caio/clang-p2996
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
1daf2994de49d1ecba4bee4e6842aa8a564cbc96
clang-p2996/llvm/test/CodeGen/AMDGPU/merge-image-load.mir
Konrad Kusiak 4fa8a5487e [AMDGPU] Add sanity check that fixes bad shift operation in AMD backend 
There is a problem with the
SILoadStoreOptimizer::dmasksCanBeCombined() function that can lead to
UB.

This boolean function decides if two masks can be combined into 1. The
idea here is that the bits which are "on" in one mask, don't overlap
with the "on" bits of the other. Consider an example (10 bits for
simplicity):

Mask 1: 0101101000
Mask 2: 0000000110

Those can be combined into a single mask: 0101101110.

To check if such an operation is possible, the code takes the mask
which is greater and counts how many 0s there are, starting from the
LSB and stopping at the first 1. Then, it shifts 1u by this number and
compares it with the smaller mask. The problem is that when both masks
are 0, the counter will find 32 zeroes in the first mask and will try
to do a shift by 32 positions which leads to UB.

The fix is a simple sanity check, if the bigger mask is 0 or not.

https://reviews.llvm.org/D155051
2023-08-11 15:26:35 -04:00

488 lines
27 KiB
YAML
Raw Blame History

# RUN: llc -march=amdgcn -mcpu=gfx900 -verify-machineinstrs -run-pass si-load-store-opt -o - %s | FileCheck -check-prefix=GFX9 %s
# GFX9-LABEL: name: image_load_merged_v1v3
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), align 4, addrspace 4)
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
# GFX9: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
name: image_load_merged_v1v3
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 1, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 14, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_merged_v1v3_reversed
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), addrspace 4)
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY %8.sub3
# GFX9: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub0_sub1_sub2
name: image_load_merged_v1v3_reversed
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_merged_v2v2
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_64 = COPY %8.sub0_sub1
# GFX9: %{{[0-9]+}}:vreg_64 = COPY killed %8.sub2_sub3
name: image_load_merged_v2v2
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vreg_64 = IMAGE_LOAD_V2_V4 %5:vreg_128, %3:sgpr_256, 3, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
%7:vreg_64 = IMAGE_LOAD_V2_V4 %5:vreg_128, %3:sgpr_256, 12, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_merged_v2v2_reversed
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_64 = COPY %8.sub2_sub3
# GFX9: %{{[0-9]+}}:vreg_64 = COPY killed %8.sub0_sub1
name: image_load_merged_v2v2_reversed
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vreg_64 = IMAGE_LOAD_V2_V4 %5:vreg_128, %3:sgpr_256, 12, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
%7:vreg_64 = IMAGE_LOAD_V2_V4 %5:vreg_128, %3:sgpr_256, 3, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_merged_v3v1
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = COPY %8.sub0_sub1_sub2
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY killed %8.sub3
name: image_load_merged_v3v1
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
%7:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
...
---
# GFX9-LABEL: name: image_load_merged_v3v1_reversed
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), align 4, addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = COPY %8.sub1_sub2_sub3
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY killed %8.sub0
name: image_load_merged_v3v1_reversed
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 14, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
%7:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 1, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
...
---
# GFX9-LABEL: name: image_load_divided_merged
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), addrspace 4)
name: image_load_divided_merged
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%8:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%9:vreg_96 = IMAGE_LOAD_V3_V4 %7:vreg_128, %3:sgpr_256, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
%10:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%11:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_divided_not_merged
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_divided_not_merged
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vreg_128 = COPY %2
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
IMAGE_STORE_V4_V4 %4:vreg_128, %5:vreg_128, %3:sgpr_256, 15, -1, 0, 0, 0, 0, 0, 0, implicit $exec :: (store (s128))
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_dmask_overlapped_not_merged
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 4, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_dmask_overlapped_not_merged
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 4, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_dmask_zero_not_merged
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 0, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 0, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_dmask_zero_not_merged
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 0, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 0, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_dmask_not_disjoint_not_merged
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 4, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 11, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_dmask_not_disjoint_not_merged
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5:vreg_128, %3:sgpr_256, 4, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5:vreg_128, %3:sgpr_256, 11, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_0
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %6, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_0
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 1, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%7:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%8:vreg_96 = IMAGE_LOAD_V3_V4 %6, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_1
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %6, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %6, %4, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_1
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%5:vgpr_32 = COPY %2.sub3
%6:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%7:vgpr_32 = IMAGE_LOAD_V1_V4 %6, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%8:vreg_96 = IMAGE_LOAD_V3_V4 %6, %4, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_10
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 1, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_10
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 1, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_3
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 1, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_3
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 1, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_4
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 1, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_4
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 1, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_5
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 1, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_5
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 1, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_6
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 1, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_6
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 1, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_7
# GFX9: %{{[0-9]+}}:vreg_64 = IMAGE_LOAD_V2_V4 %5, %3, 8, 0, 0, 0, 1, 0, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_7
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vreg_64 = IMAGE_LOAD_V2_V4 %5, %3, 8, 0, 0, 0, 1, 0, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_8
# GFX9: %{{[0-9]+}}:vreg_64 = IMAGE_LOAD_V2_V4 %5, %3, 8, 0, 0, 0, 0, 1, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_8
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vreg_64 = IMAGE_LOAD_V2_V4 %5, %3, 8, 0, 0, 0, 0, 1, -1, 0, implicit $exec :: (dereferenceable load (s64), align 16, addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_not_merged_9
# GFX9: %{{[0-9]+}}:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
# GFX9: %{{[0-9]+}}:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
name: image_load_not_merged_9
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_V1_V4 %5, %3, 8, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_V3_V4 %5, %3, 7, 0, 0, 0, 0, 0, 0, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_mip_merged_v1v3
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_MIP_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, 0, implicit $exec, implicit $exec :: (dereferenceable load (s128), align 4, addrspace 4)
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
# GFX9: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
name: image_load_mip_merged_v1v3
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_MIP_V1_V4 %5:vreg_128, %3:sgpr_256, 1, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_MIP_V3_V4 %5:vreg_128, %3:sgpr_256, 14, 0, 0, 0, 0, 0, -1, 0, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_mip_pck_merged_v1v3
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_MIP_PCK_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, implicit $exec, implicit $exec :: (dereferenceable load (s128), align 4, addrspace 4)
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
# GFX9: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
name: image_load_mip_pck_merged_v1v3
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_MIP_PCK_V1_V4 %5:vreg_128, %3:sgpr_256, 1, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_MIP_PCK_V3_V4 %5:vreg_128, %3:sgpr_256, 14, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_mip_pck_sgn_merged_v1v3
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_MIP_PCK_SGN_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, implicit $exec, implicit $exec :: (dereferenceable load (s128), align 4, addrspace 4)
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
# GFX9: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
name: image_load_mip_pck_sgn_merged_v1v3
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_MIP_PCK_SGN_V1_V4 %5:vreg_128, %3:sgpr_256, 1, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_MIP_PCK_SGN_V3_V4 %5:vreg_128, %3:sgpr_256, 14, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_pck_merged_v1v3
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_PCK_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, implicit $exec, implicit $exec :: (dereferenceable load (s128), align 4, addrspace 4)
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
# GFX9: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
name: image_load_pck_merged_v1v3
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_PCK_V1_V4 %5:vreg_128, %3:sgpr_256, 1, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_PCK_V3_V4 %5:vreg_128, %3:sgpr_256, 14, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
# GFX9-LABEL: name: image_load_pck_sgn_merged_v1v3
# GFX9: %{{[0-9]+}}:vreg_128 = IMAGE_LOAD_PCK_SGN_V4_V4 %5, %3, 15, 0, 0, 0, 0, 0, -1, implicit $exec, implicit $exec :: (dereferenceable load (s128), align 4, addrspace 4)
# GFX9: %{{[0-9]+}}:vgpr_32 = COPY %8.sub0
# GFX9: %{{[0-9]+}}:vreg_96 = COPY killed %8.sub1_sub2_sub3
name: image_load_pck_sgn_merged_v1v3
body: |
bb.0.entry:
%0:sgpr_64 = COPY $sgpr0_sgpr1
%1:sreg_64_xexec = S_LOAD_DWORDX2_IMM %0, 36, 0
%2:sgpr_128 = COPY $sgpr96_sgpr97_sgpr98_sgpr99
%3:sgpr_256 = S_LOAD_DWORDX8_IMM %1, 208, 0
%4:vgpr_32 = COPY %2.sub3
%5:vreg_128 = BUFFER_LOAD_DWORDX4_OFFSET %2:sgpr_128, 0, 0, 0, 0, implicit $exec :: (dereferenceable invariant load (s128))
%6:vgpr_32 = IMAGE_LOAD_PCK_SGN_V1_V4 %5:vreg_128, %3:sgpr_256, 1, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s32), addrspace 4)
%7:vreg_96 = IMAGE_LOAD_PCK_SGN_V3_V4 %5:vreg_128, %3:sgpr_256, 14, 0, 0, 0, 0, 0, -1, implicit $exec :: (dereferenceable load (s96), align 16, addrspace 4)
...
---
Reference in New Issue View Git Blame Copy Permalink