[Optimization] default compile rdma, reduce cudagraph buffer size in mm, fix some config bug (#5121)

* default compile rdma, reduce cudagraph buffer size in mm, fix some config logic

* update

* update

* fix bug

* enhance rdma compile

* fix
This commit is contained in:
Yuanle Liu
2025-11-20 17:19:47 +08:00
committed by GitHub
parent 6fa34102e8
commit 7ac25935c7
8 changed files with 126 additions and 37 deletions

View File

@@ -164,7 +164,7 @@ jobs:
python -m pip install -r requirements.txt
python -m pip install wheel
# 编译RDMA
export ENABLE_FD_RDMA=1
export FD_ENABLE_RDMA_COMPILE=1
bash build.sh 1 python false [${COMPILE_ARCH}]
ls ./dist/*.whl
'