[CPU] Refine default config for the CPU backend (#19539)

Signed-off-by: jiang1.li <jiang1.li@intel.com>
This commit is contained in:
Li, Jiang
2025-06-13 13:27:39 +08:00
committed by GitHub
parent bb4a0decef
commit 6458721108
4 changed files with 60 additions and 25 deletions
+12 -3
View File
@@ -24,13 +24,22 @@ numactl -C "$CORE_RANGE" -N "$NUMA_NODE" docker build --tag cpu-test-"$NUMA_NODE
numactl -C "$CORE_RANGE" -N "$NUMA_NODE" docker build --build-arg VLLM_CPU_DISABLE_AVX512="true" --tag cpu-test-"$NUMA_NODE"-avx2 --target vllm-test -f docker/Dockerfile.cpu .
# Run the image, setting --shm-size=4g for tensor parallel.
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --env VLLM_CPU_OMP_THREADS_BIND="$OMP_CORE_RANGE" --shm-size=4g --name cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --env VLLM_CPU_OMP_THREADS_BIND="$OMP_CORE_RANGE" --shm-size=4g --name cpu-test-"$NUMA_NODE"-avx2 cpu-test-"$NUMA_NODE"-avx2
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --env VLLM_CPU_OMP_THREADS_BIND="$OMP_CORE_RANGE" --env VLLM_CPU_CI_ENV=1 --shm-size=4g --name cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"
docker run -itd --cpuset-cpus="$CORE_RANGE" --cpuset-mems="$NUMA_NODE" --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --env VLLM_CPU_OMP_THREADS_BIND="$OMP_CORE_RANGE" --env VLLM_CPU_CI_ENV=1 --shm-size=4g --name cpu-test-"$NUMA_NODE"-avx2 cpu-test-"$NUMA_NODE"-avx2
function cpu_tests() {
set -e
export NUMA_NODE=$2
# list packages
docker exec cpu-test-"$NUMA_NODE"-avx2 bash -c "
set -e
pip list"
docker exec cpu-test-"$NUMA_NODE" bash -c "
set -e
pip list"
# offline inference
docker exec cpu-test-"$NUMA_NODE"-avx2 bash -c "
set -e
@@ -72,7 +81,7 @@ function cpu_tests() {
set -e
python3 -m vllm.entrypoints.openai.api_server --model facebook/opt-125m --dtype half &
timeout 600 bash -c 'until curl localhost:8000/v1/models; do sleep 1; done' || exit 1
python3 benchmarks/benchmark_serving.py \
VLLM_CPU_CI_ENV=0 python3 benchmarks/benchmark_serving.py \
--backend vllm \
--dataset-name random \
--model facebook/opt-125m \