Repository navigation
fix: Vector API sums match Rust (float order, I64 prefix overflow) - #545
Merged
Merged
Conversation
Float sums add sequentially: lane-wise reduceLanes(ADD) has no defined order. The I64 sum vectorizes only when every element is in [-2^k, 2^k) with n * 2^k <= 2^63, so no prefix can overflow; otherwise it runs the checked per-element loop, dropping [MAX, 1, -1] as Rust does. The range test uses shift/xor/or because NEON has no 64-bit min/max (the min/max variant ran 2x slower than scalar; this one is within 10%). Closes #535 Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Closes #535.
reduceLanes(ADD)has no defined order).I64sum vectorizes only when every element is in[-2^k, 2^k)withn * 2^k <= 2^63, so no prefix can overflow; otherwise the checked per-element loop runs ([MAX, 1, -1]is dropped, as in Rust). The range test uses shift/xor/or because NEON has no 64-bit min/max.U64unchanged (prefix overflow == total overflow for unsigned).🤖 Generated with Claude Code