derek.holloway
7502bf64a4
CodeQL / Analyze (actions) (push) Canceled after 0s
CodeQL / Analyze (python) (push) Canceled after 0s
pre-commit / pre-commit (push) Canceled after 0s
Multi-Arch CI / setup (push) Canceled after 0s
Publish build_manylinux_x86_64 images / publish_build_manylinux_x86_64 (push) Canceled after 0s
Setup (Multi-Arch) / setup (push) Canceled after 0s
Multi-Arch CI / Linux::${{ fromJSON(needs.setup.outputs.linux_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubi10 image / publish_no_rocm_image_ubi10 (push) Canceled after 0s
Multi-Arch CI / Windows::${{ fromJSON(needs.setup.outputs.windows_build_config || '{}').build_variant_label || 'skip' }} (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Multi-Arch CI / Manifest Diff (push) Canceled after 0s
TheRock Manifest Diff Report / Generate Manifest Diff Report (push) Canceled after 0s
Multi-Arch CI / CI Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_ocl_rt images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04_rocgdb images / publish_no_rocm_image_ubuntu24_04_rocgdb (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_ocl_rt (push) Canceled after 0s
Publish TheRock Docker image / build-and-push-image (push) Canceled after 0s
Unit Tests / Unit Tests :: ubuntu-24.04 (push) Canceled after 0s
Unit Tests / Unit Tests :: windows-2022 (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_media (push) Canceled after 0s
Unit Tests / Unit Tests Summary (push) Canceled after 0s
Publish no_rocm_image_ubuntu24_04 images / publish_no_rocm_image_ubuntu24_04_openmpi (push) Canceled after 0s
643 lines
29 KiB
YAML
643 lines
29 KiB
YAML
# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py
|
|
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=tahiti -run-pass=instruction-select -verify-machineinstrs \
|
|
# RUN: -global-isel-abort=0 -o - %s | FileCheck -check-prefix=WAVE64 %s
|
|
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=fiji -run-pass=instruction-select -verify-machineinstrs \
|
|
# RUN: -global-isel-abort=0 -o - %s | FileCheck -check-prefix=WAVE64 %s
|
|
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1010 -mattr=+wavefrontsize32 -run-pass=instruction-select \
|
|
# RUN: -global-isel-abort=0 -verify-machineinstrs -o - %s | FileCheck -check-prefixes=WAVE32,PREGFX11 %s
|
|
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 -mattr=+wavefrontsize32,+real-true16 -run-pass=instruction-select \
|
|
# RUN: -global-isel-abort=0 -verify-machineinstrs -o - %s | FileCheck -check-prefixes=WAVE32,GFX11-TRUE16 %s
|
|
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1100 -mattr=+wavefrontsize32,-real-true16 -run-pass=instruction-select \
|
|
# RUN: -global-isel-abort=0 -verify-machineinstrs -o - %s | FileCheck -check-prefixes=WAVE32,GFX11-FAKE16 %s
|
|
|
|
---
|
|
|
|
name: or_s1_vcc_vcc_vcc
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1
|
|
; WAVE64-LABEL: name: or_s1_vcc_vcc_vcc
|
|
; WAVE64: liveins: $vgpr0, $vgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE64-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
|
|
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
|
|
; WAVE64-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
|
|
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_s1_vcc_vcc_vcc
|
|
; WAVE32: liveins: $vgpr0, $vgpr1
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE32-NEXT: [[V_MOV_B32_e32_:%[0-9]+]]:vgpr_32 = V_MOV_B32_e32 0, implicit $exec
|
|
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY]], [[V_MOV_B32_e32_]], implicit $exec
|
|
; WAVE32-NEXT: [[V_CMP_EQ_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_EQ_U32_e64 [[COPY1]], [[V_MOV_B32_e32_]], implicit $exec
|
|
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_EQ_U32_e64_]], [[V_CMP_EQ_U32_e64_1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
%0:vgpr(s32) = COPY $vgpr0
|
|
%1:vgpr(s32) = COPY $vgpr1
|
|
%2:vgpr(s32) = G_CONSTANT i32 0
|
|
%3:vcc(s1) = G_ICMP intpred(eq), %0, %2
|
|
%4:vcc(s1) = G_ICMP intpred(eq), %1, %2
|
|
%5:vcc(s1) = G_OR %3, %4
|
|
S_ENDPGM 0, implicit %5
|
|
...
|
|
|
|
# Should fail to select
|
|
---
|
|
|
|
name: or_s1_sgpr_sgpr_sgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $sgpr0, $sgpr1
|
|
; WAVE64-LABEL: name: or_s1_sgpr_sgpr_sgpr
|
|
; WAVE64: liveins: $sgpr0, $sgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE64-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_s1_sgpr_sgpr_sgpr
|
|
; WAVE32: liveins: $sgpr0, $sgpr1
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
%0:sgpr(s32) = COPY $sgpr0
|
|
%1:sgpr(s32) = COPY $sgpr1
|
|
%2:sgpr(s1) = G_TRUNC %0
|
|
%3:sgpr(s1) = G_TRUNC %1
|
|
%4:sgpr(s1) = G_OR %2, %3
|
|
S_ENDPGM 0, implicit %4
|
|
...
|
|
|
|
---
|
|
|
|
name: or_s16_sgpr_sgpr_sgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $sgpr0, $sgpr1
|
|
; WAVE64-LABEL: name: or_s16_sgpr_sgpr_sgpr
|
|
; WAVE64: liveins: $sgpr0, $sgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE64-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_s16_sgpr_sgpr_sgpr
|
|
; WAVE32: liveins: $sgpr0, $sgpr1
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
%0:sgpr(s32) = COPY $sgpr0
|
|
%1:sgpr(s32) = COPY $sgpr1
|
|
%2:sgpr(s16) = G_TRUNC %0
|
|
%3:sgpr(s16) = G_TRUNC %1
|
|
%4:sgpr(s16) = G_OR %2, %3
|
|
S_ENDPGM 0, implicit %4
|
|
...
|
|
|
|
---
|
|
|
|
name: or_s16_vgpr_vgpr_vgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1
|
|
; WAVE64-LABEL: name: or_s16_vgpr_vgpr_vgpr
|
|
; WAVE64: liveins: $vgpr0, $vgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE64-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
|
|
;
|
|
; PREGFX11-LABEL: name: or_s16_vgpr_vgpr_vgpr
|
|
; PREGFX11: liveins: $vgpr0, $vgpr1
|
|
; PREGFX11-NEXT: {{ $}}
|
|
; PREGFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; PREGFX11-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; PREGFX11-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
|
|
; PREGFX11-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
|
|
;
|
|
; GFX11-TRUE16-LABEL: name: or_s16_vgpr_vgpr_vgpr
|
|
; GFX11-TRUE16: liveins: $vgpr0, $vgpr1
|
|
; GFX11-TRUE16-NEXT: {{ $}}
|
|
; GFX11-TRUE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-TRUE16-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; GFX11-TRUE16-NEXT: [[COPY2:%[0-9]+]]:vgpr_16 = COPY [[COPY]].lo16
|
|
; GFX11-TRUE16-NEXT: [[COPY3:%[0-9]+]]:vgpr_16 = COPY [[COPY1]].lo16
|
|
; GFX11-TRUE16-NEXT: [[V_OR_B16_t16_e64_:%[0-9]+]]:vgpr_16 = V_OR_B16_t16_e64 0, [[COPY2]], 0, [[COPY3]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: S_ENDPGM 0, implicit [[V_OR_B16_t16_e64_]]
|
|
;
|
|
; GFX11-FAKE16-LABEL: name: or_s16_vgpr_vgpr_vgpr
|
|
; GFX11-FAKE16: liveins: $vgpr0, $vgpr1
|
|
; GFX11-FAKE16-NEXT: {{ $}}
|
|
; GFX11-FAKE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-FAKE16-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; GFX11-FAKE16-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
|
|
%0:vgpr(s32) = COPY $vgpr0
|
|
%1:vgpr(s32) = COPY $vgpr1
|
|
%2:vgpr(s16) = G_TRUNC %0
|
|
%3:vgpr(s16) = G_TRUNC %1
|
|
%4:vgpr(s16) = G_OR %2, %3
|
|
S_ENDPGM 0, implicit %4
|
|
...
|
|
|
|
---
|
|
|
|
name: or_s32_sgpr_sgpr_sgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $sgpr0, $sgpr1
|
|
; WAVE64-LABEL: name: or_s32_sgpr_sgpr_sgpr
|
|
; WAVE64: liveins: $sgpr0, $sgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE64-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_s32_sgpr_sgpr_sgpr
|
|
; WAVE32: liveins: $sgpr0, $sgpr1
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
%0:sgpr(s32) = COPY $sgpr0
|
|
%1:sgpr(s32) = COPY $sgpr1
|
|
%2:sgpr(s32) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
---
|
|
|
|
name: or_s64_sgpr_sgpr_sgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE64-LABEL: name: or_s64_sgpr_sgpr_sgpr
|
|
; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
|
|
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_s64_sgpr_sgpr_sgpr
|
|
; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
|
|
; WAVE32-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
%0:sgpr(s64) = COPY $sgpr0_sgpr1
|
|
%1:sgpr(s64) = COPY $sgpr2_sgpr3
|
|
%2:sgpr(s64) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
---
|
|
|
|
name: or_v2s16_sgpr_sgpr_sgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $sgpr0, $sgpr1
|
|
; WAVE64-LABEL: name: or_v2s16_sgpr_sgpr_sgpr
|
|
; WAVE64: liveins: $sgpr0, $sgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE64-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_v2s16_sgpr_sgpr_sgpr
|
|
; WAVE32: liveins: $sgpr0, $sgpr1
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_32 = COPY $sgpr0
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_32 = COPY $sgpr1
|
|
; WAVE32-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32 = S_OR_B32 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
%0:sgpr(<2 x s16>) = COPY $sgpr0
|
|
%1:sgpr(<2 x s16>) = COPY $sgpr1
|
|
%2:sgpr(<2 x s16>) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
---
|
|
|
|
name: or_v2s32_sgpr_sgpr_sgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE64-LABEL: name: or_v2s32_sgpr_sgpr_sgpr
|
|
; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
|
|
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_v2s32_sgpr_sgpr_sgpr
|
|
; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
|
|
; WAVE32-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
%0:sgpr(<2 x s32>) = COPY $sgpr0_sgpr1
|
|
%1:sgpr(<2 x s32>) = COPY $sgpr2_sgpr3
|
|
%2:sgpr(<2 x s32>) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
---
|
|
|
|
name: or_v4s16_sgpr_sgpr_sgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE64-LABEL: name: or_v4s16_sgpr_sgpr_sgpr
|
|
; WAVE64: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
|
|
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_v4s16_sgpr_sgpr_sgpr
|
|
; WAVE32: liveins: $sgpr0_sgpr1, $sgpr2_sgpr3
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:sreg_64 = COPY $sgpr0_sgpr1
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:sreg_64 = COPY $sgpr2_sgpr3
|
|
; WAVE32-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64 = S_OR_B64 [[COPY]], [[COPY1]], implicit-def dead $scc
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
%0:sgpr(<4 x s16>) = COPY $sgpr0_sgpr1
|
|
%1:sgpr(<4 x s16>) = COPY $sgpr2_sgpr3
|
|
%2:sgpr(<4 x s16>) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
---
|
|
|
|
name: or_s32_vgpr_vgpr_vgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1
|
|
; WAVE64-LABEL: name: or_s32_vgpr_vgpr_vgpr
|
|
; WAVE64: liveins: $vgpr0, $vgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE64-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_s32_vgpr_vgpr_vgpr
|
|
; WAVE32: liveins: $vgpr0, $vgpr1
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE32-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
|
|
%0:vgpr(s32) = COPY $vgpr0
|
|
%1:vgpr(s32) = COPY $vgpr1
|
|
%2:vgpr(s32) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
---
|
|
|
|
name: or_v2s16_vgpr_vgpr_vgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1
|
|
; WAVE64-LABEL: name: or_v2s16_vgpr_vgpr_vgpr
|
|
; WAVE64: liveins: $vgpr0, $vgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE64-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
|
|
;
|
|
; WAVE32-LABEL: name: or_v2s16_vgpr_vgpr_vgpr
|
|
; WAVE32: liveins: $vgpr0, $vgpr1
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE32-NEXT: [[V_OR_B32_e64_:%[0-9]+]]:vgpr_32 = V_OR_B32_e64 [[COPY]], [[COPY1]], implicit $exec
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[V_OR_B32_e64_]]
|
|
%0:vgpr(<2 x s16>) = COPY $vgpr0
|
|
%1:vgpr(<2 x s16>) = COPY $vgpr1
|
|
%2:vgpr(<2 x s16>) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
|
|
# This should fail to select
|
|
---
|
|
|
|
name: or_s64_vgpr_vgpr_vgpr
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; WAVE64-LABEL: name: or_s64_vgpr_vgpr_vgpr
|
|
; WAVE64: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
|
|
; WAVE64-NEXT: [[OR:%[0-9]+]]:vgpr(s64) = G_OR [[COPY]], [[COPY1]]
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[OR]](s64)
|
|
;
|
|
; WAVE32-LABEL: name: or_s64_vgpr_vgpr_vgpr
|
|
; WAVE32: liveins: $vgpr0_vgpr1, $vgpr2_vgpr3
|
|
; WAVE32-NEXT: {{ $}}
|
|
; WAVE32-NEXT: [[COPY:%[0-9]+]]:vgpr(s64) = COPY $vgpr0_vgpr1
|
|
; WAVE32-NEXT: [[COPY1:%[0-9]+]]:vgpr(s64) = COPY $vgpr2_vgpr3
|
|
; WAVE32-NEXT: [[OR:%[0-9]+]]:vgpr(s64) = G_OR [[COPY]], [[COPY1]]
|
|
; WAVE32-NEXT: S_ENDPGM 0, implicit [[OR]](s64)
|
|
%0:vgpr(s64) = COPY $vgpr0_vgpr1
|
|
%1:vgpr(s64) = COPY $vgpr2_vgpr3
|
|
%2:vgpr(s64) = G_OR %0, %1
|
|
S_ENDPGM 0, implicit %2
|
|
...
|
|
|
|
---
|
|
|
|
name: or_s1_vcc_copy_to_vcc
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1
|
|
; WAVE64-LABEL: name: or_s1_vcc_copy_to_vcc
|
|
; WAVE64: liveins: $vgpr0, $vgpr1
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; WAVE64-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
|
|
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
|
|
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
;
|
|
; PREGFX11-LABEL: name: or_s1_vcc_copy_to_vcc
|
|
; PREGFX11: liveins: $vgpr0, $vgpr1
|
|
; PREGFX11-NEXT: {{ $}}
|
|
; PREGFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; PREGFX11-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; PREGFX11-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; PREGFX11-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; PREGFX11-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
|
|
; PREGFX11-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
|
|
; PREGFX11-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; PREGFX11-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
;
|
|
; GFX11-TRUE16-LABEL: name: or_s1_vcc_copy_to_vcc
|
|
; GFX11-TRUE16: liveins: $vgpr0, $vgpr1
|
|
; GFX11-TRUE16-NEXT: {{ $}}
|
|
; GFX11-TRUE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-TRUE16-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; GFX11-TRUE16-NEXT: [[COPY2:%[0-9]+]]:vgpr_16 = COPY [[COPY]].lo16
|
|
; GFX11-TRUE16-NEXT: [[COPY3:%[0-9]+]]:vgpr_16 = COPY [[COPY1]].lo16
|
|
; GFX11-TRUE16-NEXT: [[V_AND_B16_t16_e64_:%[0-9]+]]:vgpr_16 = V_AND_B16_t16_e64 0, 1, 0, [[COPY2]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[V_CMP_NE_U16_t16_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U16_t16_e64 0, 0, 0, [[V_AND_B16_t16_e64_]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[V_AND_B16_t16_e64_1:%[0-9]+]]:vgpr_16 = V_AND_B16_t16_e64 0, 1, 0, [[COPY3]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[V_CMP_NE_U16_t16_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U16_t16_e64 0, 0, 0, [[V_AND_B16_t16_e64_1]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U16_t16_e64_]], [[V_CMP_NE_U16_t16_e64_1]], implicit-def dead $scc
|
|
; GFX11-TRUE16-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
;
|
|
; GFX11-FAKE16-LABEL: name: or_s1_vcc_copy_to_vcc
|
|
; GFX11-FAKE16: liveins: $vgpr0, $vgpr1
|
|
; GFX11-FAKE16-NEXT: {{ $}}
|
|
; GFX11-FAKE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-FAKE16-NEXT: [[COPY1:%[0-9]+]]:vgpr_32 = COPY $vgpr1
|
|
; GFX11-FAKE16-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[V_AND_B32_e32_1:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY1]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_1]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; GFX11-FAKE16-NEXT: S_ENDPGM 0, implicit [[S_OR_B32_]]
|
|
%0:vgpr(s32) = COPY $vgpr0
|
|
%1:vgpr(s32) = COPY $vgpr1
|
|
%2:vgpr(s1) = G_TRUNC %0
|
|
%3:vgpr(s1) = G_TRUNC %1
|
|
%4:vcc(s1) = COPY %2
|
|
%5:vcc(s1) = COPY %3
|
|
%6:vcc(s1) = G_OR %4, %5
|
|
S_ENDPGM 0, implicit %6
|
|
...
|
|
|
|
# The selector for the copy of the or result may constrain the result
|
|
# register of the or, losing that it is a VCCRegBank context.
|
|
|
|
# Works for wave32, should fail for wave64
|
|
---
|
|
name: copy_select_constrain_vcc_result_reg_wave32
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $sgpr0
|
|
|
|
; WAVE64-LABEL: name: copy_select_constrain_vcc_result_reg_wave32
|
|
; WAVE64: liveins: $vgpr0, $sgpr0
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B64_]]
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[COPY1]]
|
|
;
|
|
; PREGFX11-LABEL: name: copy_select_constrain_vcc_result_reg_wave32
|
|
; PREGFX11: liveins: $vgpr0, $sgpr0
|
|
; PREGFX11-NEXT: {{ $}}
|
|
; PREGFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; PREGFX11-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; PREGFX11-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; PREGFX11-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; PREGFX11-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; PREGFX11-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; PREGFX11-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; PREGFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B32_]]
|
|
; PREGFX11-NEXT: S_ENDPGM 0, implicit [[COPY1]]
|
|
;
|
|
; GFX11-TRUE16-LABEL: name: copy_select_constrain_vcc_result_reg_wave32
|
|
; GFX11-TRUE16: liveins: $vgpr0, $sgpr0
|
|
; GFX11-TRUE16-NEXT: {{ $}}
|
|
; GFX11-TRUE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-TRUE16-NEXT: [[COPY1:%[0-9]+]]:vgpr_16 = COPY [[COPY]].lo16
|
|
; GFX11-TRUE16-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; GFX11-TRUE16-NEXT: [[V_AND_B16_t16_e64_:%[0-9]+]]:vgpr_16 = V_AND_B16_t16_e64 0, 1, 0, [[COPY1]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[V_CMP_NE_U16_t16_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U16_t16_e64 0, 0, 0, [[V_AND_B16_t16_e64_]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; GFX11-TRUE16-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U16_t16_e64_]], [[V_CMP_NE_U32_e64_]], implicit-def dead $scc
|
|
; GFX11-TRUE16-NEXT: [[COPY2:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B32_]]
|
|
; GFX11-TRUE16-NEXT: S_ENDPGM 0, implicit [[COPY2]]
|
|
;
|
|
; GFX11-FAKE16-LABEL: name: copy_select_constrain_vcc_result_reg_wave32
|
|
; GFX11-FAKE16: liveins: $vgpr0, $sgpr0
|
|
; GFX11-FAKE16-NEXT: {{ $}}
|
|
; GFX11-FAKE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-FAKE16-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; GFX11-FAKE16-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; GFX11-FAKE16-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; GFX11-FAKE16-NEXT: [[COPY1:%[0-9]+]]:sreg_32_xm0 = COPY [[S_OR_B32_]]
|
|
; GFX11-FAKE16-NEXT: S_ENDPGM 0, implicit [[COPY1]]
|
|
%1:vgpr(s32) = COPY $vgpr0
|
|
%0:vgpr(s1) = G_TRUNC %1(s32)
|
|
%sgpr0:sgpr(s32) = COPY $sgpr0
|
|
%2:sgpr(s1) = G_TRUNC %sgpr0
|
|
%6:sgpr(s32) = G_CONSTANT i32 0
|
|
%7:sgpr(p1) = G_IMPLICIT_DEF
|
|
%9:vcc(s1) = COPY %0(s1)
|
|
%10:vcc(s1) = COPY %2(s1)
|
|
%8:vcc(s1) = G_OR %9, %10
|
|
%3:sreg_32_xm0(s1) = COPY %8(s1)
|
|
S_ENDPGM 0, implicit %3
|
|
|
|
...
|
|
|
|
# Works for wave64, should fail for wave32
|
|
---
|
|
name: copy_select_constrain_vcc_result_reg_wave64
|
|
legalized: true
|
|
regBankSelected: true
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $sgpr0
|
|
|
|
; WAVE64-LABEL: name: copy_select_constrain_vcc_result_reg_wave64
|
|
; WAVE64: liveins: $vgpr0, $sgpr0
|
|
; WAVE64-NEXT: {{ $}}
|
|
; WAVE64-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; WAVE64-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; WAVE64-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; WAVE64-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; WAVE64-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_64_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; WAVE64-NEXT: [[S_OR_B64_:%[0-9]+]]:sreg_64_xexec = S_OR_B64 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; WAVE64-NEXT: S_ENDPGM 0, implicit [[S_OR_B64_]]
|
|
;
|
|
; PREGFX11-LABEL: name: copy_select_constrain_vcc_result_reg_wave64
|
|
; PREGFX11: liveins: $vgpr0, $sgpr0
|
|
; PREGFX11-NEXT: {{ $}}
|
|
; PREGFX11-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; PREGFX11-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; PREGFX11-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; PREGFX11-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; PREGFX11-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; PREGFX11-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; PREGFX11-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; PREGFX11-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_OR_B32_]]
|
|
; PREGFX11-NEXT: S_ENDPGM 0, implicit [[COPY1]]
|
|
;
|
|
; GFX11-TRUE16-LABEL: name: copy_select_constrain_vcc_result_reg_wave64
|
|
; GFX11-TRUE16: liveins: $vgpr0, $sgpr0
|
|
; GFX11-TRUE16-NEXT: {{ $}}
|
|
; GFX11-TRUE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-TRUE16-NEXT: [[COPY1:%[0-9]+]]:vgpr_16 = COPY [[COPY]].lo16
|
|
; GFX11-TRUE16-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; GFX11-TRUE16-NEXT: [[V_AND_B16_t16_e64_:%[0-9]+]]:vgpr_16 = V_AND_B16_t16_e64 0, 1, 0, [[COPY1]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[V_CMP_NE_U16_t16_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U16_t16_e64 0, 0, 0, [[V_AND_B16_t16_e64_]], 0, implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; GFX11-TRUE16-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; GFX11-TRUE16-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U16_t16_e64_]], [[V_CMP_NE_U32_e64_]], implicit-def dead $scc
|
|
; GFX11-TRUE16-NEXT: [[COPY2:%[0-9]+]]:sreg_64_xexec = COPY [[S_OR_B32_]]
|
|
; GFX11-TRUE16-NEXT: S_ENDPGM 0, implicit [[COPY2]]
|
|
;
|
|
; GFX11-FAKE16-LABEL: name: copy_select_constrain_vcc_result_reg_wave64
|
|
; GFX11-FAKE16: liveins: $vgpr0, $sgpr0
|
|
; GFX11-FAKE16-NEXT: {{ $}}
|
|
; GFX11-FAKE16-NEXT: [[COPY:%[0-9]+]]:vgpr_32 = COPY $vgpr0
|
|
; GFX11-FAKE16-NEXT: %sgpr0:sreg_32 = COPY $sgpr0
|
|
; GFX11-FAKE16-NEXT: [[V_AND_B32_e32_:%[0-9]+]]:vgpr_32 = V_AND_B32_e32 1, [[COPY]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[V_CMP_NE_U32_e64_:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[V_AND_B32_e32_]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[S_AND_B32_:%[0-9]+]]:sreg_32 = S_AND_B32 1, %sgpr0, implicit-def dead $scc
|
|
; GFX11-FAKE16-NEXT: [[V_CMP_NE_U32_e64_1:%[0-9]+]]:sreg_32_xm0_xexec = V_CMP_NE_U32_e64 0, [[S_AND_B32_]], implicit $exec
|
|
; GFX11-FAKE16-NEXT: [[S_OR_B32_:%[0-9]+]]:sreg_32_xm0_xexec = S_OR_B32 [[V_CMP_NE_U32_e64_]], [[V_CMP_NE_U32_e64_1]], implicit-def dead $scc
|
|
; GFX11-FAKE16-NEXT: [[COPY1:%[0-9]+]]:sreg_64_xexec = COPY [[S_OR_B32_]]
|
|
; GFX11-FAKE16-NEXT: S_ENDPGM 0, implicit [[COPY1]]
|
|
%1:vgpr(s32) = COPY $vgpr0
|
|
%0:vgpr(s1) = G_TRUNC %1(s32)
|
|
%sgpr0:sgpr(s32) = COPY $sgpr0
|
|
%2:sgpr(s1) = G_TRUNC %sgpr0
|
|
%6:sgpr(s32) = G_CONSTANT i32 0
|
|
%7:sgpr(p1) = G_IMPLICIT_DEF
|
|
%9:vcc(s1) = COPY %0(s1)
|
|
%10:vcc(s1) = COPY %2(s1)
|
|
%8:vcc(s1) = G_OR %9, %10
|
|
%3:sreg_64_xexec(s1) = COPY %8(s1)
|
|
S_ENDPGM 0, implicit %3
|
|
|
|
...
|