[llvm-branch-commits] [llvm] [AMDGPU] Add HWUI pressure heuristics to coexec strategy (PR #184929)

via llvm-branch-commits llvm-branch-commits at lists.llvm.org
Thu Mar 5 19:10:33 PST 2026


github-actions[bot] wrote:

<!--PREMERGE ADVISOR COMMENT: Windows-->
# :window: Windows x64 Test Results

* 131299 tests passed
* 2962 tests skipped
* 2 tests failed

## Failed Tests
(click on a test name to see its output)

### LLVM
<details>
<summary>LLVM.CodeGen/AMDGPU/coexec-sched-effective-stall.mir</summary>

```
Exit Code: 1

Command Output (stdout):
--
# RUN: at line 2
c:\_work\llvm-project\llvm-project\build\bin\llc.exe -mtriple=amdgcn -mcpu=gfx1250 -run-pass=machine-scheduler -verify-misched C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir -o - | c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe -check-prefix=DEFAULT C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\llc.exe' -mtriple=amdgcn -mcpu=gfx1250 -run-pass=machine-scheduler -verify-misched 'C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir' -o -
# note: command had no output on stdout or stderr
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe' -check-prefix=DEFAULT 'C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir'
# note: command had no output on stdout or stderr
# RUN: at line 3
c:\_work\llvm-project\llvm-project\build\bin\llc.exe -mtriple=amdgcn -mcpu=gfx1250 -run-pass=machine-scheduler -amdgpu-sched-strategy=coexec -verify-misched C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir -o - | c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe -check-prefix=COEXEC C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\llc.exe' -mtriple=amdgcn -mcpu=gfx1250 -run-pass=machine-scheduler -amdgpu-sched-strategy=coexec -verify-misched 'C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir' -o -
# note: command had no output on stdout or stderr
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe' -check-prefix=COEXEC 'C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir'
# .---command stderr------------
# | C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir:40:17: error: COEXEC-NEXT: is not on the line after the previous match
# |  ; COEXEC-NEXT: [[DEF5:%[0-9]+]]:vreg_512_align2 = IMPLICIT_DEF
# |                 ^
# | <stdin>:143:2: note: 'next' match was here
# |  %5:vreg_512_align2 = IMPLICIT_DEF
# |  ^
# | <stdin>:141:33: note: previous match ended here
# |  %4:vgpr_32_lo256 = IMPLICIT_DEF
# |                                 ^
# | <stdin>:142:1: note: non-matching line after previous match is here
# |  early-clobber %13:vreg_256_align2 = V_WMMA_SCALE_F32_16X16X128_F8F6F4_f8_f8_w32_threeaddr %0, %1, 0, %2, %3, %4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec
# | ^
# | 
# | Input file: <stdin>
# | Check file: C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-sched-effective-stall.mir
# | 
# | -dump-input=help explains the following input dump.
# | 
# | Input was:
# | <<<<<<
# |          .
# |          .
# |          .
# |        138:  %1:vreg_512_align2 = IMPLICIT_DEF 
# |        139:  %2:vreg_256_align2 = IMPLICIT_DEF 
# |        140:  %3:vgpr_32_lo256 = IMPLICIT_DEF 
# |        141:  %4:vgpr_32_lo256 = IMPLICIT_DEF 
# |        142:  early-clobber %13:vreg_256_align2 = V_WMMA_SCALE_F32_16X16X128_F8F6F4_f8_f8_w32_threeaddr %0, %1, 0, %2, %3, %4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, implicit $exec 
# |        143:  %5:vreg_512_align2 = IMPLICIT_DEF 
# | next:40      !~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  error: match on wrong line
# |        144:  %6:vreg_512_align2 = IMPLICIT_DEF 
# |        145:  %7:vreg_256_align2 = IMPLICIT_DEF 
# |        146:  %8:vgpr_32_lo256 = IMPLICIT_DEF 
# |        147:  %9:vgpr_32_lo256 = IMPLICIT_DEF 
# |        148:  %10:vreg_64_align2 = IMPLICIT_DEF 
# |          .
# |          .
# |          .
# | >>>>>>
# `-----------------------------
# error: command failed with exit status: 1

--

```
</details>
<details>
<summary>LLVM.CodeGen/AMDGPU/coexec-scheduler.ll</summary>

```
Exit Code: 1

Command Output (stdout):
--
# RUN: at line 2
c:\_work\llvm-project\llvm-project\build\bin\llc.exe -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -amdgpu-sched-strategy=coexec --verify-misched  < C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-scheduler.ll | c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe -check-prefix=COEXEC C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-scheduler.ll
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\llc.exe' -mtriple=amdgcn-amd-amdhsa -mcpu=gfx1250 -amdgpu-sched-strategy=coexec --verify-misched
# note: command had no output on stdout or stderr
# executed command: 'c:\_work\llvm-project\llvm-project\build\bin\filecheck.exe' -check-prefix=COEXEC 'C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-scheduler.ll'
# .---command stderr------------
# | C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-scheduler.ll:276:16: error: COEXEC-NEXT: is not on the line after the previous match
# | ; COEXEC-NEXT: v_mov_b32_e32 v29, v0
# |                ^
# | <stdin>:204:2: note: 'next' match was here
# |  v_mov_b32_e32 v29, v0
# |  ^
# | <stdin>:201:50: note: previous match ended here
# |  v_dual_mov_b32 v27, v0 :: v_dual_mov_b32 v28, v0
# |                                                  ^
# | <stdin>:202:1: note: non-matching line after previous match is here
# |  s_wait_kmcnt 0x0
# | ^
# | 
# | Input file: <stdin>
# | Check file: C:\_work\llvm-project\llvm-project\llvm\test\CodeGen\AMDGPU\coexec-scheduler.ll
# | 
# | -dump-input=help explains the following input dump.
# | 
# | Input was:
# | <<<<<<
# |           .
# |           .
# |           .
# |         199:  v_dual_mov_b32 v23, v0 :: v_dual_mov_b32 v24, v0 
# |         200:  v_dual_mov_b32 v25, v0 :: v_dual_mov_b32 v26, v0 
# |         201:  v_dual_mov_b32 v27, v0 :: v_dual_mov_b32 v28, v0 
# |         202:  s_wait_kmcnt 0x0 
# |         203:  s_bitcmp1_b32 s0, 0 
# |         204:  v_mov_b32_e32 v29, v0 
# | next:276      !~~~~~~~~~~~~~~~~~~~~  error: match on wrong line
# |         205:  s_cselect_b32 s0, -1, 0 
# |         206:  v_mov_b32_e32 v30, v0 
# |         207:  s_xor_b32 s0, s0, -1 
# |         208:  v_mov_b32_e32 v31, v0 
# |         209:  v_cndmask_b32_e64 v32, 0, 1, s0 
# |           .
# |           .
# |           .
# | >>>>>>
# `-----------------------------
# error: command failed with exit status: 1

--

```
</details>

If these failures are unrelated to your changes (for example tests are broken or flaky at HEAD), please open an issue at https://github.com/llvm/llvm-project/issues and add the `infrastructure` label.

https://github.com/llvm/llvm-project/pull/184929


More information about the llvm-branch-commits mailing list