[Metal] Use safe block_size in accumulate#36
Merged
anicusan merged 1 commit intoJuliaGPU:mainfrom May 20, 2025
Merged
Conversation
Member
|
Oof, it really can't be simple, can it... Thanks for this. It means we'll need to do some smarter querying later on |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
1024 is the hard maximum number of threads per threadgroup, but the actual maximum depends on the
maxTotalThreadsPerThreadgroupproperty of the kernel'sMTLComputePipelineState. This makes always attempting to create a block/threadgroup with 1024 will cause errors in some situations.This PR sets
block_sizeto 256 like with all the other algorithms as a safe value.As an example, the last failure in JuliaGPU/Metal.jl#590 is because of this.