Repository navigation
fix: avoid 4-bit IVF-PQ accumulator overflow - #111
Merged
Merged
Conversation
Use exact f32 scanning when the number of subquantizers can overflow a u16 accumulator. Apply the bound to row-major, transposed, and FastScan layouts before integer accumulation begins. Cover m=256, 258, and 512 with exactly representable vectors and a prefix whose distance range matches the LUT, independently of range calibration. Exercise in-memory, budgeted, persisted, and batch searches. Validation: Rust 1.95 core tests (497 passed, 2 ignored), rustfmt, and Clippy.
adrian-wang
force-pushed
the
fix-ivfpq-4bit-overflow
branch
from
October 9, 2026 07:26
66fe7f1 to
f1c2a78
Compare
Contributor
|
+1 |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Use exact f32 scanning when the number of subquantizers can overflow a u16 accumulator. Apply the bound to row-major, transposed, and FastScan layouts before integer accumulation begins.
Cover m=256, 258, and 512 with exactly representable vectors and a prefix whose distance range matches the LUT, independently of range calibration. Exercise in-memory, budgeted, persisted, and batch searches.
Validation: Rust 1.95 core tests (497 passed, 2 ignored), rustfmt, and Clippy.