biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[CUDA graphs] Enable full cuda graphs with FA3 AoT scheduling (#20301)

Browse Source 
Signed-off-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Woosuk Kwon
2025-07-01 09:07:36 -07:00
committed by GitHub
parent 314af8617c
commit 8acb4badee
2 changed files with 54 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
cmake/external_projects/vllm_flash_attn.cmake
View File

@@ -38,7 +38,7 @@ else()
FetchContent_Declare(
vllm-flash-attn
GIT_REPOSITORY https://github.com/vllm-project/flash-attention.git
GIT_TAG 5f3644181c7a15345ce20bfc65af117d3601b524
GIT_TAG 1c2624e53c078854e0637ee566c72fe2107e75f4
GIT_PROGRESS TRUE
# Don't share the vllm-flash-attn build between build types
BINARY_DIR ${CMAKE_BINARY_DIR}/vllm-flash-attn