[Feature] Add --gpu-max-threads-per-block=1024

8864e4f7 · zhuwenwen · 21723d7a · 8864e4f7
Commit 8864e4f7 authored Apr 29, 2026 by zhuwenwen
Hide whitespace changes
Inline Side-by-side

Showing with 2 additions and 1 deletion

cmake/utils.cmake cmake/utils.cmake +2 -1

No files found.
--- a/cmake/utils.cmake
+++ b/cmake/utils.cmake
@@ -123,7 +123,8 @@ function (get_torch_gpu_compiler_flags OUT_GPU_FLAGS GPU_LANG)
      "-U__HIP_NO_HALF_CONVERSIONS__"
      "-U__HIP_NO_HALF_OPERATORS__"
      "-Werror=unused-variable"
-      "-fno-gpu-rdc")
+      "-fno-gpu-rdc"
+      "--gpu-max-threads-per-block=1024")
  endif()
  set(${OUT_GPU_FLAGS} ${GPU_FLAGS} PARENT_SCOPE)