CMake build elf without PTX (#3739)

51c31bc1 · Simon Mo · GitHub · 3ad438c6 · 51c31bc1
Unverified Commit 51c31bc1 authored Mar 29, 2024 by Simon Mo Committed by GitHub Mar 30, 2024
Show whitespace changes
Inline Side-by-side

Showing with 4 additions and 1 deletion

cmake/utils.cmake cmake/utils.cmake +4 -1

No files found.
--- a/cmake/utils.cmake
+++ b/cmake/utils.cmake
@@ -240,9 +240,12 @@ macro(override_gpu_arches GPU_ARCHES GPU_LANG GPU_SUPPORTED_ARCHES)
      endif()
      if (_SM)
-        set(_VIRT "")
+        # -real suffix let CMake to only generate elf code for the kernels.
+        # we want this, otherwise the added ptx (default) will increase binary size.
+        set(_VIRT "-real")
        set(_CODE_ARCH ${_SM})
      else()
+        # -virtual suffix let CMake to generate ptx code for the kernels.
        set(_VIRT "-virtual")
        set(_CODE_ARCH ${_CODE})
      endif()