recipes: add DeepSeek-V4 GB200 decode/prefill bench yamls#124
Open
esmeetu wants to merge 3 commits intoNVIDIA:mainfrom
Open
recipes: add DeepSeek-V4 GB200 decode/prefill bench yamls#124esmeetu wants to merge 3 commits intoNVIDIA:mainfrom
esmeetu wants to merge 3 commits intoNVIDIA:mainfrom
Conversation
Adds three new benchmark recipes under recipes/vllm/deepseek-v4-pro/GB200/8k1k/: - decode-bench-gb200-dep8.yaml: decode-only DEP=8 with DecodeBenchConnector - decode-bench-gb200-tep8.yaml: decode-only TEP=8 sweep - prefill-bench-gb200-dep8.yaml: prefill-only DEP=8
esmeetu
requested review from
hjjq,
kedarpotdar-nv,
kyleliang-nv and
qiching
as code owners
|
FYI @alec-flowers - not meant to be merged but these are rather some data points on prefill setup & decode setups @ certain concurrencies. |
…lization - max-num-seqs: 400 -> 384 - max-cudagraph-capture-size: 400 -> 384 - gpu-memory-utilization: 0.98 -> 0.96
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
decode-bench-gb200-dep8.yaml(decode-only DEP=8, megamoe, concurrency sweep 1024/2048/3072).decode-bench-gb200-tep8.yaml(decode-only TEP=8, megamoe, concurrency sweep 128/256/512).prefill-bench-gb200-dep8.yaml(prefill-only DEP8 flashinfer_nvlink_one_sided).All three live under
recipes/vllm/deepseek-v4-pro/GB200/8k1k/, alongside the existing disagg recipes.Test plan
srtctl dry-run -f recipes/vllm/deepseek-v4-pro/GB200/8k1k/decode-bench-gb200-dep8.yamlsrtctl dry-run -f recipes/vllm/deepseek-v4-pro/GB200/8k1k/decode-bench-gb200-tep8.yamlsrtctl dry-run -f recipes/vllm/deepseek-v4-pro/GB200/8k1k/prefill-bench-gb200-dep8.yaml