Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Make it easy to profile workers with nsight #3162

Merged
merged 5 commits into from
Mar 4, 2024

Conversation

pcmoritz
Copy link
Collaborator

@pcmoritz pcmoritz commented Mar 3, 2024

This PR makes it easy to profile models with higher tensor parallelism with Ray. It can for example be used with

python benchmarks/benchmark_latency.py --model "mistralai/Mixtral-8x7B-Instruct-v0.1" --tensor-parallel-size 2 --num-iters 10 --ray-workers-use-nsight --input-len 2200 --output-len 200

The profiling data will end up in /tmp/ray/session_latest/logs/nsight

vllm/engine/arg_utils.py Outdated Show resolved Hide resolved
pcmoritz and others added 3 commits March 3, 2024 09:21
Co-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>
Copy link
Member

@zhuohan123 zhuohan123 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! Thanks for the contribution

@zhuohan123 zhuohan123 merged commit 17c3103 into vllm-project:main Mar 4, 2024
22 checks passed
dtransposed pushed a commit to afeldman-nm/vllm that referenced this pull request Mar 26, 2024
Co-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>
Temirulan pushed a commit to Temirulan/vllm-whisper that referenced this pull request Sep 6, 2024
Co-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants