xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-03-20 11:39:12 +08:00
Code Issues Packages Projects Releases Wiki Activity

Merge 1133c53c0e6814d07c2754d9d4049b5583ca5ab1 into 254f6b986720c92ddf97fbb1a6a6465da8e87e29

Browse Source
This commit is contained in:
Michael Goin 2025-12-25 08:09:11 +08:00 committed by GitHub
commit 932e928da7
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 6 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
cmake/external_projects/vllm_flash_attn.cmake
View File

@ -62,6 +62,12 @@ install(CODE "set(CMAKE_INSTALL_PREFIX \"\${CMAKE_INSTALL_PREFIX}/vllm/\")" ALL_
FetchContent_MakeAvailable(vllm-flash-attn)
message(STATUS "vllm-flash-attn is available at ${vllm-flash-attn_SOURCE_DIR}")
# Suppress ptxas warnings for flash attention compilation
if(VLLM_GPU_LANG STREQUAL "CUDA")
# Add ptxas flags to suppress C7520 warnings about wgmma.mma_async serialization
set(CMAKE_CUDA_FLAGS "${CMAKE_CUDA_FLAGS} -Xptxas --diag_suppress=7520")
endif()
# Restore the install prefix
install(CODE "set(CMAKE_INSTALL_PREFIX \"\${OLD_CMAKE_INSTALL_PREFIX}\")" ALL_COMPONENTS)
install(CODE "set(CMAKE_INSTALL_LOCAL_ONLY TRUE)" ALL_COMPONENTS)