MLA RoPE + quantization fused kernel: shape generalization for MHA / GQA #1924
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
kahyunnam
force-pushed
the
knam/RoPe-fusion
branch
from
fa658e2 to
bd2a338
Compare
kahyunnam
force-pushed
the
knam/RoPe-fusion
branch
from
bd2a338 to
0fc5d11
Compare
kahyunnam
changed the title
|
@nvpohanh for another set of eyes. |
| TensorView q_nope_out, TensorView k_nope_out, TensorView cos_sin_cache, | ||
| TensorView pos_ids, double quant_scale_q, double quant_scale_kv, | ||
| bool interleave) { | ||
| void rope_quantize(TensorView q_rope_in, TensorView k_rope_in, TensorView q_nope_in, |
There was a problem hiding this comment.
A doc string explaining what the method does would be helpful here.
| if __name__ == "__main__": | ||
| # Run all benchmarks and generate individual plots | ||
| print("Running MLA benchmark...") | ||
| benchmark_mla.run(print_data=False, show_plots=True, save_path=".") |
There was a problem hiding this comment.
Time permitting it would be nice to compare against existing Flashinfer Rope + Quant in torch native as baseline for these measurements
| mutates_args=("q_rope_out", "k_rope_out", "q_nope_out", "k_nope_out"), | ||
| ) | ||
| def _mla_rope_quantize( | ||
| def _rope_quantize( |
There was a problem hiding this comment.
Nit: Please add doc string here as well.
pavanimajety
approved these changes
Oct 16, 2025
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Outdated suggestions cannot be applied.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.
📌 Description
Generalize the existing MLA RoPE+Q fused kernels to support GQA/MHA problem shapes.
🔍 Test Results
pytest -v tests/attention/test_rope.py::test_generalized_rope_quantizepython benchmarks/bench_rope_quantize_fp8.py🚀 Pull Request Checklist
Thank you for contributing to FlashInfer! Before we review your pull request, please make sure the following items are complete.
✅ Pre-commit Checks
pre-commitby runningpip install pre-commit(or used your preferred method).pre-commit install.pre-commit run --all-filesand fixed any reported issues.🧪 Tests
unittest, etc.).