-
Notifications
You must be signed in to change notification settings - Fork 955
Add CUDA graph capture/replay for qwen 3.5 moe decode method #18809
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Open
Gasoonjia
wants to merge
34
commits into
main
Choose a base branch
from
cuda-graph
base: main
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
Open
Changes from 20 commits
Commits
Show all changes
34 commits
Select commit
Hold shift + click to select a range
5465d8b
Replace chunked FLA with recurrent gated delta rule for T=1 decode
Gasoonjia a6ebe8a
Runtime dispatch: recurrent (T=1) vs chunked (T>1) inside triton_op
Gasoonjia fc5018e
Revert model.py, export.py, main.cpp to main branch
Gasoonjia c90a8e8
Add tests for recurrent (T=1) and multi-T dispatch
Gasoonjia ce3e9ca
lint fix - 2
Gasoonjia 8d35c65
lint fix - 2
Gasoonjia 709deb0
Merge branch 'main' into recurrent-fla
Gasoonjia eff976d
lint fix - 3
Gasoonjia 7dd4280
Optimize recurrent kernel: parallelize over V tiles
Gasoonjia 3a1ee31
Dual-method PTE with GPU-resident state for Qwen3.5 MoE
63c162e
Use share_mutable_buffers to eliminate select_scatter overhead
47d6b98
Merge branch 'main' into recurrent-fla
Gasoonjia 375e5c0
lint
Gasoonjia 2b36797
remove reduntdant updates
Gasoonjia c06d58b
Cross-method AOTI constant sharing for KV cache
6945b2a
Fix cross-method AOTI constant sharing and add dual-method runner
Gasoonjia ea51d0d
Remove debug printf and decode_only flag
Gasoonjia a0a62f1
Lint formatting fixes
Gasoonjia ca69871
Improve CUDA backend error handling and add dual-method runner fallback
7c148f7
Add CUDA graph capture/replay for decode method
ee75c2e
Merge branch 'main' into cuda-graph
Gasoonjia 10e7aad
lint and reformat
Gasoonjia 9042f36
Merge branch 'main' into cuda-graph
Gasoonjia 84d1587
Merge branch 'main' into cuda-graph
Gasoonjia e00a499
solve claude
Gasoonjia aa7bb82
Merge branch 'main' into cuda-graph
Gasoonjia cef386b
Merge branch 'main' into cuda-graph
Gasoonjia 2d32422
Merge branch 'main' into cuda-graph
Gasoonjia 1270870
Merge branch 'main' into cuda-graph
Gasoonjia 8fc7355
solve stride out of scope
Gasoonjia 2c46ed2
Merge branch 'main' into cuda-graph
Gasoonjia 855eb93
Merge branch 'main' into cuda-graph
Gasoonjia 4237d17
remove unused env var
Gasoonjia 9b4705e
Merge branch 'main' into cuda-graph
Gasoonjia File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
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
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
Oops, something went wrong.
Oops, something went wrong.
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.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why does it have to be specified here, can we enable it internally when user specifies
--cuda_graph?