Skip to content

Dsv4p add sm120 recipe-Rtx Pro 6000 - #724

Open
yiminghub2024 wants to merge 4 commits into
vllm-project:mainfrom
yiminghub2024:dsv4p-sm120
Open

Dsv4p add sm120 recipe-Rtx Pro 6000#724
yiminghub2024 wants to merge 4 commits into
vllm-project:mainfrom
yiminghub2024:dsv4p-sm120

Conversation

@yiminghub2024

Copy link
Copy Markdown
Contributor

Dsv4p add sm120(rtx pro 6000) recipe

Added RTX PRO 6000 8x configuration with details.

Signed-off-by: yiminghub2024 <482890@qq.com>
Signed-off-by: yiminghub2024 <482890@qq.com>
Updated deployment strategies and hardware configurations for DeepSeek V4 Pro, including support and overrides for GPU configurations.

Signed-off-by: yiminghub2024 <482890@qq.com>
@vercel

vercel Bot commented Aug 1, 2026

Copy link
Copy Markdown
Contributor

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated (UTC)
vllm-recipes Ready Ready Preview Aug 1, 2026 2:56pm

Request Review

@gemini-code-assist gemini-code-assist Bot left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Code Review

This pull request introduces support for the RTX PRO 6000 Blackwell Server Edition (8×96GB, SM120) by adding a new hardware profile in taxonomy.yaml and configuring its deployment strategies, hardware overrides, and guide instructions in models/deepseek-ai/DeepSeek-V4-Pro.yaml. Feedback on the changes points out a critical compatibility issue: the --linear-backend deep_gemm argument specified under the blackwell_sm120 overrides is not supported on SM120 architectures and will cause runtime failures, so it should be removed.

Comment thread models/deepseek-ai/DeepSeek-V4-Pro.yaml
Signed-off-by: yiminghub2024 <482890@qq.com>
@yiminghub2024

Copy link
Copy Markdown
Contributor Author

@esmeetu hi ,please take a look , thanks

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

1 participant