[CI/Build] Modify Dockerfile build for ARM64 & GH200 #11302
Closed
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: drikster80 <[email protected]>
Signed-off-by: drikster80 <[email protected]>
Signed-off-by: drikster80 <[email protected]>
…/causal-conv1d/mamba/flashinfer/bitsandbytes
# Conflicts: # docs/source/serving/deploying_with_docker.rst
…l && rm arm64 platform condition in requirements && add python3 use_existing_torch.py before build in run-gh200-test.sh
…& outlines == 0.1.11 # Requires pytorch && rm torch install before docker build
|
👋 Hi! Thank you for contributing to the vLLM project. Once the PR is approved and ready to go, your PR reviewer(s) can run CI to test the changes comprehensively before merging. To run CI, PR reviewers can do one of these:
🚀 |
mergify
bot
added
documentation
cennn
changed the title
|
close as we are testing it in #11351 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
From PR: 10499,11212
Fix Issue: 2021
The Dockerfile build process for ARM64 systems is regarded as a build specific to a particular PyTorch version. Before initiating the build, the use_existing_pytorch.py script is employed. Additionally, torch and torchvision are installed from the nightly build prior to installing any other requirements. This approach maximally avoids the overwrite of torch by other packages listed in the requirements. Besides, this step enhances the consistency between the build executed within the docker environment and the user's manual installation from the source code.
The following command was utilized for the building and has been verified.
python3 use_existing_torch.py && docker build . --target vllm-openai --platform "linux/arm64" -t cenncenn/vllm-gh200-openai:v0.6.4.post1 --build-arg max_jobs=66 --build-arg nvcc_threads=2 --build-arg torch_cuda_arch_list="9.0+PTX" --build-arg vllm_fa_cmake_gpu_arches="90-real" --build-arg RUN_WHEEL_CHECK='false'The following tutorial was employed for building from the source code and has been verified.
use-an-existing-pytorch-installation
The changes have been tested on the Nvidia GH200 platform with models meta-llama/Llama-3.1-8B and Qwen/Qwen2.5-0.5B-Instruct.