Shibi-bala commented on issue #12637:
URL: https://github.com/apache/lucene/issues/12637#issuecomment-1763742706
@msfroh this impacts the ability to snapshot since you can't read old
`userData`. Check out the test in my PR:
https://github.com/apache/lucene/pull/12626
--
This is an automa
zhaih commented on PR #12651:
URL: https://github.com/apache/lucene/pull/12651#issuecomment-1763700781
I reran'd the benchmark and still get the similar perf and same recall.
(Just to make sure the later edits have not messed up things)
--
This is an automated message from the Apache Git
zhaih commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1360085712
##
lucene/core/src/test/org/apache/lucene/util/hnsw/HnswGraphTestCase.java:
##
@@ -454,77 +454,6 @@ public void testSearchWithSelectiveAcceptOrds() throws
IOException {
zhaih commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1360085215
##
lucene/core/src/java/org/apache/lucene/util/hnsw/OnHeapHnswGraph.java:
##
@@ -158,50 +185,82 @@ public int entryNode() {
return entryNode;
}
+ /**
+ * WA
zhaih commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1360084970
##
lucene/core/src/java/org/apache/lucene/util/hnsw/HnswGraphSearcher.java:
##
@@ -284,6 +285,13 @@ int graphNextNeighbor(HnswGraph graph) throws IOException {
retu
zhaih commented on PR #12660:
URL: https://github.com/apache/lucene/pull/12660#issuecomment-1763695626
Thanks @msokolov, I'll take a look soon.
> so ideally we'd compare total time to merge using single-threaded vs using
this.
The most fair comparison here I posted might be the
MarcusSorealheis commented on PR #12626:
URL: https://github.com/apache/lucene/pull/12626#issuecomment-1763688716
Hi @Shibi-bala and great to see you here. Let's sync up this week and maybe
we can help move this PR forward. It's a good catch, so thank you.
--
This is an automated message
vsop-479 closed issue #12669: Make OrdinalMap maps docID to global ordinal
directly？
URL: https://github.com/apache/lucene/issues/12669
--
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
vsop-479 commented on issue #12669:
URL: https://github.com/apache/lucene/issues/12669#issuecomment-1763633876
> need to store maxDoc global ordinals (one per doc) instead of valueCount
global ordinals.
@jpountz Thanks for remind that. I will close this issue.
--
This is an automa
zhaih commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1359985681
##
lucene/core/src/java/org/apache/lucene/util/hnsw/OnHeapHnswGraph.java:
##
@@ -158,50 +185,82 @@ public int entryNode() {
return entryNode;
}
+ /**
+ * WA
zhaih commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1359984597
##
lucene/core/src/java/org/apache/lucene/util/hnsw/OnHeapHnswGraph.java:
##
@@ -99,27 +122,31 @@ public void addNode(int level, int node) {
entryNode = node;
zhaih commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1359984155
##
lucene/core/src/java/org/apache/lucene/util/hnsw/OnHeapHnswGraph.java:
##
@@ -74,23 +88,32 @@ public final class OnHeapHnswGraph extends HnswGraph
implements Account
dweiss opened a new pull request, #12684:
URL: https://github.com/apache/lucene/pull/12684
Code coverage tests have been failing with an exception thrown from jacoco's
premain:
```
Caused by: java.security.AccessControlException: access denied
("java.lang.RuntimePermission" "createCla
msokolov commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1359929577
##
lucene/core/src/java/org/apache/lucene/util/hnsw/InitializedHnswGraphBuilder.java:
##
@@ -0,0 +1,78 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) und
msokolov commented on PR #12660:
URL: https://github.com/apache/lucene/pull/12660#issuecomment-1763471838
OK, I posted https://github.com/apache/lucene/pull/12683. I'm curious about
the performance measurements. We want to know how much lock contention there is
so ideally we'd compare total
msokolov opened a new pull request, #12683:
URL: https://github.com/apache/lucene/pull/12683
### Description
adds row locks to OnHeapHnswGraph
seems to be broken in the way it adds new nodes to upper levels such that
they become discoverable before they have neighbors?
I
uschindler commented on PR #12634:
URL: https://github.com/apache/lucene/pull/12634#issuecomment-1763463645
I backported this one to 9.x.
--
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 speci
uschindler commented on PR #12632:
URL: https://github.com/apache/lucene/pull/12632#issuecomment-1763463569
I backported this one to 9.x.
--
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 speci
uschindler commented on PR #12680:
URL: https://github.com/apache/lucene/pull/12680#issuecomment-1763463280
I backported this one to 9.x. As I cleaned up commits, this one was
committed in 9.x as part of #12681
--
This is an automated message from the Apache Git Service.
To respond to th
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763462800
I backported this one to 9.x. For that I selected the separate commits,
cherrypicked and squashed them with my tortoise GUI (sorry!).
--
This is an automated message from the Apache
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763454286
I can take care of backporting, also your previous commits. For maintenenace
it would be good as the code differs dramatically now between main and 9.x and
makes merging hard.
--
T
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763426882
> > > ok, all done. "asfgit" can do proper merge commit :) thanks
@uschindler !
> >
> >
> > The only downside is that backporting gets harder. So squashed PRs are
easier to
msokolov commented on code in PR #12651:
URL: https://github.com/apache/lucene/pull/12651#discussion_r1359884135
##
lucene/core/src/java/org/apache/lucene/util/hnsw/HnswGraphSearcher.java:
##
@@ -284,6 +285,13 @@ int graphNextNeighbor(HnswGraph graph) throws IOException {
r
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763402811
> > ok, all done. "asfgit" can do proper merge commit :) thanks @uschindler !
>
> The only downside is that backporting gets harder. So squashed PRs are
easier to handle in that rega
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763402251
I haven't been backporting recent vector changes as they are fairly
aggressive. That's just my thoughts. If we want to do that, we have to do it
proper and backport test, benchmarks, loggi
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763399661
> ok, all done. "asfgit" can do proper merge commit :) thanks @uschindler !
The only downside is that backporting gets harder. So squashed PRs are
easier to handle in that regar
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763386185
Will you backport?
--
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
easyice commented on code in PR #12658:
URL: https://github.com/apache/lucene/pull/12658#discussion_r1357058015
##
lucene/core/src/java/org/apache/lucene/util/bkd/BKDWriter.java:
##
@@ -519,9 +526,8 @@ private Runnable writeFieldNDims(
// compute the min/max for this slice
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763380049
ok, all done. "asfgit" can do proper merge commit :) thanks @uschindler !
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub a
asfgit merged PR #12681:
URL: https://github.com/apache/lucene/pull/12681
--
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.apac
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763378774
> At least we should have the option to select both variants!
Yeah thats strange, It doesnt make sense to remove the option from the merge
button. I will use "git push" to workaround
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763378449
> and hope to avoid more conflicts (it is a conflict nightmare...)
This is why I tend to make one PR after the other. So I don't want to start
another test cleanup one before th
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763378070
OK, i just wanted to do "enough" on the tests to make some progress. If we
just add _128, _256, _512 variants without addressing the testing, then we
create a potential problem.
I w
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763377432
> I understand the git issue now. Why do we configure the merge button as
"squash"? I'd like to default it to "merge" because I value my sanity!
At least we should have the opti
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763373537
I understand the git issue now. Why do we configure the merge button as
"squash"? I'd like to default it to "merge" because I value my sanity!
--
This is an automated message from the Ap
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763372114
wow, merge conflicts, really? git is terrible sometimes. literally didn't
merge anything that isn't in this branch already.
--
This is an automated message from the Apache Git Service.
T
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763372042
> I don't mind if sysprop code moves, as long as you can ensure no
classloading deadlocks
No classloading deadlock possible. We have a Holder for that already:
https://github.c
rmuir commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763371551
I don't mind if sysprop code moves, as long as you can ensure no
classloading deadlocks
--
This is an automated message from the Apache Git Service.
To respond to the message, please log
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763371401
Otherwise this is a great improvement for testing, we just need to fine-tune
it! ❤️
--
This is an automated message from the Apache Git Service.
To respond to the message, please lo
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763370765
In addition to that, maybe we should change TestVectorUtilSupport to spawn 3
child JVMs and don't set the property top-level?
The problem is currently any of those:
- If we e
rmuir merged PR #12680:
URL: https://github.com/apache/lucene/pull/12680
--
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.apach
benwtrent commented on code in PR #12657:
URL: https://github.com/apache/lucene/pull/12657#discussion_r1359865451
##
lucene/core/src/java/org/apache/lucene/codecs/lucene95/IncrementalHnswGraphMerger.java:
##
@@ -0,0 +1,189 @@
+/*
+ * Licensed to the Apache Software Foundation (A
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763361313
> > This also makes this test reproducible from random seed regardless of
the hardware, as `SPECIES_PREFERRED` is not used at all in tests. From a test
perspective, it is like a forbi
ChrisHegarty commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763359841
> for me, when investigating a modification, this works easily enough:
>
> ```
> $ for bits in 128 256 512; do ./gradlew -p lucene/core test --tests
TestVectorUtilSupport -
ChrisHegarty commented on code in PR #12681:
URL: https://github.com/apache/lucene/pull/12681#discussion_r1359857545
##
lucene/core/src/java20/org/apache/lucene/internal/vectorization/PanamaVectorUtilSupport.java:
##
@@ -16,165 +16,155 @@
*/
package org.apache.lucene.internal
ChrisHegarty commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359852957
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +124,21 @@ static VectorizationProvider lookup(boolean
uschindler merged PR #12677:
URL: https://github.com/apache/lucene/pull/12677
--
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 code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359852363
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +124,21 @@ static VectorizationProvider lookup(boolean te
ChrisHegarty commented on code in PR #12677:
URL: https://github.com/apache/lucene/pull/12677#discussion_r1359850988
##
lucene/core/src/java/org/apache/lucene/internal/vectorization/VectorizationProvider.java:
##
@@ -120,24 +124,21 @@ static VectorizationProvider lookup(boolean
uschindler commented on PR #12681:
URL: https://github.com/apache/lucene/pull/12681#issuecomment-1763331631
> This also makes this test reproducible from random seed regardless of the
hardware, as `SPECIES_PREFERRED` is not used at all in tests. From a test
perspective, it is like a forbidd
50 matches
Mail list logo
