Expose entry hints through launch#27
Merged
maleadt merged 4 commits intoJuliaGPU:mainfrom Jan 13, 2026
Merged
Conversation
maleadt
requested changes
Jan 13, 2026
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Exposes and encodes the
occupancyandnum_ctasentry hints as keyword arguments ofct.launchandcompile.cutile-python exposes defaults through the kernel function decorator, and has an experimental feature for finding the best config from some search space through autotune_launch.
This is used in AttentionFMHA.py, and would be needed for a fair performance comparison.
occupancyhas little effect on a set of simple kernels I have tested it on, and I've only seen decreased performance fromnum_ctas(0.2-0.5x) in vadd and matmul kernels, but it should hopefully make #16 faster.