[Hopper] optimize decoding performance for headdim 128 fp8 #96
Conversation
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
Signed-off-by: Jonas Kuebler <[email protected]>
jmkuebler
force-pushed
the
headdim128_fp8_optim
branch
from
af207d5 to
9b9c5f8
Compare
Author
|
@LucasWilkinson could you please check |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 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Outdated suggestions cannot be applied.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.
Similar to #91 this PR optimizes the FP8 tiles sizes for headdim 128. This applies for example to
Qwen3-Coder-30B-A3B-Instructwhere it reduces the context-length dependent inter-token latency by ~1.39x against BF16.With the current mainline config, FP8 decoding is not faster at all than BF16 for Qwen3-Coder-30B-A3B-Instruct.