[Feature]: multiple gpus specification

[XYaoooo]

🚀 The feature, motivation and pitch

hi develpers, could vllm support multiple gpus specification like cuda:0,1 (vllm device) in initiation in future.

Alternatives

No response

Additional context

No response

Before submitting a new issue...

• Make sure you already searched for relevant issues, and asked the chatbot living at the bottom right corner of the documentation page, which can answer lots of frequently asked questions.

[DarkLight1337]

You can set CUDA_VISIBLE_DEVICES environment variable to control the GPUs that are detected by vLLM.

[XYaoooo]

You can set CUDA_VISIBLE_DEVICES environment variable to control the GPUs that are detected by vLLM.

Thanks for your advice, I know this can work.
But I personally prefer the way to specify it like "cuda:6,7" (vllm device), more brief.
I am not sure how much work it will bring to developers.

[DarkLight1337]

cc @youkaichao

[youkaichao]

if you are using ray, you can specify it using the env var:

vllm/vllm/envs.py

Line 90 in da833b0

VLLM_RAY_BUNDLE_INDICES: str = ""

[XYaoooo]

if you are using ray, you can specify it using the env var:

vllm/vllm/envs.py

Line 90 in da833b0

VLLM_RAY_BUNDLE_INDICES: str = ""

Thanks for your advice.

[github-actions]

This issue has been automatically marked as stale because it has not had any activity within 90 days. It will be automatically closed if no further activity occurs within 30 days. Leave a comment if you feel this issue should remain open. Thank you!

[github-actions]

This issue has been automatically closed due to inactivity. Please feel free to reopen if you feel it is still relevant. Thank you!