], [[UMULH9]]
; GFX8: [[MV3:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL9]](s32), [[ADD11]](s32)
; GFX8: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ne), [[MV2]](s64), [[C]]
; GFX8: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV1]](s64), [[MV3]](s64)
; GFX8: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; GFX8: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[ICMP]](s1)
; GFX8: [[AND:%[0-9]+]]:_(s64) = G_AND [[ANYEXT]], [[C1]]
; GFX8: [[ANYEXT1:%[0-9]+]]:_(s64) = G_ANYEXT [[ICMP1]](s1)
; GFX8: [[AND1:%[0-9]+]]:_(s64) = G_AND [[ANYEXT1]], [[C1]]
; GFX8: [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[AND]](s64), [[AND1]](s64)
; GFX8: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
; GFX8: $vgpr4_vgpr5_vgpr6_vgpr7 = COPY [[BUILD_VECTOR1]](<2 x s64>)
; GFX9-LABEL: name: test_umulo_v2s64
; GFX9: [[COPY:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
; GFX9: [[COPY1:%[0-9]+]]:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
; GFX9: [[UV:%[0-9]+]]:_(s64), [[UV1:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY]](<2 x s64>)
; GFX9: [[UV2:%[0-9]+]]:_(s64), [[UV3:%[0-9]+]]:_(s64) = G_UNMERGE_VALUES [[COPY1]](<2 x s64>)
; GFX9: [[UV4:%[0-9]+]]:_(s32), [[UV5:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
; GFX9: [[UV6:%[0-9]+]]:_(s32), [[UV7:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
; GFX9: [[MUL:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV6]]
; GFX9: [[MUL1:%[0-9]+]]:_(s32) = G_MUL [[UV4]], [[UV7]]
; GFX9: [[UMULH:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV6]]
; GFX9: [[UADDO:%[0-9]+]]:_(s32), [[UADDO1:%[0-9]+]]:_(s1) = G_UADDO [[MUL]], [[MUL1]]
; GFX9: [[ZEXT:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO1]](s1)
; GFX9: [[UADDO2:%[0-9]+]]:_(s32), [[UADDO3:%[0-9]+]]:_(s1) = G_UADDO [[UADDO]], [[UMULH]]
; GFX9: [[ZEXT1:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO3]](s1)
; GFX9: [[ADD:%[0-9]+]]:_(s32) = G_ADD [[ZEXT]], [[ZEXT1]]
; GFX9: [[MUL2:%[0-9]+]]:_(s32) = G_MUL [[UV5]], [[UV7]]
; GFX9: [[UMULH1:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV6]]
; GFX9: [[UMULH2:%[0-9]+]]:_(s32) = G_UMULH [[UV4]], [[UV7]]
; GFX9: [[UADDO4:%[0-9]+]]:_(s32), [[UADDO5:%[0-9]+]]:_(s1) = G_UADDO [[MUL2]], [[UMULH1]]
; GFX9: [[ZEXT2:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO5]](s1)
; GFX9: [[UADDO6:%[0-9]+]]:_(s32), [[UADDO7:%[0-9]+]]:_(s1) = G_UADDO [[UADDO4]], [[UMULH2]]
; GFX9: [[ZEXT3:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO7]](s1)
; GFX9: [[ADD1:%[0-9]+]]:_(s32) = G_ADD [[ZEXT2]], [[ZEXT3]]
; GFX9: [[UADDO8:%[0-9]+]]:_(s32), [[UADDO9:%[0-9]+]]:_(s1) = G_UADDO [[UADDO6]], [[ADD]]
; GFX9: [[ZEXT4:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO9]](s1)
; GFX9: [[ADD2:%[0-9]+]]:_(s32) = G_ADD [[ADD1]], [[ZEXT4]]
; GFX9: [[UMULH3:%[0-9]+]]:_(s32) = G_UMULH [[UV5]], [[UV7]]
; GFX9: [[ADD3:%[0-9]+]]:_(s32) = G_ADD [[UMULH3]], [[ADD2]]
; GFX9: [[MV:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO8]](s32), [[ADD3]](s32)
; GFX9: [[C:%[0-9]+]]:_(s64) = G_CONSTANT i64 0
; GFX9: [[UV8:%[0-9]+]]:_(s32), [[UV9:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV]](s64)
; GFX9: [[UV10:%[0-9]+]]:_(s32), [[UV11:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV2]](s64)
; GFX9: [[MUL3:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV10]]
; GFX9: [[MUL4:%[0-9]+]]:_(s32) = G_MUL [[UV9]], [[UV10]]
; GFX9: [[MUL5:%[0-9]+]]:_(s32) = G_MUL [[UV8]], [[UV11]]
; GFX9: [[UMULH4:%[0-9]+]]:_(s32) = G_UMULH [[UV8]], [[UV10]]
; GFX9: [[ADD4:%[0-9]+]]:_(s32) = G_ADD [[MUL4]], [[MUL5]]
; GFX9: [[ADD5:%[0-9]+]]:_(s32) = G_ADD [[ADD4]], [[UMULH4]]
; GFX9: [[MV1:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL3]](s32), [[ADD5]](s32)
; GFX9: [[ICMP:%[0-9]+]]:_(s1) = G_ICMP intpred(ne), [[MV]](s64), [[C]]
; GFX9: [[UV12:%[0-9]+]]:_(s32), [[UV13:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
; GFX9: [[UV14:%[0-9]+]]:_(s32), [[UV15:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
; GFX9: [[MUL6:%[0-9]+]]:_(s32) = G_MUL [[UV13]], [[UV14]]
; GFX9: [[MUL7:%[0-9]+]]:_(s32) = G_MUL [[UV12]], [[UV15]]
; GFX9: [[UMULH5:%[0-9]+]]:_(s32) = G_UMULH [[UV12]], [[UV14]]
; GFX9: [[UADDO10:%[0-9]+]]:_(s32), [[UADDO11:%[0-9]+]]:_(s1) = G_UADDO [[MUL6]], [[MUL7]]
; GFX9: [[ZEXT5:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO11]](s1)
; GFX9: [[UADDO12:%[0-9]+]]:_(s32), [[UADDO13:%[0-9]+]]:_(s1) = G_UADDO [[UADDO10]], [[UMULH5]]
; GFX9: [[ZEXT6:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO13]](s1)
; GFX9: [[ADD6:%[0-9]+]]:_(s32) = G_ADD [[ZEXT5]], [[ZEXT6]]
; GFX9: [[MUL8:%[0-9]+]]:_(s32) = G_MUL [[UV13]], [[UV15]]
; GFX9: [[UMULH6:%[0-9]+]]:_(s32) = G_UMULH [[UV13]], [[UV14]]
; GFX9: [[UMULH7:%[0-9]+]]:_(s32) = G_UMULH [[UV12]], [[UV15]]
; GFX9: [[UADDO14:%[0-9]+]]:_(s32), [[UADDO15:%[0-9]+]]:_(s1) = G_UADDO [[MUL8]], [[UMULH6]]
; GFX9: [[ZEXT7:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO15]](s1)
; GFX9: [[UADDO16:%[0-9]+]]:_(s32), [[UADDO17:%[0-9]+]]:_(s1) = G_UADDO [[UADDO14]], [[UMULH7]]
; GFX9: [[ZEXT8:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO17]](s1)
; GFX9: [[ADD7:%[0-9]+]]:_(s32) = G_ADD [[ZEXT7]], [[ZEXT8]]
; GFX9: [[UADDO18:%[0-9]+]]:_(s32), [[UADDO19:%[0-9]+]]:_(s1) = G_UADDO [[UADDO16]], [[ADD6]]
; GFX9: [[ZEXT9:%[0-9]+]]:_(s32) = G_ZEXT [[UADDO19]](s1)
; GFX9: [[ADD8:%[0-9]+]]:_(s32) = G_ADD [[ADD7]], [[ZEXT9]]
; GFX9: [[UMULH8:%[0-9]+]]:_(s32) = G_UMULH [[UV13]], [[UV15]]
; GFX9: [[ADD9:%[0-9]+]]:_(s32) = G_ADD [[UMULH8]], [[ADD8]]
; GFX9: [[MV2:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[UADDO18]](s32), [[ADD9]](s32)
; GFX9: [[UV16:%[0-9]+]]:_(s32), [[UV17:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV1]](s64)
; GFX9: [[UV18:%[0-9]+]]:_(s32), [[UV19:%[0-9]+]]:_(s32) = G_UNMERGE_VALUES [[UV3]](s64)
; GFX9: [[MUL9:%[0-9]+]]:_(s32) = G_MUL [[UV16]], [[UV18]]
; GFX9: [[MUL10:%[0-9]+]]:_(s32) = G_MUL [[UV17]], [[UV18]]
; GFX9: [[MUL11:%[0-9]+]]:_(s32) = G_MUL [[UV16]], [[UV19]]
; GFX9: [[UMULH9:%[0-9]+]]:_(s32) = G_UMULH [[UV16]], [[UV18]]
; GFX9: [[ADD10:%[0-9]+]]:_(s32) = G_ADD [[MUL10]], [[MUL11]]
; GFX9: [[ADD11:%[0-9]+]]:_(s32) = G_ADD [[ADD10]], [[UMULH9]]
; GFX9: [[MV3:%[0-9]+]]:_(s64) = G_MERGE_VALUES [[MUL9]](s32), [[ADD11]](s32)
; GFX9: [[ICMP1:%[0-9]+]]:_(s1) = G_ICMP intpred(ne), [[MV2]](s64), [[C]]
; GFX9: [[BUILD_VECTOR:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[MV1]](s64), [[MV3]](s64)
; GFX9: [[C1:%[0-9]+]]:_(s64) = G_CONSTANT i64 1
; GFX9: [[ANYEXT:%[0-9]+]]:_(s64) = G_ANYEXT [[ICMP]](s1)
; GFX9: [[AND:%[0-9]+]]:_(s64) = G_AND [[ANYEXT]], [[C1]]
; GFX9: [[ANYEXT1:%[0-9]+]]:_(s64) = G_ANYEXT [[ICMP1]](s1)
; GFX9: [[AND1:%[0-9]+]]:_(s64) = G_AND [[ANYEXT1]], [[C1]]
; GFX9: [[BUILD_VECTOR1:%[0-9]+]]:_(<2 x s64>) = G_BUILD_VECTOR [[AND]](s64), [[AND1]](s64)
; GFX9: $vgpr0_vgpr1_vgpr2_vgpr3 = COPY [[BUILD_VECTOR]](<2 x s64>)
; GFX9: $vgpr4_vgpr5_vgpr6_vgpr7 = COPY [[BUILD_VECTOR1]](<2 x s64>)
%0:_(<2 x s64>) = COPY $vgpr0_vgpr1_vgpr2_vgpr3
%1:_(<2 x s64>) = COPY $vgpr4_vgpr5_vgpr6_vgpr7
%2:_(<2 x s64>), %3:_(<2 x s1>) = G_UMULO %0, %1
%4:_(<2 x s64>) = G_ZEXT %3
$vgpr0_vgpr1_vgpr2_vgpr3 = COPY %2
$vgpr4_vgpr5_vgpr6_vgpr7 = COPY %4
...
---
name: test_umulo_s24 name: test_umulo_s24
body: | body: |
bb.0: bb.0:
liveins: $vgpr0, $vgpr1 liveins: $vgpr0, $vgpr1
; GFX8-LABEL: name: test_umulo_s24 ; GFX8-LABEL: name: test_umulo_s24
; GFX8: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0 ; GFX8: [[COPY:%[0-9]+]]:_(s32) = COPY $vgpr0
; GFX8: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1 ; GFX8: [[COPY1:%[0-9]+]]:_(s32) = COPY $vgpr1
▲ Show 20 Lines • Show All 399 Lines • Show Last 20 Lines