jpountz commented on PR #14017:
URL: https://github.com/apache/lucene/pull/14017#issuecomment-2497337140
With Combined tasks in the file (call is 3-polymorphic in the baseline,
bimorphic in the modified version):
```
TaskQPS baseline StdDevQPS
my_m
jpountz opened a new pull request, #14017:
URL: https://github.com/apache/lucene/pull/14017
We currently use `SlowImpactsEnum` for terms whose `docFreq` is less than
128 because it's convenient as these terms don't have impacts anyway. But a
recent slowdown on nightly benchmarks suggests th
jpountz commented on PR #13999:
URL: https://github.com/apache/lucene/pull/13999#issuecomment-2497242523
I opened #14017.
--
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
viliam-durina opened a new pull request, #14016:
URL: https://github.com/apache/lucene/pull/14016
This PR moves the locking logic from `HnswConcurrentMergeBuilder` to
`HnswGraphBuilder`, which automatically picks the single-threaded vs.
concurrent searcher based on whether a lock is used. T
ChrisHegarty opened a new issue, #14020:
URL: https://github.com/apache/lucene/issues/14020
Fails with: `java.lang.AssertionError: expected:<5> but was:<3>`
Reproduces with:
```
./gradlew test --tests
TestSoftDeletesDirectoryReaderWrapper.testAvoidWrappingReadersWithoutSoftDelet
msfroh commented on PR #13987:
URL: https://github.com/apache/lucene/pull/13987#issuecomment-2498931094
Closing in favor of https://github.com/apache/lucene/pull/14012
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use th
msfroh closed pull request #13987: Update lastDoc in ScoreCachingWrappingScorer
URL: https://github.com/apache/lucene/pull/13987
--
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.
viswanathk commented on issue #14002:
URL: https://github.com/apache/lucene/issues/14002#issuecomment-2498525370
> Interesting. The "theoretical" max depth of this stack would be the size
of this graph. I suppose the stack does get large, which would explain a high
no. of Array::grow calls?
msfroh commented on PR #14012:
URL: https://github.com/apache/lucene/pull/14012#issuecomment-2499037808
Thanks folks! I made the suggested change and took this PR out of draft.
I think this one makes more sense than my previous one over at
https://github.com/apache/lucene/pull/13987
jpountz opened a new pull request, #14021:
URL: https://github.com/apache/lucene/pull/14021
Currently, `WANDSCorer` considers that a hit is a match if the sum of
maximum scores across clauses is more than or equal to the minimum competitive
score. We can do better by computing scores of lea
jpountz commented on issue #14004:
URL: https://github.com/apache/lucene/issues/14004#issuecomment-2497680059
> Or does that 50% check trump the flooring?
It does trump the flooring indeed. The reasoning is that even if a user is
happy to spend lots of hardware resources on merging, t
jpountz commented on PR #14021:
URL: https://github.com/apache/lucene/pull/14021#issuecomment-2497708293
```
TaskQPS baseline StdDevQPS
my_modified_version StdDevPct diff p-value
OrStopWords 33.45 (6.
viliam-durina opened a new pull request, #14018:
URL: https://github.com/apache/lucene/pull/14018
(no comment)
--
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 unsubscrib
jpountz commented on issue #14004:
URL: https://github.com/apache/lucene/issues/14004#issuecomment-2498203861
> But this doesn't seem intrinsic/important -- +1 to allow it to pick a
range of segments to merge at once?
I ran some quick simulations with maxMergeAtOnce > segsPerTier and
jpountz merged PR #14017:
URL: https://github.com/apache/lucene/pull/14017
--
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: issues-unsubscr...@lucene.apa
jpountz merged PR #13989:
URL: https://github.com/apache/lucene/pull/13989
--
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: issues-unsubscr...@lucene.apa
jpountz commented on PR #13989:
URL: https://github.com/apache/lucene/pull/13989#issuecomment-2498257183
Thank you!
--
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 unsub
jpountz commented on PR #13987:
URL: https://github.com/apache/lucene/pull/13987#issuecomment-2498260663
> In that case, the unit test that I added can be removed
This works for me. Sorry for putting you on the wrong track by suggesting
that a test is added, it took me a while to unde
aleboulanger commented on issue #3304:
URL: https://github.com/apache/lucene/issues/3304#issuecomment-2498305210
any news please ?
--
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 com
jpountz commented on code in PR #14012:
URL: https://github.com/apache/lucene/pull/14012#discussion_r1856772113
##
lucene/core/src/java/org/apache/lucene/search/ScoreCachingWrappingScorer.java:
##
@@ -31,8 +31,7 @@
*/
public final class ScoreCachingWrappingScorer extends Scor
jpountz commented on PR #13999:
URL: https://github.com/apache/lucene/pull/13999#issuecomment-2497224592
So it went from bimorphic to polymorphic indeed, we had 2 iterator impls:
`Lucene101PostingsReader$BlockImpactsDocsEnum` and `SlowImpactsEnum` and we
added `DisjunctionDISIApproximation`
dungba88 commented on PR #14009:
URL: https://github.com/apache/lucene/pull/14009#issuecomment-2499786440
I have a preliminary benchmark here (top-k=100, fanout=0) using Cohere 768
dataset.
shatejas commented on code in PR #14009:
URL: https://github.com/apache/lucene/pull/14009#discussion_r1857976423
##
lucene/core/src/java/org/apache/lucene/search/RerankKnnFloatVectorQuery.java:
##
@@ -0,0 +1,117 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under o
23 matches
Mail list logo
