Bluetopia opened a new issue, #12666:
URL: https://github.com/apache/lucene/issues/12666
### Description
Using:
https://javadoc.io/doc/org.apache.lucene/lucene-core/latest/org/apache/lucene/analysis/Analyzer.html
which should hopefully be up to date. The current state of the source f
mayya-sharipova commented on code in PR #12582:
URL: https://github.com/apache/lucene/pull/12582#discussion_r1357345297
##
lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99ScalarQuantizedVectorsWriter.java:
##
@@ -0,0 +1,782 @@
+/*
+ * Licensed to the Apache Softwa
benwtrent commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1357376455
##
lucene/core/src/java/org/apache/lucene/codecs/lucene95/IncrementalHnswGraphMerger.java:
##
@@ -0,0 +1,209 @@
+/*
+ * Licensed to the Apache Software Foundation (A
benwtrent commented on code in PR #12582:
URL: https://github.com/apache/lucene/pull/12582#discussion_r1357380984
##
lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99HnswVectorsFormat.java:
##
@@ -0,0 +1,209 @@
+/*
+ * Licensed to the Apache Software Foundation (AS
rmuir opened a new pull request, #12667:
URL: https://github.com/apache/lucene/pull/12667
Following up to @dweiss work, this gives us the same benchmarks as
https://github.com/rmuir/vectorbench, just without the code duplication and
maintenance hassle.
Each method is simply invoked w
rmuir commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1760526957
You can still tell what's happening too due to the log messages. When each
benchmark runs, you see a single message:
xxxScalar() methods:
```
WARNING: Java vector incubator mod
dungba88 commented on code in PR #12624:
URL: https://github.com/apache/lucene/pull/12624#discussion_r1353826324
##
lucene/core/src/java/org/apache/lucene/util/fst/BytesStore.java:
##
@@ -21,19 +21,18 @@
import java.util.List;
import org.apache.lucene.store.DataInput;
import
zhaih commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1357264961
##
lucene/core/src/java/org/apache/lucene/util/hnsw/OnHeapHnswGraph.java:
##
@@ -40,31 +41,29 @@ public final class OnHeapHnswGraph extends HnswGraph
implements Account
xjtushilei commented on PR #12660:
URL: https://github.com/apache/lucene/pull/12660#issuecomment-1760636017
Good Job
--
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 uns
sohami commented on code in PR #12606:
URL: https://github.com/apache/lucene/pull/12606#discussion_r1357659905
##
lucene/core/src/java/org/apache/lucene/search/IndexSearcher.java:
##
@@ -420,13 +418,12 @@ public int count(Query query) throws IOException {
}
/**
- * Ret
gf2121 commented on PR #12661:
URL: https://github.com/apache/lucene/pull/12661#issuecomment-1760704047
It seems i can reproduce the slow down locally with `wikimediumall`, and
this patch do not work :(
I'll keep digging
```
TaskQPS baseline StdD
gf2121 commented on PR #12661:
URL: https://github.com/apache/lucene/pull/12661#issuecomment-1760721571
I made an experiment that run luceneutil without `-r`, which means we read
old format with [candidate
code](https://github.com/apache/lucene/blob/92c13b8e59c5e4cf6de5e49863e0b102d5063987/
jmazanec15 commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1357750648
##
lucene/core/src/java/org/apache/lucene/codecs/lucene95/IncrementalHnswGraphMerger.java:
##
@@ -0,0 +1,209 @@
+/*
+ * Licensed to the Apache Software Foundation (
zhaih commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1357779437
##
lucene/core/src/java/org/apache/lucene/codecs/lucene95/IncrementalHnswGraphMerger.java:
##
@@ -0,0 +1,209 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF)
gf2121 commented on PR #12661:
URL: https://github.com/apache/lucene/pull/12661#issuecomment-1760938463
2000 PK task repeats per JVM:
```
TaskQPS baseline StdDevQPS
my_modified_version StdDevPct diff p-value
jpountz closed pull request #12664: Optimize reading data from postings/impacts
enums.
URL: https://github.com/apache/lucene/pull/12664
--
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 c
jpountz commented on PR #12664:
URL: https://github.com/apache/lucene/pull/12664#issuecomment-1761003837
I'll split this PR into smaller ones, so that it's easier to understand the
impact of each change.
--
This is an automated message from the Apache Git Service.
To respond to the messag
jpountz opened a new pull request, #12668:
URL: https://github.com/apache/lucene/pull/12668
The code was written as if frequencies should be lazily decoding, except
that when refilling buffers freqs were getting eagerly decoded instead of
lazily.
--
This is an automated message from the
dungba88 commented on issue #12543:
URL: https://github.com/apache/lucene/issues/12543#issuecomment-1761024337
Copied from the PR:
It seems Tantivy segregate the building and the traverse of FST as 2
different entity. The FST Builder will just write the FST to a DataOutput and
not al
jpountz commented on PR #12668:
URL: https://github.com/apache/lucene/pull/12668#issuecomment-1761022817
Results on `wikibigall`:
```
TaskQPS baseline StdDevQPS
my_modified_version StdDevPct diff p-value
Coun
jpountz commented on PR #12668:
URL: https://github.com/apache/lucene/pull/12668#issuecomment-1761035614
For reference this is extracted from #12664.
--
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
gf2121 commented on PR #12661:
URL: https://github.com/apache/lucene/pull/12661#issuecomment-1761040593
I indeed can not understand how this version of `readMSBVLong` can be slower
than `readVLong` when JVM is warmed up enough. I wonder if there could be some
higher level reasons causing th
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761054475
I have not yet checked the benchmakr fraework at all (earlier PR), but the
idea here is cool:
--
This is an automated message from the Apache Git Service.
To respond to the mess
gf2121 commented on PR #12661:
URL: https://github.com/apache/lucene/pull/12661#issuecomment-1761076967
Maybe this is the
[todo](https://github.com/apache/lucene/blob/06341ffe1dfdd2c263c1ea2fe8da0ba6e719d00f/lucene/core/src/java/org/apache/lucene/codecs/lucene90/blocktree/SegmentTermsEnum.ja
jpountz merged PR #12668:
URL: https://github.com/apache/lucene/pull/12668
--
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
vsop-479 opened a new issue, #12669:
URL: https://github.com/apache/lucene/issues/12669
### Description
OrdinalMap maps segment ordinal to global ordinal. When collecting, we need
read segment ordinal by docID, and get global ordinal by segment ordinal.
If OrdinalMap maps docID to
jpountz opened a new pull request, #12670:
URL: https://github.com/apache/lucene/pull/12670
When we initially introduced support for dynamic pruning, we had an
implementation of WAND that would almost exclusively use `advance()`. Now that
we switched to MAXSCORE and rely much more on `nextD
jpountz commented on PR #12670:
URL: https://github.com/apache/lucene/pull/12670#issuecomment-1761288055
Results on wikibigall. Both the baseline and the contender have #12668.
```
TaskQPS baseline StdDevQPS
my_modified_version StdDev
dweiss commented on code in PR #12667:
URL: https://github.com/apache/lucene/pull/12667#discussion_r1358126818
##
help/jmh.txt:
##
@@ -0,0 +1,15 @@
+benchmarks
+==
+
+See the README.txt folder in lucene/benchmark-jmh for how to run benchmarks
using the build system.
Re
benwtrent commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1358143088
##
lucene/core/src/java/org/apache/lucene/codecs/lucene95/IncrementalHnswGraphMerger.java:
##
@@ -0,0 +1,209 @@
+/*
+ * Licensed to the Apache Software Foundation (A
javanna commented on code in PR #12606:
URL: https://github.com/apache/lucene/pull/12606#discussion_r1358146590
##
lucene/core/src/java/org/apache/lucene/search/IndexSearcher.java:
##
@@ -420,13 +418,12 @@ public int count(Query query) throws IOException {
}
/**
- * Re
shubhamvishu commented on PR #12636:
URL: https://github.com/apache/lucene/pull/12636#issuecomment-1761387046
@benwtrent Makes sense to meI'll see if I could address those part of
code in this PR without a new abstraction.
--
This is an automated message from the Apache Git Service.
T
rmuir commented on code in PR #12667:
URL: https://github.com/apache/lucene/pull/12667#discussion_r1358163111
##
help/jmh.txt:
##
@@ -0,0 +1,15 @@
+benchmarks
+==
+
+See the README.txt folder in lucene/benchmark-jmh for how to run benchmarks
using the build system.
Rev
shubhamvishu opened a new pull request, #12671:
URL: https://github.com/apache/lucene/pull/12671
### Description
This is a followup PR of #12548
Changes in this PR :
- Move some of the private static classes or functions etc out of
`DoubleValuesSource` class to keep
shubhamvishu commented on PR #12548:
URL: https://github.com/apache/lucene/pull/12548#issuecomment-1761428061
@msokolov @benwtrent Raised a followup PR #12671 for this and also moved
some other relevant static classes etc out of DVS(it now looks much better to
me). Thanks!
--
This is an
benwtrent commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1358208212
##
lucene/core/src/java/org/apache/lucene/util/hnsw/InitializedHnswGraphBuilder.java:
##
@@ -0,0 +1,78 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) un
jpountz opened a new pull request, #12672:
URL: https://github.com/apache/lucene/pull/12672
This changes the following:
- fewer docs indexed in non-nightly runs,
- `QueryUtils#checkFirstSkipTo` uses the `ScorerSupplier` API to convey it
will only check one doc,
- `QueryUtils#chec
jpountz commented on PR #12672:
URL: https://github.com/apache/lucene/pull/12672#issuecomment-1761457934
On my machine the test goes from ~30s to ~5s. This is probably still too
much, but at least these changes should not meaningfully decrease coverage.
--
This is an automated message fro
jpountz commented on issue #12669:
URL: https://github.com/apache/lucene/issues/12669#issuecomment-1761462109
> the global ordinals can not keep monotonic, which may harm to memory.
The other downside is that you then need to store `maxDoc` global ordinals
(one per doc) instead of `va
ChrisHegarty commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761486768
This looks great. I'll checkout the branch and play a little with, then
report back.
--
This is an automated message from the Apache Git Service.
To respond to the message, pleas
jpountz commented on issue #12648:
URL: https://github.com/apache/lucene/issues/12648#issuecomment-1761523732
I found the issue, this is due to codec randomization. This test fails when
assigned the `SimpleText` codec and one of the segment IDs has the `\n` or `\`
byte. This forces `SimpleT
jpountz opened a new pull request, #12673:
URL: https://github.com/apache/lucene/pull/12673
This test sometimes fails because `SimpleText` has a non-deterministic size
for its segment info file, due to escape characters. The test now enforces the
default codec, and checks that segments have
rmuir commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761564948
I also removed some of the hacky self-tests the benchmarks were doing
(comparing results across different impls). The reason is: when experimenting,
you can easily validate the correctness
ChrisHegarty commented on code in PR #12667:
URL: https://github.com/apache/lucene/pull/12667#discussion_r1358302452
##
lucene/benchmark-jmh/README.txt:
##
@@ -15,15 +15,16 @@ java --module-path lucene\benchmark-jmh\build\benchmarks
--module org.apache.luc
You can pass any J
ChrisHegarty commented on code in PR #12667:
URL: https://github.com/apache/lucene/pull/12667#discussion_r1358302452
##
lucene/benchmark-jmh/README.txt:
##
@@ -15,15 +15,16 @@ java --module-path lucene\benchmark-jmh\build\benchmarks
--module org.apache.luc
You can pass any J
benwtrent commented on code in PR #12582:
URL: https://github.com/apache/lucene/pull/12582#discussion_r1358319824
##
lucene/core/src/java/org/apache/lucene/util/ScalarQuantizer.java:
##
@@ -0,0 +1,267 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
jimczi commented on code in PR #12582:
URL: https://github.com/apache/lucene/pull/12582#discussion_r1358281601
##
lucene/core/src/java/org/apache/lucene/util/ScalarQuantizer.java:
##
@@ -0,0 +1,267 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ *
rmuir commented on code in PR #12667:
URL: https://github.com/apache/lucene/pull/12667#discussion_r1358317343
##
lucene/benchmark-jmh/README.txt:
##
@@ -15,15 +15,16 @@ java --module-path lucene\benchmark-jmh\build\benchmarks
--module org.apache.luc
You can pass any JMH opti
rmuir commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761582752
I updated the docs, please have another look.
--
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
benwtrent commented on code in PR #12582:
URL: https://github.com/apache/lucene/pull/12582#discussion_r1358322632
##
lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99ScalarQuantizedVectorsWriter.java:
##
@@ -0,0 +1,782 @@
+/*
+ * Licensed to the Apache Software Fou
gsmiller merged PR #12643:
URL: https://github.com/apache/lucene/pull/12643
--
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.ap
gsmiller merged PR #12642:
URL: https://github.com/apache/lucene/pull/12642
--
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.ap
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761613992
Hi Robert,
could you remove this exclusion:
https://github.com/dweiss/lucene/blob/e2e1b569fba1593ff4c3381f3494f552db2b4c74/lucene/benchmark-jmh/build.gradle#L37-L38
It is
jimczi commented on code in PR #12582:
URL: https://github.com/apache/lucene/pull/12582#discussion_r1358363771
##
lucene/core/src/java/org/apache/lucene/util/ScalarQuantizer.java:
##
@@ -0,0 +1,267 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ *
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761642871
> @dweiss is it enough to have RUNTIME_JAVA_HOME set when using Gradle to
run benchmark with also preview versions not yet supported by Gradle?
I noticed you need to run it on y
rmuir commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761663375
> Hi Robert, could you remove this exclusion:
https://github.com/dweiss/lucene/blob/e2e1b569fba1593ff4c3381f3494f552db2b4c74/lucene/benchmark-jmh/build.gradle#L37-L38
>
> It is no lo
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761668041
> > Hi Robert, could you remove this exclusion:
https://github.com/dweiss/lucene/blob/e2e1b569fba1593ff4c3381f3494f552db2b4c74/lucene/benchmark-jmh/build.gradle#L37-L38
> > It is no
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761682258
Sorry for the chaotic comments, the `failOnMissingClasses = false` line can
be removed.
I checked it with roberts branch:
```
> Task :lucene:benchmark-jmh:forbiddenApisMai
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761684636
I just committed the change.
--
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 specif
dweiss commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761712726
> ./gradlew :lucene:benchmark-jmh:run, which runs with the main gradle or
alternative JDK as configured in Gradle.
I mentioned this on the original issue - I don't think forking a J
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761713255
Let me fix...
--
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
dweiss commented on code in PR #12667:
URL: https://github.com/apache/lucene/pull/12667#discussion_r1358421683
##
help/jmh.txt:
##
@@ -0,0 +1,15 @@
+benchmarks
+==
+
+See the README.txt folder in lucene/benchmark-jmh for how to run benchmarks
using the build system.
Re
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761718313
Fixed. If you start the benchamrk now in module mode with the command line
provided by "assemble" it works correct:
```
> Task :lucene:benchmark-jmh:assemble
JMH bench
dweiss commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761724093
> is it enough to have RUNTIME_JAVA_HOME set when using Gradle to run
benchmark with also preview versions not yet supported by Gradle?
The code will compile with the provided RUNTI
dweiss commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761727033
> The jdk.incubator.vector must be removed from the module-info.java file.
This was intentional, otherwise the example benchmark wouldn't compile...
But since the vectors-specific c
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761741178
> > The jdk.incubator.vector must be removed from the module-info.java file.
>
> This was intentional, otherwise the example benchmark wouldn't compile...
But since the vectors-
jpountz merged PR #12673:
URL: https://github.com/apache/lucene/pull/12673
--
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 closed issue #12648: TestSizeBoundedForceMerge.testByteSizeLimit test
failure
URL: https://github.com/apache/lucene/issues/12648
--
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
iverase commented on code in PR #12625:
URL: https://github.com/apache/lucene/pull/12625#discussion_r1358473230
##
lucene/core/src/java/org/apache/lucene/index/TermsHashPerField.java:
##
@@ -255,6 +255,81 @@ final void writeBytes(int stream, byte[] b, int offset,
int len) {
cpoerschke opened a new pull request, #12674:
URL: https://github.com/apache/lucene/pull/12674
### Description
Building upon the #448 pull request upgrading to 1.9.4 -- this could be
follow-up or alternative.
#11277
--
This is an automated message from the Apache Git Service
cpoerschke commented on PR #448:
URL: https://github.com/apache/lucene/pull/448#issuecomment-1761792386
> > @jzonthemtn not sure I have the knowledge or chops to do this upgrade...
>
> I'll push an update!
I'm also interested in this -- #12674 so far, work-in-progress.
--
Thi
rmuir commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761796246
Yeah, i think its better not to compile benchmarks with incubator/preview
etc? I think benchmarks should just call lucene methods, and lucene code should
do the right thing with handling i
jpountz opened a new issue, #12675:
URL: https://github.com/apache/lucene/issues/12675
### Description
We generally prefer scores to not depend on the order of clauses, e.g.
BM25's multi-term sim scorer, disjunctive and conjunctive queries all sum up
scores into a double to reduce ac
cpoerschke commented on code in PR #12674:
URL: https://github.com/apache/lucene/pull/12674#discussion_r1358513093
##
lucene/analysis/opennlp/src/test/org/apache/lucene/analysis/opennlp/TestOpenNLPChunkerFilterFactory.java:
##
@@ -58,7 +58,7 @@ public class TestOpenNLPChunkerFil
cpoerschke commented on code in PR #12674:
URL: https://github.com/apache/lucene/pull/12674#discussion_r1358515284
##
lucene/licenses/opennlp-tools-NOTICE.txt:
##
@@ -1,6 +1,11 @@
-
-Apache OpenNLP Tools
-Copyright 2015 The Apache Software Foundation
+Apache OpenNLP
+Copyright 2
cpoerschke commented on code in PR #12674:
URL: https://github.com/apache/lucene/pull/12674#discussion_r1358579503
##
lucene/analysis/opennlp/src/tools/test-model-data/README.txt:
##
@@ -3,4 +3,4 @@ Training data derived from Reuters corpus in very unscientific
way.
Tagging do
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761940947
I ran all benchmarks in module mode (second line of assemble output) on my
AVX-256 laptop:
Prozessor: Intel(R) Core(TM) i7-8565U CPU @ 1.80GHz, 1992 MHz, 4 Kern(e),
8 logisch
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1761946068
> Yeah, i think its better not to compile benchmarks with incubator/preview
etc? I think benchmarks should just call lucene methods, and lucene code should
do the right thing with han
dweiss commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1762019283
> The problem was: if you run the benchmark in module mode, the requirement
added the module without respecting if the extra runtime parameter was given.
Yep, I understand. Did it a
rmuir commented on PR #12672:
URL: https://github.com/apache/lucene/pull/12672#issuecomment-1762048057
Thanks for improving the speed of this test, last time i ran tests it took
over 200seconds on my computer, i just did not look into it.
--
This is an automated message from the Apache Gi
rmuir commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1762051354
oh @gf2121 I missed that you added this, thank you! I am looking at it.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and
rmuir commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1762055897
@gf2121 @ChrisHegarty you can see the issue from his assembler output with
the failed intel optimization:
the current code does 2 x 256-bit vpmull on ymm registers, the proposed
simplif
rmuir commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1762063378
@gf2121 i think we could diagnose it further with
https://github.com/travisdowns/avx-turbo
--
This is an automated message from the Apache Git Service.
To respond to the message, please
rmuir commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1762068277
I compiled the code and ran it easily, just `git clone + make`. You do have
to run it as root to get the useful output, I took a risk on my machine:
```
think:avx-turbo[master]$ sudo
uschindler opened a new pull request, #12676:
URL: https://github.com/apache/lucene/pull/12676
This improves logging a bit:
- if java version is too old, but user explicitely enabled the incubator
module
- inform user about missing integer support
Robert and I hit this when tets
rmuir commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1762073195
I guess you will have to probably `modprobe msr` first. I already have the
`msr` module loaded for other nefarious purposes.
--
This is an automated message from the Apache Git Service.
rmuir commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1762084009
And at least the theory makes sense, this integer multiply is definitely
"avx512 heavy", so if u have a cpu susceptible to throttling, better to do
256bit multiplies that we do today. I gu
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1762118934
> Did it also happen when it was started with -jar ("classpath" mode)? I
don't think it should, right?
Naaah, that was the most confusing part.
--
This is an automated messag
benwtrent commented on PR #12657:
URL: https://github.com/apache/lucene/pull/12657#issuecomment-1762131240
Just being paranoid, I tested and verified that recall is absolutely
unchanged between these changes.
baseline:
```
0.500 0.10 10 10 4 50 20
uschindler merged PR #12676:
URL: https://github.com/apache/lucene/pull/12676
--
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.
uschindler commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1762144670
@rmuir I merged main barnch here, so you can retest the better logging of
#12676
--
This is an automated message from the Apache Git Service.
To respond to the message, please log
benwtrent commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1358794152
##
lucene/core/src/test/org/apache/lucene/util/hnsw/HnswGraphTestCase.java:
##
@@ -1147,15 +1139,15 @@ void assertVectorsEqual(AbstractMockVectorValues u,
AbstractM
rmuir commented on PR #12667:
URL: https://github.com/apache/lucene/pull/12667#issuecomment-1762210828
The logging works correctly for my use-case (accidentally running java 17).
Scalar method has no warnings, Vector method prints: `WARNING: Java vector
incubator module was enabled by comma
uschindler commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1358910134
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +122,22 @@ static VectorizationProvider lookup(boolean te
zhaih opened a new pull request, #12678:
URL: https://github.com/apache/lucene/pull/12678
### Description
This fix failure test:
https://jenkins.thetaphi.de/job/Lucene-main-Windows/13300/
The cause is the writer is not committed nor force merged, but the test is
depending on t
dweiss commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359203222
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +122,22 @@ static VectorizationProvider lookup(boolean testMo
uschindler commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359204137
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +122,22 @@ static VectorizationProvider lookup(boolean te
uschindler commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359204137
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +122,22 @@ static VectorizationProvider lookup(boolean te
uschindler commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359204492
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +122,22 @@ static VectorizationProvider lookup(boolean te
uschindler commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359204862
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +122,22 @@ static VectorizationProvider lookup(boolean te
901 - 1000 of 20534 matches
Mail list logo
