[PD Disaggregation] Add unittest for splitwise deployment with using rdma (#5189)
Some checks failed
CE Compile Job / ce_job_pre_check (push) Has been cancelled
CE Compile Job / print_ce_job_pre_check_outputs (push) Has been cancelled
CE Compile Job / FD-Clone-Linux (push) Has been cancelled
CE Compile Job / Show Code Archive Output (push) Has been cancelled
CE Compile Job / BUILD_SM8090 (push) Has been cancelled
CE Compile Job / BUILD_SM8689 (push) Has been cancelled
CE Compile Job / CE_UPLOAD (push) Has been cancelled
Deploy GitHub Pages / deploy (push) Has been cancelled

* Add splitwise deployment with using rdma
* clean cuda
This commit is contained in:
Juncai
2025-11-27 14:27:17 +08:00
committed by GitHub
parent 373b5c3807
commit ce9a49f6bf
9 changed files with 723 additions and 139 deletions

View File

@@ -105,6 +105,7 @@ jobs:
FD_CACHE_QUEUE_PORT=$((8098 + DEVICE_PORT * 100))
FD_ROUTER_PORT=$((8048 + DEVICE_PORT * 100))
FD_CONNECTOR_PORT=$((8038 + DEVICE_PORT * 100))
FD_RDMA_PORT=$((8028 + DEVICE_PORT * 100))
echo "Test ENV Parameter:"
echo "========================================================="
echo "FLASK_PORT=${FLASK_PORT}"
@@ -114,6 +115,7 @@ jobs:
echo "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}"
echo "FD_ROUTER_PORT=${FD_ROUTER_PORT}"
echo "FD_CONNECTOR_PORT=${FD_CONNECTOR_PORT}"
echo "FD_RDMA_PORT=${FD_RDMA_PORT}"
echo "DEVICES=${DEVICES}"
echo "========================================================="
@@ -149,9 +151,15 @@ jobs:
docker rm -f ${runner_name} || true
fi
export RDMA_DEVICES=$(find /dev/infiniband/uverbs* -maxdepth 1 -not -type d | xargs -I{} echo '--device {}:{}')
docker run --rm --net=host \
--name ${runner_name} \
--cap-add=SYS_PTRACE --shm-size=64G \
--cap-add=SYS_PTRACE --cap-add=IPC_LOCK \
--shm-size=64G \
${RDMA_DEVICES} \
--device=/dev/infiniband/rdma_cm \
--ulimit memlock=-1:-1 \
-v $(pwd):/workspace -w /workspace \
-v "${CACHE_DIR}/gitconfig:/etc/gitconfig:ro" \
-v "${CACHE_DIR}/.cache:/root/.cache" \
@@ -165,6 +173,8 @@ jobs:
-e "FD_CACHE_QUEUE_PORT=${FD_CACHE_QUEUE_PORT}" \
-e "FD_ROUTER_PORT=${FD_ROUTER_PORT}" \
-e "FD_CONNECTOR_PORT=${FD_CONNECTOR_PORT}" \
-e "FD_RDMA_PORT=${FD_RDMA_PORT}" \
-e "CLEAN_CUDA=1" \
-e TZ="Asia/Shanghai" \
-e "fd_wheel_url=${fd_wheel_url}" \
-e "BASE_REF=${BASE_REF}" \