334 lines
22 KiB
YAML
334 lines
22 KiB
YAML
# NOTE: Assertions have been autogenerated by utils/update_mir_test_checks.py UTC_ARGS: --version 5
|
|
# RUN: llc -mtriple=amdgcn-amd-amdhsa -mcpu=gfx942 -run-pass=greedy,amdgpu-rewrite-agpr-copy-mfma -o - %s | FileCheck %s
|
|
|
|
---
|
|
name: test_rewrite_mfma_src2_is_subreg_0
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_is_subreg_0
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
undef %5.sub0_sub1:areg_128_align2 = COPY %4
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, %5
|
|
GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
SI_RETURN
|
|
...
|
|
|
|
---
|
|
name: test_rewrite_mfma_src2_is_subreg_1
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_is_subreg_1
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
undef %5.sub0_sub1:areg_128_align2 = COPY %4
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, %5
|
|
GLOBAL_STORE_DWORDX4 %0, %5, 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
SI_RETURN
|
|
...
|
|
|
|
---
|
|
name: test_rewrite_mfma_src2_is_subreg_chain_mfma_full_def
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_is_subreg_chain_mfma_full_def
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:areg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: dead %other_use:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_1:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_1]]
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3735561 /* reguse:AReg_64_Align2 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
undef %4.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use:vreg_64_align2 = COPY %4.sub0_sub1
|
|
%5:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
%6:areg_64_align2 = COPY %5
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3735561 /* reguse:AReg_64_Align2 */, %6:areg_64_align2
|
|
GLOBAL_STORE_DWORDX2 %0, %6, 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
SI_RETURN
|
|
...
|
|
|
|
---
|
|
name: test_rewrite_mfma_src2_is_subreg_chain_mfma_sub_def
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_is_subreg_chain_mfma_sub_def
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:areg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: dead %other_use:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
|
|
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_1:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]], 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_2:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_2]]
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3735561 /* reguse:AReg_64_Align2 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use:vreg_64_align2 = COPY %4
|
|
undef %5.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4, 0, 0, 0, implicit $mode, implicit $exec
|
|
%6:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %5.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
%7:areg_64_align2 = COPY %6
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3735561 /* reguse:AReg_64_Align2 */, %7
|
|
GLOBAL_STORE_DWORDX2 %0, %7, 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
SI_RETURN
|
|
|
|
...
|
|
|
|
---
|
|
name: test_rewrite_mfma_src2_is_subreg_chain_mfma_sub_def_copy_subreg_insert
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_is_subreg_chain_mfma_sub_def_copy_subreg_insert
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:areg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_1:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]], 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: dead %other_use:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_2:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: undef [[COPY3:%[0-9]+]].sub0_sub1:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_2]]
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
|
|
undef %5.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use:vreg_64_align2 = COPY %5.sub0_sub1
|
|
%6:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %5.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
undef %8.sub0_sub1:areg_128_align2 = COPY %6
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, %8:areg_128_align2
|
|
GLOBAL_STORE_DWORDX4 %0, %8, 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
SI_RETURN
|
|
|
|
...
|
|
|
|
---
|
|
name: test_rewrite_mfma_src2_is_subreg_chain_mfma_sub_def_copy_subreg_extract
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_is_subreg_chain_mfma_sub_def_copy_subreg_extract
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX2_:%[0-9]+]]:areg_64_align2 = GLOBAL_LOAD_DWORDX2 [[COPY]], 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX2_]], 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0
|
|
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_1:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]], 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: dead %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_2:%[0-9]+]]:areg_64_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_1]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: [[COPY3:%[0-9]+]]:agpr_32 = COPY [[V_MFMA_F64_4X4X4F64_e64_2]]
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2162697 /* reguse:AGPR_32 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORD [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s32), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_64_align2 = GLOBAL_LOAD_DWORDX2 %0, 0, 0, implicit $exec :: (load (s64), addrspace 1)
|
|
%4:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use0:vreg_64_align2 = COPY %4.sub0
|
|
undef %5.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use1:vreg_64_align2 = COPY %5.sub0_sub1
|
|
%6:vreg_64_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %5.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
%8:agpr_32 = COPY %6
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 2162697 /* reguse:AGPR_32 */, %8:agpr_32
|
|
GLOBAL_STORE_DWORD %0, %8, 0, 0, implicit $exec :: (store (s32), addrspace 1)
|
|
SI_RETURN
|
|
|
|
...
|
|
|
|
# Multiple MFMAs in the chain use the same register with different
|
|
# subregisters.
|
|
---
|
|
name: test_rewrite_mfma_src2_chain_different_subregs_same_reg_full_copy_use
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_chain_different_subregs_same_reg_full_copy_use
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = IMPLICIT_DEF
|
|
; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: dead %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
|
|
; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
|
|
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_128_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]]
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX4 [[COPY]], [[COPY3]], 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
%4.sub0_sub1 = IMPLICIT_DEF
|
|
%other_use0:vreg_64_align2 = COPY %4.sub0_sub1
|
|
%4.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use1:vreg_64_align2 = COPY %4.sub2_sub3
|
|
%other_use2:vreg_64 = COPY %4.sub1_sub2
|
|
%6:areg_128_align2 = COPY %4
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 6291465 /* reguse:AReg_128_Align2 */, %6:areg_128_align2
|
|
GLOBAL_STORE_DWORDX4 %0, %6, 0, 0, implicit $exec :: (store (s128), addrspace 1)
|
|
SI_RETURN
|
|
...
|
|
|
|
# Multiple MFMAs in the chain use the same register with different
|
|
# subregisters.
|
|
---
|
|
name: test_rewrite_mfma_src2_chain_different_subregs_same_reg_extract_use_0
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_chain_different_subregs_same_reg_extract_use_0
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = IMPLICIT_DEF
|
|
; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
|
|
; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
|
|
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3473417 /* reguse:AReg_64 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
%4.sub0_sub1 = IMPLICIT_DEF
|
|
%other_use0:vreg_64_align2 = COPY %4.sub0_sub1
|
|
%4.sub0_sub1:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use1:vreg_64_align2 = COPY %4.sub2_sub3
|
|
%other_use2:vreg_64 = COPY %4.sub1_sub2
|
|
%6:areg_64 = COPY %4.sub1_sub2
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3473417 /* reguse:AReg_64 */, %6:areg_64
|
|
GLOBAL_STORE_DWORDX2 %0, %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
SI_RETURN
|
|
...
|
|
|
|
---
|
|
name: test_rewrite_mfma_src2_chain_different_subregs_same_reg_extract_use_1
|
|
tracksRegLiveness: true
|
|
body: |
|
|
bb.0:
|
|
liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
|
|
; CHECK-LABEL: name: test_rewrite_mfma_src2_chain_different_subregs_same_reg_extract_use_1
|
|
; CHECK: liveins: $vgpr0, $vgpr1, $vgpr2, $vgpr3, $vgpr4, $vgpr5
|
|
; CHECK-NEXT: {{ $}}
|
|
; CHECK-NEXT: [[COPY:%[0-9]+]]:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
; CHECK-NEXT: [[COPY1:%[0-9]+]]:av_64_align2 = COPY $vgpr0_vgpr1
|
|
; CHECK-NEXT: [[COPY2:%[0-9]+]]:av_64_align2 = COPY $vgpr2_vgpr3
|
|
; CHECK-NEXT: [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]]:areg_128_align2 = GLOBAL_LOAD_DWORDX4 [[COPY]], 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
; CHECK-NEXT: undef [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[GLOBAL_LOAD_DWORDX4_]].sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: [[V_MFMA_F64_4X4X4F64_e64_:%[0-9]+]].sub0_sub1:areg_128_align2 = IMPLICIT_DEF
|
|
; CHECK-NEXT: dead [[GLOBAL_LOAD_DWORDX4_:%[0-9]+]].sub2_sub3:areg_128_align2 = V_MFMA_F64_4X4X4F64_e64 [[COPY1]], [[COPY2]], [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
; CHECK-NEXT: dead %other_use0:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub0_sub1
|
|
; CHECK-NEXT: %other_use1:vreg_64_align2 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub2_sub3
|
|
; CHECK-NEXT: dead %other_use2:vreg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
|
|
; CHECK-NEXT: [[COPY3:%[0-9]+]]:areg_64 = COPY [[V_MFMA_F64_4X4X4F64_e64_]].sub1_sub2
|
|
; CHECK-NEXT: INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3473417 /* reguse:AReg_64 */, [[COPY3]]
|
|
; CHECK-NEXT: GLOBAL_STORE_DWORDX2 [[COPY]], %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
; CHECK-NEXT: SI_RETURN
|
|
%0:vreg_64_align2 = COPY $vgpr4_vgpr5
|
|
%1:av_64_align2 = COPY $vgpr0_vgpr1
|
|
%2:av_64_align2 = COPY $vgpr2_vgpr3
|
|
%3:vreg_128_align2 = GLOBAL_LOAD_DWORDX4 %0, 0, 0, implicit $exec :: (load (s128), addrspace 1)
|
|
undef %4.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %3.sub0_sub1, 0, 0, 0, implicit $mode, implicit $exec
|
|
%4.sub0_sub1 = IMPLICIT_DEF
|
|
%3.sub2_sub3:vreg_128_align2 = V_MFMA_F64_4X4X4F64_vgprcd_e64 %1, %2, %4.sub2_sub3, 0, 0, 0, implicit $mode, implicit $exec
|
|
%other_use0:vreg_64_align2 = COPY %4.sub0_sub1
|
|
%other_use1:vreg_64_align2 = COPY %4.sub2_sub3
|
|
%other_use2:vreg_64 = COPY %4.sub1_sub2
|
|
%6:areg_64 = COPY %4.sub1_sub2
|
|
INLINEASM &"; use $0", 1 /* sideeffect attdialect */, 3473417 /* reguse:AReg_64 */, %6:areg_64
|
|
GLOBAL_STORE_DWORDX2 %0, %other_use1, 0, 0, implicit $exec :: (store (s64), addrspace 1)
|
|
SI_RETURN
|
|
...
|