[Spec][DSV4] perf: Compile draft hc_head during CUDA graph capture - #32220
Open
mattteochen wants to merge 36 commits into
Open
mattteochen wants to merge 36 commits into
mattteochen wants to merge 36 commits into
Conversation
Contributor
|
Caution The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased. |
mattteochen
changed the title
mattteochen
marked this pull request as ready for review
July 23, 2026 14:07
Contributor
|
Caution The consumer version of Gemini Code Assist on GitHub has been sunset. All code review activity has officially ceased. |
5 tasks
41 tasks
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
The DSV4 MTP draft model's
hc_headruns as a sequence of small eager kernels during CUDA graph capture, adding launch overhead in the low-token-count regime targeted by speculative decoding.Modifications
Reuse the shared PyTorch
hc_head_torchimplementation and compile it during CUDA graph capture on CUDA. Preserve the eager path outside capture and for empty or non-CUDA inputs.CI States
Latest PR Test (Base): ❌ Run #30799521326
Latest PR Test (Extra): ❌ Run #30799521089