zhaih commented on code in PR #13124:
URL: https://github.com/apache/lucene/pull/13124#discussion_r1503101967
##
lucene/core/src/java/org/apache/lucene/index/ConcurrentMergeScheduler.java:
##
@@ -910,4 +936,58 @@ public void setSuppressExceptions(ConcurrentMergeScheduler
cms)
mccullocht commented on PR #13133:
URL: https://github.com/apache/lucene/pull/13133#issuecomment-1964602685
@rmuir @uschindler Would it be sufficient to add a factory function to
`VectorizationProvider` for this like `GroupVIntUtil.Decoder
createGroupVIntDecoder(MemorySegment segment)`? I
stefanvodita commented on PR #13017:
URL: https://github.com/apache/lucene/pull/13017#issuecomment-1964882814
Thanks for updating the PR. Before merging, can you also add an entry to
[CHANGES.txt, as a bug fix under
uschindler commented on PR #13133:
URL: https://github.com/apache/lucene/pull/13133#issuecomment-1964649237
Hi, you can try it out. Two important comments:
- In Java 21, MemorySegment is still a "preview" class, so it can't be used
in any public method signatures. So actually to pass the
gsmiller commented on PR #12559:
URL: https://github.com/apache/lucene/pull/12559#issuecomment-1964636964
Closing this out since we fixed the underlying root cause last year (sorry,
don't have the PR in front of me but can dig it up if necessary).
--
This is an automated message from the
gsmiller closed pull request #12559: Choose sparse values in IntTaxonomyFacets
when FacetsCollector has em…
URL: https://github.com/apache/lucene/pull/12559
--
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
uschindler commented on PR #13133:
URL: https://github.com/apache/lucene/pull/13133#issuecomment-1964672015
> I'm surprised by how slow this is with AVX off given that this can be
implemented with SSE2 :(.
This is why we try to avoid the incubating vector API as much as possible.
dweiss commented on PR #13108:
URL: https://github.com/apache/lucene/pull/13108#issuecomment-1965293218
Just FYI - I've returned to trying to add a nightly smoketester workflow
check. The main branch required some changes to the script as well, which sort
of proves Rob's point that we need
dweiss commented on PR #13108:
URL: https://github.com/apache/lucene/pull/13108#issuecomment-1965313190
I don't think it's going to be so easy - some things just didn't work for me
with local artifacts. I'll follow up, perhaps tomorrow.
--
This is an automated message from the Apache Git
hossman closed issue #13105: (Byte|Float)KnnVectorFieldSource unusable if
segment has no vector values
URL: https://github.com/apache/lucene/issues/13105
--
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
uschindler commented on PR #13108:
URL: https://github.com/apache/lucene/pull/13108#issuecomment-1965306394
I checked the script. Forward port should be more or less:
- copy file to main branch
- change BASE_VERSION constant in the script
The new script is no longer a search
rmuir commented on PR #13133:
URL: https://github.com/apache/lucene/pull/13133#issuecomment-1965433852
> I'm surprised by how slow this is with AVX off given that this can be
implemented with SSE2 :(.
Yes, it is surprising: we found the same situation with VectorUtil byte[]
methods.
dweiss commented on PR #13108:
URL: https://github.com/apache/lucene/pull/13108#issuecomment-1965938522
I think we can merge this?
--
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
stefanvodita commented on PR #13108:
URL: https://github.com/apache/lucene/pull/13108#issuecomment-1964103920
> If you run with Lucene 9.10.0 (just released last week), it should work.
I think so, but the 9.10.0 RC artefacts were no longer available at
msokolov commented on PR #13124:
URL: https://github.com/apache/lucene/pull/13124#issuecomment-1964229191
ooh exciting! I left some comments in a related issue that were maybe a
little clueless given all the discussion here that I missed until now. Still
I'm happy about the direction this
stefanvodita commented on code in PR #13108:
URL: https://github.com/apache/lucene/pull/13108#discussion_r1502581266
##
dev-tools/scripts/smokeTestRelease.py:
##
@@ -633,9 +636,12 @@ def verifyUnpacked(java, artifact, unpackPath,
gitRevision, version, testArgs):
stefanvodita commented on code in PR #13108:
URL: https://github.com/apache/lucene/pull/13108#discussion_r1502581745
##
dev-tools/scripts/smokeTestRelease.py:
##
@@ -911,33 +917,45 @@ def crawl(downloadedFiles, urlString, targetDir,
exclusions=set()):
rmuir commented on PR #13133:
URL: https://github.com/apache/lucene/pull/13133#issuecomment-1964303622
Hi, a couple suggestions:
1. Somehow, we need to avoid Vector API code inside the MemorySegment code.
Just because MemorySegment is available, does not mean Vector API is usable,
uschindler commented on PR #13133:
URL: https://github.com/apache/lucene/pull/13133#issuecomment-1964318304
I have to agree with Robert. We can't glue that code inside
MemorySegmentIndexInput.
The only way to do this is to move the vector optimized decoding to a
special
19 matches
Mail list logo