[CI/Build] Limit wheel size check to CUDA 13 (#50357)
Signed-off-by: Tyler Michael Smith <tlrmchlsmth@gmail.com> Co-authored-by: OpenAI Codex <codex@openai.com>
This commit is contained in:
co-authored by
OpenAI Codex
parent
e5f48dfda2
commit
1ad5182ba9
@@ -15,7 +15,6 @@ DOCKER_BUILDKIT=1 docker build . \
|
||||
-t gh200-test \
|
||||
--build-arg max_jobs=66 \
|
||||
--build-arg nvcc_threads=2 \
|
||||
--build-arg RUN_WHEEL_CHECK=false \
|
||||
--build-arg torch_cuda_arch_list="9.0+PTX"
|
||||
|
||||
# Setup cleanup
|
||||
|
||||
+3
-2
@@ -585,13 +585,14 @@ COPY --from=extensions-build /tmp/ep_kernels_workspace/dist /tmp/ep_kernels_work
|
||||
RUN sha256sum /tmp/ep_kernels_workspace/dist/*.whl \
|
||||
> /tmp/ep_kernels_workspace/dist/wheels.sha256
|
||||
|
||||
# Check the size of the wheel if RUN_WHEEL_CHECK is true
|
||||
# Check the size of the CUDA 13 wheel uploaded to PyPI
|
||||
COPY .buildkite/check-wheel-size.py check-wheel-size.py
|
||||
# sync the default value with .buildkite/check-wheel-size.py
|
||||
ARG VLLM_MAX_SIZE_MB=500
|
||||
ENV VLLM_MAX_SIZE_MB=$VLLM_MAX_SIZE_MB
|
||||
ARG RUN_WHEEL_CHECK=true
|
||||
RUN if [ "$RUN_WHEEL_CHECK" = "true" ]; then \
|
||||
RUN if [ "$RUN_WHEEL_CHECK" = "true" ] && \
|
||||
[ "${CUDA_VERSION%%.*}" = "13" ]; then \
|
||||
python3 check-wheel-size.py dist; \
|
||||
else \
|
||||
echo "Skipping wheel size check."; \
|
||||
|
||||
@@ -384,8 +384,7 @@ A docker container can be built for aarch64 systems such as the Nvidia Grace-Hop
|
||||
-t vllm/vllm-gh200-openai:latest \
|
||||
--build-arg max_jobs=66 \
|
||||
--build-arg nvcc_threads=2 \
|
||||
--build-arg torch_cuda_arch_list="9.0 10.0+PTX" \
|
||||
--build-arg RUN_WHEEL_CHECK=false
|
||||
--build-arg torch_cuda_arch_list="9.0 10.0+PTX"
|
||||
```
|
||||
|
||||
For (G)B300, we recommend using CUDA 13, as shown in the following command.
|
||||
@@ -398,7 +397,6 @@ For (G)B300, we recommend using CUDA 13, as shown in the following command.
|
||||
--build-arg BUILD_BASE_IMAGE=nvidia/cuda:13.0.2-devel-ubuntu22.04 \
|
||||
--build-arg max_jobs=256 \
|
||||
--build-arg nvcc_threads=2 \
|
||||
--build-arg RUN_WHEEL_CHECK=false \
|
||||
--build-arg torch_cuda_arch_list='9.0 10.0+PTX' \
|
||||
--platform "linux/arm64" \
|
||||
--tag vllm/vllm-gb300-openai:latest \
|
||||
|
||||
Reference in New Issue
Block a user