airborne12 opened a new pull request, #70374:
URL: https://github.com/apache/doris/pull/70374
### What problem does this PR solve?
Issue Number: None
Related PR: #68758 (added the shared run for SNII)
Problem Summary: Every scanner of a split segment evaluates the segment's
index predicates over the whole segment, so a 6.3M-row segment cut at 2M rows
per scanner runs them three times. CLucene and BKD readers now share concurrent
identical queries as SNII readers already do, and search() shares its
whole-segment result the same way. At 64 clients a search() workload served 139
instead of 57 queries per second.
### Release note
Queries with inverted index predicates use less CPU when a large segment is
read by several scanners.
### Check List (For Author)
- Test <!-- At least one of them must be included. -->
- [x] Regression test
- [x] Unit Test
- [x] Manual test (add detailed scripts or steps below)
- [ ] No need to test or manual test. Explain why:
- [ ] This is a refactor/code format and no logic has been changed.
- [ ] Previous test can cover this change.
- [ ] No code files have been changed.
- [ ] Other reason <!-- Add your reason? -->
- Behavior changed:
- [ ] No.
- [x] Yes. CLucene and BKD queries now share concurrent identical runs,
as SNII ones do.
- Does this need documentation?
- [x] No.
- [ ] Yes.
### Check List (For Reviewer who merge this PR)
- [ ] Confirm the release note
- [ ] Confirm test cases
- [ ] Confirm document
- [ ] Add branch pick label
🤖 Generated with [Claude Code](https://claude.com/claude-code)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
