Speed up string view comparison (up to 3x) #9250
Merged
Conversation
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
Contributor
Author
|
run benchmark comparison_kernels |
|
🤖 |
|
🤖: Benchmark completed Details
|
Contributor
Author
Seems a good direction |
Contributor
|
@zhuqi-lucas as you have spent time optimzing string view comparison in other PRs, do you have time to help review this PR as well? |
Contributor
Author
|
run benchmark comparison_kernels |
|
🤖 |
Contributor
Nice improvement, i plan to review this PR in few days. |
|
🤖: Benchmark completed Details
|
Contributor
Author
|
run benchmark comparison_kernels |
|
🤖 |
Contributor
Author
|
run benchmark comparison_kernels |
Contributor
Author
|
run benchmark arrow_reader_clickbench |
Dandandan
changed the title
Dandandan
changed the title
Contributor
Author
I found some more time for improvement 🚀 |
Dandandan
commented
Jan 23, 2026
|
🤖: Benchmark completed Details
|
|
🤖 |
Contributor
Author
|
On my machine (Apple M2) it makes more of a difference, but still overall improvements (also I think the prefix change might make more of a difference.in real life due to avoiding random access). |
|
🤖: Benchmark completed Details
|
|
🤖 |
Contributor
Author
|
|
🤖: Benchmark completed Details
|
Contributor
Author
|
Improvements on clickbench seem to correlate well with |
Contributor
|
Previous optimization attempt, mostly for the empty string case: #7767 |
Contributor
Author
|
run benchmark arrow_reader_clickbench |
|
🤖 |
|
🤖: Benchmark completed Details
|
zhuqi-lucas
approved these changes
Jan 23, 2026
Contributor
zhuqi-lucas
left a comment
zhuqi-lucas
left a comment
There was a problem hiding this comment.
LGTM, great work @Dandandan , the benchmark result is promising, and the code is more clean.
| /// - The combination ensures that a single `u128` comparison correctly orders by inline data | ||
| /// first (lexicographically), then by length (numerically). | ||
| #[inline(always)] | ||
| pub fn inline_key_fast(raw: u128) -> u128 { |
Contributor
There was a problem hiding this comment.
The datafusion seems also use this function, we also need to run datafusion based this change to make sure it's not breaking something.
| if l_len == 0 && r_len == 0 { | ||
|
|
||
| // Both are empty | ||
| if l_len == 0 { |
Contributor
There was a problem hiding this comment.
I found this will compile to same code finally, but it's more clean here, i agree.
Contributor
|
I have a few more tests I would like to propose: |
Dandandan
pushed a commit
that referenced
this pull request
Jan 24, 2026
# Which issue does this PR close? - Follow on to #9250 # Rationale for this change While (posthumously) reviewing #9250 from @Dandandan and @zhuqi-lucas I noticed that some of the special case branches are not covered. # What changes are included in this PR? Add some more tests to cover all the special cases # Are these changes tested? Yes, only tests # Are there any user-facing changes? <!-- If there are user-facing changes then we may require documentation to be updated before approving the PR. If there are any breaking changes to public APIs, please call them out. -->
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
5 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.
Which issue does this PR close?
Rationale for this change
I saw this was a bit hot in profiles.
Details
What changes are included in this PR?
Are these changes tested?
Are there any user-facing changes?