[SW-238029] Fix max_batch_size handling - Lllama perf degradation fix#1839
Merged
bartekkuncer merged 1 commit intohabana_mainfrom Aug 29, 2025
Merged
[SW-238029] Fix max_batch_size handling - Lllama perf degradation fix#1839bartekkuncer merged 1 commit intohabana_mainfrom
bartekkuncer merged 1 commit intohabana_mainfrom
Conversation
Llama Perf degradation seen with Gemma3 suport: #1616. : max_batch_size was initialized incorrectly for the profile_run due to mm_registry checking instead of actual multimodal models. Fix to only initialized to 1 when multimodal(mrope or mm_optimized) model is in use. Llama v3.1 70B 2048/128 BF16 2xcard - perf drop 170 tps to 150 tps. With this fix, it's back to 170tps --------- Co-authored-by: Iryna Boiko <iboiko@habana.ai>
jiminha
requested review from
PatrykWo,
afierka-intel,
deepvars,
jikunshang,
kzawora-intel,
madamczyk-intel,
mgawarkiewicz-intel,
michalkuligowski,
mswiniarsk,
vivekgoe and
wpyszka
as code owners
|
/run-gaudi-tests |
michalkuligowski
approved these changes
Aug 28, 2025
shepark
pushed a commit
that referenced
this pull request
Aug 31, 2025
…#1839) Llama Perf degradation seen with Gemma3 suport: #1616. max_batch_size was initialized incorrectly for the profile_run due to mm_registry checking instead of actual multimodal models. Fix to only initialized to 1 when multimodal(mrope or mm_optimized) model is in use. Llama v3.1 70B 2048/128 BF16 2xcard - perf drop 170 tps to 150 tps. With this fix, it's back to 170tps Co-authored-by: Iryna Boiko <iboiko@habana.ai>
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
4 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Llama Perf degradation seen with Gemma3 suport: #1616.
max_batch_size was initialized incorrectly for the profile_run due to mm_registry checking instead of actual multimodal models. Fix to only initialized to 1 when multimodal(mrope or mm_optimized) model is in use.
Llama v3.1 70B 2048/128 BF16 2xcard - perf drop 170 tps to 150 tps.
With this fix, it's back to 170tps