Closed
Conversation
…n some tests locally to make sure things are working, but I havent run benchmarks to identify the critical point, so its likely that the switch to the vectorized version at dim=8 is premature
var77
added a commit
that referenced
this pull request
Oct 8, 2024
* Fix data_producer deadloc, remove unwraps * Bump versions
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Initial implementations of AVX2 vectorized distance functions for linux-x86_64 (L1, L2, L2SQ, cosine). It should be fairly easy to extend it to BSD and intel macs, windows will require a separate implementation. I tested the code, and I'm reasonably confident it works. I have yet to profile it, but I wanted to get eyes on it before going much further. The choice of dim=8 to switch to the vectorized path was convenient for testing but is almost certainly premature. My hope is that this will provide a significant performance increase on vectors of higher dimension. I added a CMake file to detect AVX to support building distfunc.c with avx2 enabled as well which makes up the bulk of this PR