Accelerate The Inference Speed Of Wan2.2 On NVIDIA Thor T5000 - #1489
Open
Michael20070814 wants to merge 1 commit into
Open
Accelerate The Inference Speed Of Wan2.2 On NVIDIA Thor T5000#1489Michael20070814 wants to merge 1 commit into
Michael20070814 wants to merge 1 commit into
Conversation
Add NVFP4 cuBLASLt projections and fused split-N FFN epilogues, CuTeDSL sparse BF16/FP8 attention, and Thor-specific build and configuration support. Preserve upstream interfaces and include regression coverage. Co-Authored-By: Claude Code <noreply@anthropic.com>
Michael20070814
force-pushed
the
try/fuse_qkv
branch
from
September 6, 2026 08:58
5dc2f64 to
2e6a756
Compare
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Replace the triton's self-attention backend with cutsdsl's.
Implement a more efficient quantization kernel for converting FP16 to FP4
Split the FFN to 2 parts
Use the cuBlaslt for the QKV GEMM
Tune the QKV/FFN's GEMM algorithm for the best shape
Before:
After:
