Repository navigation
small floating-point performance improvements - #1068
Merged
Merged
Conversation
Contributor
Author
|
Another instruction-selection improvement from @samth: clear bits before |
Mark `fl-make-rectangular` as accepting unboxed arguments, which is needed to comunicate unboxing potential to `let` bindings. On x86_64, use MOVAPD instead of MOVSD for register-to-register moves, since that avoids creating a false dependency on some processors to preserve high bits, and similarly clear bits of the desitination register before CVTSI2SD and some related instructions. These improvements and related tests were provided by @samth.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Mark
fl-make-rectangularas accepting unboxed arguments, which is needed to comunicate unboxing potential toletbindings.On x86_64, use
movapdinstead ofmovsdfor register-to-register moves, since that avoids creating a false dependency on some processors to preserve high bits.