[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:
Tyler Michael Smith
2026-07-29 20:10:42 -07:00
committed by GitHub
co-authored by OpenAI Codex
parent e5f48dfda2
commit 1ad5182ba9
3 changed files with 4 additions and 6 deletions
@@ -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
View File
@@ -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 \