test: add regression test for ivf/pq search#5476
Merged
westonpace merged 1 commit intolance-format:mainfrom Jan 5, 2026
Merged
test: add regression test for ivf/pq search#5476westonpace merged 1 commit intolance-format:mainfrom
westonpace merged 1 commit intolance-format:mainfrom
Conversation
|
Codex usage limits have been reached for code reviews. Please check with the admins of this repo to increase the limits by adding credits. |
wjones127
reviewed
Dec 15, 2025
Contributor
There was a problem hiding this comment.
Two dimensions not tested here that seem worthwhile are:
num_bitsin PQ (4-bit vs 8-bit)data_typein the vector (f16 vs f32)
These setting will trigger different code paths, which I think are important to cover in regression testing.
There are some setting here like k which I don't think really hit different code paths as much as they just change a variable in the same code path.
Member
Author
There was a problem hiding this comment.
I think rust-level benchmarks would sufficiently (and probably with more fidelity) capture those parameters since they primarily affect the compute driven PQ search time. These parameters would also require training completely new indexes which is what dominates the benchmark time.
Member
Author
There was a problem hiding this comment.
We have pq_dist_table and 4bit_pq_dist_table which cover num_bits and distance_type. However, these only look at f32. I'll add an issue to add data_type to these benchmarks (I want to review them anyways since the lance-index benches don't pass reliably and in sufficient time for me)
wjones127
approved these changes
Dec 15, 2025
Member
Author
|
@bench-bot benchmark |
Member
Author
Benchmark Results for PR #5476Commit: Summary
✅ All Benchmarks Within Normal RangeNo benchmarks had |z-score| > 2.0 All ResultsView all 123 benchmark results
Generated by bench-bot 🤖 |
jackye1995
pushed a commit
to jackye1995/lance
that referenced
this pull request
Jan 21, 2026
The test covers 5 parameters (cache, k, nprobes, refine factor, dataset size) for a total of 64 tests. It takes ~3-4 minutes to run on my system. There are other parameters (distance type, number of dimensions, PQ-v-SQ, etc.) but these should only really affect compute time and/or recall and are probably better tested in rust benchmarks. Recall benchmarks will be done separately as they require real data.
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.
The test covers 5 parameters (cache, k, nprobes, refine factor, dataset size) for a total of 64 tests. It takes ~3-4 minutes to run on my system.
There are other parameters (distance type, number of dimensions, PQ-v-SQ, etc.) but these should only really affect compute time and/or recall and are probably better tested in rust benchmarks. Recall benchmarks will be done separately as they require real data.