xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-10 00:06:06 +08:00
Code Issues Packages Projects Releases Wiki Activity

[CPU][CI] Re-enable the CPU CI tests (#19046)

Browse Source 
Signed-off-by: jiang.li <jiang1.li@intel.com>
This commit is contained in:
Li, Jiang 2025-06-03 11:46:47 +08:00 committed by GitHub
parent 4ce42f9204
commit 8655f47f37
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
3 changed files with 29 additions and 26 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
.buildkite/scripts/hardware_ci/run-cpu-test.sh
View File

@ -8,67 +8,65 @@ set -ex
CORE_RANGE=${CORE_RANGE:-48-95}
NUMA_NODE=${NUMA_NODE:-1}
export CMAKE_BUILD_PARALLEL_LEVEL=32
# Setup cleanup
remove_docker_container() {
set -e;
docker rm -f cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2-"$NUMA_NODE" || true;
docker image rm cpu-test-"$BUILDKITE_BUILD_NUMBER" cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2 || true;
docker rm -f cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"-avx2 || true;
}
trap remove_docker_container EXIT
remove_docker_container
# Try building the docker image
numactl -C "$CORE_RANGE" -N "$NUMA_NODE" docker build --tag cpu-test-"$BUILDKITE_BUILD_NUMBER" --target vllm-test -f docker/Dockerfile.cpu .
numactl -C "$CORE_RANGE" -N "$NUMA_NODE" docker build --build-arg VLLM_CPU_DISABLE_AVX512="true" --tag cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2 --target vllm-test -f docker/Dockerfile.cpu .
numactl -C "$CORE_RANGE" -N "$NUMA_NODE" docker build --tag cpu-test-"$NUMA_NODE" --target vllm-test -f docker/Dockerfile.cpu .
numactl -C "$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 --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --cpuset-cpus="$CORE_RANGE" \
--cpuset-mems="$NUMA_NODE" --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --shm-size=4g --name cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" cpu-test-"$BUILDKITE_BUILD_NUMBER"
--cpuset-mems="$NUMA_NODE" --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --shm-size=4g --name cpu-test-"$NUMA_NODE" cpu-test-"$NUMA_NODE"
docker run -itd --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --cpuset-cpus="$CORE_RANGE" \
--cpuset-mems="$NUMA_NODE" --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --shm-size=4g --name cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2-"$NUMA_NODE" cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2
--cpuset-mems="$NUMA_NODE" --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --shm-size=4g --name cpu-test-"$NUMA_NODE"-avx2 cpu-test-"$NUMA_NODE"-avx2
function cpu_tests() {
set -e
export NUMA_NODE=$2
export BUILDKITE_BUILD_NUMBER=$3
# offline inference
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2-"$NUMA_NODE" bash -c "
docker exec cpu-test-"$NUMA_NODE"-avx2 bash -c "
set -e
python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m"
# Run basic model test
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
docker exec cpu-test-"$NUMA_NODE" bash -c "
set -e
pytest -v -s tests/kernels/test_cache.py -m cpu_model
pytest -v -s tests/kernels/test_mla_decode_cpu.py -m cpu_model
pytest -v -s tests/models/decoder_only/language -m cpu_model
pytest -v -s tests/models/embedding/language -m cpu_model
pytest -v -s tests/models/encoder_decoder/language -m cpu_model
pytest -v -s tests/models/decoder_only/audio_language -m cpu_model
pytest -v -s tests/models/decoder_only/vision_language -m cpu_model"
pytest -v -s tests/kernels/attention/test_cache.py -m cpu_model
pytest -v -s tests/kernels/attention/test_mla_decode_cpu.py -m cpu_model
pytest -v -s tests/models/language/generation -m cpu_model
pytest -v -s tests/models/language/pooling -m cpu_model
pytest -v -s tests/models/multimodal/generation --ignore=tests/models/multimodal/generation/test_mllama.py -m cpu_model"
# Run compressed-tensor test
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
docker exec cpu-test-"$NUMA_NODE" bash -c "
set -e
pytest -s -v \
tests/quantization/test_compressed_tensors.py::test_compressed_tensors_w8a8_static_setup \
tests/quantization/test_compressed_tensors.py::test_compressed_tensors_w8a8_dynamic_per_token"
# Run AWQ test
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
docker exec cpu-test-"$NUMA_NODE" bash -c "
set -e
pytest -s -v \
tests/quantization/test_ipex_quant.py"
# Run chunked-prefill and prefix-cache test
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
docker exec cpu-test-"$NUMA_NODE" bash -c "
set -e
pytest -s -v -k cpu_model \
tests/basic_correctness/test_chunked_prefill.py"
# online serving
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
docker exec cpu-test-"$NUMA_NODE" bash -c "
set -e
export VLLM_CPU_KVCACHE_SPACE=10
export VLLM_CPU_OMP_THREADS_BIND=$1
@ -83,7 +81,7 @@ function cpu_tests() {
--tokenizer facebook/opt-125m"
# Run multi-lora tests
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
docker exec cpu-test-"$NUMA_NODE" bash -c "
set -e
pytest -s -v \
tests/lora/test_qwen2vl.py"
@ -91,4 +89,4 @@ function cpu_tests() {
# All of CPU tests are expected to be finished less than 40 mins.
export -f cpu_tests
timeout 40m bash -c "cpu_tests $CORE_RANGE $NUMA_NODE $BUILDKITE_BUILD_NUMBER"
timeout 40m bash -c "cpu_tests $CORE_RANGE $NUMA_NODE"

10
docker/Dockerfile.cpu
View File

@ -75,6 +75,7 @@ RUN --mount=type=bind,source=.git,target=.git \
RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=cache,target=/root/.cache/ccache \
--mount=type=cache,target=/workspace/vllm/.deps,sharing=locked \
--mount=type=bind,source=.git,target=.git \
VLLM_TARGET_DEVICE=cpu python3 setup.py bdist_wheel
@ -85,7 +86,7 @@ WORKDIR /workspace/vllm
RUN --mount=type=cache,target=/var/cache/apt,sharing=locked \
--mount=type=cache,target=/var/lib/apt,sharing=locked \
apt-get install -y --no-install-recommends vim numactl
apt-get install -y --no-install-recommends vim numactl xz-utils
# install development dependencies (for testing)
RUN --mount=type=cache,target=/root/.cache/uv \
@ -108,8 +109,11 @@ FROM base AS vllm-test
WORKDIR /workspace/
RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=bind,src=requirements/test.txt,target=requirements/test.txt \
uv pip install -r requirements/test.txt
--mount=type=bind,src=requirements/test.in,target=requirements/test.in \
cp requirements/test.in requirements/test-cpu.in && \
sed -i '/mamba_ssm/d' requirements/test-cpu.in && \
uv pip compile requirements/test-cpu.in -o requirements/cpu-test.txt && \
uv pip install -r requirements/cpu-test.txt
RUN --mount=type=cache,target=/root/.cache/uv \
--mount=type=bind,from=vllm-build,src=/workspace/vllm/dist,target=dist \

3
vllm/distributed/parallel_state.py
View File

@ -1203,7 +1203,8 @@ def cleanup_dist_env_and_memory(shutdown_ray: bool = False):
if empty_cache is not None:
empty_cache()
try:
torch._C._host_emptyCache()
if not current_platform.is_cpu():
torch._C._host_emptyCache()
except AttributeError:
logger.warning(
"torch._C._host_emptyCache() only available in Pytorch >=2.5")