[Openmp-commits] [PATCH] D88929: [OpenMP] Change CMake Configuration to Build for Highest CUDA Architecture by Default

Ye Luo via Phabricator via Openmp-commits openmp-commits at lists.llvm.org
Tue Oct 6 20:01:06 PDT 2020


ye-luo requested changes to this revision.
ye-luo added inline comments.
This revision now requires changes to proceed.


================
Comment at: openmp/libomptarget/deviceRTLs/nvptx/CMakeLists.txt:80
+      else()
+        list(APPEND compute_capabilities ${CMAKE_MATCH_1})
+      endif()
----------------
1. Doesn't work right now. Missing comma ",${CMAKE_MATCH_1}"
2. using CUDA_ARCH as "string(REGEX MATCH" output causes problems.
3. "append" needs to protect redundant 35.
4. I think it is better to move this part of logic to `openmp/libomptarget/cmake/Modules/LibomptargetGetDependencies.cmake`
after `find_package(CUDA QUIET)`.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D88929/new/

https://reviews.llvm.org/D88929



More information about the Openmp-commits mailing list