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

[CPU] Upgrade CPU backend to torch-2.6 (#13381)

Browse Source 
Signed-off-by: jiang1.li <jiang1.li@intel.com>
Co-authored-by: Isotr0py <2037008807@qq.com>
This commit is contained in:
Li, Jiang
2025-03-12 18:41:13 +08:00
committed by GitHub
parent debd6bbf09
commit ff47aab056
9 changed files with 23 additions and 13 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
.buildkite/run-cpu-test.sh
View File

@@ -19,13 +19,14 @@ remove_docker_container
# 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 --network host -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-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" cpu-test-"$BUILDKITE_BUILD_NUMBER"
docker run -itd --entrypoint /bin/bash -v ~/.cache/huggingface:/root/.cache/huggingface --cpuset-cpus="$CORE_RANGE" \
--cpuset-mems="$NUMA_NODE" --privileged=true --network host -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --shm-size=4g --name cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2-"$NUMA_NODE" cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2
--cpuset-mems="$NUMA_NODE" --privileged=true -e HF_TOKEN --env VLLM_CPU_KVCACHE_SPACE=4 --shm-size=4g --name cpu-test-"$BUILDKITE_BUILD_NUMBER"-avx2-"$NUMA_NODE" cpu-test-"$BUILDKITE_BUILD_NUMBER"-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-"$NUMA_NODE" bash -c "
@@ -36,6 +37,7 @@ function cpu_tests() {
docker exec cpu-test-"$BUILDKITE_BUILD_NUMBER"-"$NUMA_NODE" bash -c "
set -e
pip install -r vllm/requirements/test.txt
pip install -r vllm/requirements/cpu.txt
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
@@ -85,4 +87,4 @@ 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"
timeout 40m bash -c "cpu_tests $CORE_RANGE $NUMA_NODE $BUILDKITE_BUILD_NUMBER"