Cold-Air-Networks/mirror-llvm-project
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions 7

[openmp][cmake] Match GPU triples consistently in openmp/module (#211138)

Browse source 
Fixes #211135.

`openmp/module/CMakeLists.txt:29` gates the GPU-only Fortran compile
options on `"^amdgcn|^nvptx"`, while `openmp/CMakeLists.txt:176-177`
selects host-vs-device layout for the same build using
`"^amdgpu|^amdgcn|^nvptx|^spirv64"` against both
`LLVM_DEFAULT_TARGET_TRIPLE` and `CMAKE_CXX_COMPILER_TARGET`.

All four offload cache files use the triple `amdgpu-amd-amdhsa`, which
`^amdgcn` does not match, so `-nogpulib -flto` are silently not applied
to `libomp-mod` in the recommended AMDGPU offload configurations.

Compute the test once as `LIBOMP_TARGET_IS_GPU` in
`openmp/CMakeLists.txt` and use it at both sites, rather than
duplicating a widened regex. The conditions differ, not just the
patterns: copying the regex alone would leave a build that sets
`CMAKE_CXX_COMPILER_TARGET` without `LLVM_DEFAULT_TARGET_TRIPLE` still
taking the device path in `openmp/` while `libomp-mod` misses the flags.
`libomp-mod` compiles Fortran, so `CMAKE_Fortran_COMPILER_TARGET` is
also consulted; neither site did before.

Evaluated with CMake's regex engine:

| triple | before | after |
|---|---|---|
| `amdgpu-amd-amdhsa` | 0 | 1 |
| `amdgcn-amd-amdhsa` | 1 | 1 |
| `nvptx64-nvidia-cuda` | 1 | 1 |
| `spirv64-amd-amdhsa` | 0 | 1 |
| `x86_64-unknown-linux-gnu` | 0 | 0 |

`^spirv64` only brings `module/` in line with
`openmp/CMakeLists.txt:176`, which already routes spirv64 to `device/`.
Host builds are unaffected.

Roughly five other sites carry the same `amdgcn`-only pattern
(`offload/CMakeLists.txt:30`, `flang-rt/CMakeLists.txt:126`,
`flang-rt/lib/runtime/CMakeLists.txt:317`,
`flang-rt/cmake/modules/AddFlangRT.cmake:298`,
`cmake/Modules/GetToolchainDirs.cmake:118`) and are left for a
follow-up.

No test: configure-time logic, no test mechanism exists in `openmp/`.

This affects performance-critical applications on large AMD GPU
supercomputers, including [MFC](https://github.com/MFlowCode/MFC).

All numbers above come from the validated reproducers included with this
report and are independently reproducible; they stand on their own.

This was found and root-caused with the assistance of AI tools.
This commit is contained in:
Spencer Bryngelson 2026-07-22 10:49:38 -04:00 committed by GitHub
parent 34436db53d
commit 2f5771c8ec
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 13 additions and 4 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
openmp/CMakeLists.txt
View file

@ -168,13 +168,22 @@ add_custom_target(install-openmp-stripped
add_dependencies(install-openmp openmp)
add_dependencies(install-openmp-stripped openmp)
# Whether this build targets a GPU. "amdgcn" is the legacy spelling of "amdgpu";
# both reach here, and the offload cache files use "amdgpu-amd-amdhsa".
set(LIBOMP_GPU_TRIPLE_REGEX "^amdgpu|^amdgcn|^nvptx|^spirv64")
if("${LLVM_DEFAULT_TARGET_TRIPLE}" MATCHES "${LIBOMP_GPU_TRIPLE_REGEX}" OR
"${CMAKE_CXX_COMPILER_TARGET}" MATCHES "${LIBOMP_GPU_TRIPLE_REGEX}" OR
"${CMAKE_Fortran_COMPILER_TARGET}" MATCHES "${LIBOMP_GPU_TRIPLE_REGEX}")
set(LIBOMP_TARGET_IS_GPU TRUE)
else()
set(LIBOMP_TARGET_IS_GPU FALSE)
endif()
if(LIBOMP_FORTRAN_MODULES)
add_subdirectory(module)
endif()
# Use the current compiler target to determine the appropriate runtime to build.
if("${LLVM_DEFAULT_TARGET_TRIPLE}" MATCHES "^amdgpu|^amdgcn|^nvptx|^spirv64" OR
"${CMAKE_CXX_COMPILER_TARGET}" MATCHES "^amdgpu|^amdgcn|^nvptx|^spirv64")
if(LIBOMP_TARGET_IS_GPU)
add_subdirectory(device)
else()
# Build host runtime library, after LIBOMPTARGET variables are set since they

2
openmp/module/CMakeLists.txt
View file

@ -26,7 +26,7 @@ if (CMAKE_Fortran_COMPILER_ID STREQUAL "GNU")
target_compile_options(libomp-mod PRIVATE -fno-range-check)
endif ()
if ("${LLVM_DEFAULT_TARGET_TRIPLE}" MATCHES "^amdgcn|^nvptx")
if (LIBOMP_TARGET_IS_GPU)
target_compile_options(libomp-mod PRIVATE
$<$<COMPILE_LANGUAGE:Fortran>:-nogpulib -flto>
)