Default to bf16
#465
Merged
Default to bf16
#465
Conversation
This file contains 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
Closes #421 Co-authored-by: jrcavani <[email protected]>
Closes #450 Co-authored-by: Michał Bartoszkiewicz <[email protected]>
This commit exposes the underlying SimSIMD `bf16` kernels for distance functions, that can be up to 3x faster than IEEE half-precision floats for different multiplication- intensive operations. It also brings backwards-compatibility for older CPUs, that don't yet support `bf16`. Important to note, NumPy still doesn't support half-precision, so we can't reliably guess what the name of the `bf16` type will be in the buffer protocol specifications. numpy/numpy#19808
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Outdated suggestions cannot be applied.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.
The new default behavior is to use
bf16when constructing new indices on modern CPUs. This should positively affect both throughput and recall compared tof16.