Updates copy operations to use improved vectorization #44
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
Replaces generic copy struct with AutoVectorizingCopyWithAssumedAlignment for better memory access patterns. Reduces vector layout from 8 to 4 values per read for ZOH and ActiveMask operations to optimize memory bandwidth usage.
Contributor
There was a problem hiding this comment.
Pull Request Overview
This PR updates copy operations to leverage improved vectorization for better memory access patterns. Key changes include:
- Replacing Gmem_copy_struct with AutoVectorizingCopyWithAssumedAlignment for ZOH and ActiveMask copy operations
- Reducing the vector layout from 8 to 4 values per read for ZOH and ActiveMask operations to optimize memory bandwidth usage
Comments suppressed due to low confidence (2)
csrc/src/kernel_traits.h:157
- Consider adding a brief inline comment or documentation note explaining why AutoVectorizingCopyWithAssumedAlignment with a 64-byte alignment is used for ZOH and ActiveMask, especially given that a different alignment (128) is used for the 'O' operation.
make_tiled_copy(Copy_Atom<AutoVectorizingCopyWithAssumedAlignment<64>, Element>{},
csrc/src/kernel_traits.h:159
- Ensure that tests cover the new vector layout configuration to confirm that reducing from 8 to 4 values per read does not introduce unintended behavior.
Layout<Shape<_1, _4>>{})); // Val layout, 4 vals per read
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Replaces generic copy struct with AutoVectorizingCopyWithAssumedAlignment for better memory access patterns.
Reduces vector layout from 8 to 4 values per read for ZOH and ActiveMask operations to optimize memory bandwidth usage.