shubhamvishu commented on PR #12427:
URL: https://github.com/apache/lucene/pull/12427#issuecomment-1790129043
Nice!I like it how we discovered a build issue bug(which was hidden for
sometime now I think) and fixed it. Thanks for solving @dweiss!
--
This is an automated message from th
rmuir commented on PR #12737:
URL: https://github.com/apache/lucene/pull/12737#issuecomment-1790097686
ARM Graviton3E (256-bit SVE)
Master (well #12747 for sanity):
```
Benchmark (size) Mode Cnt Score
Error Units
VectorUtilBenchmark.
rmuir commented on PR #12747:
URL: https://github.com/apache/lucene/pull/12747#issuecomment-1790041666
and i know it is annoying it runs slower, i really tried to not be overkill,
but some methods here are super simple and super stable and some are unstable.
So I gave all methods minimum 3
rmuir opened a new pull request, #12747:
URL: https://github.com/apache/lucene/pull/12747
This benchmark is too noisy across forks which makes comparisons impossible
and misleading. Especially vectorized float methods on some machines: it is
almost useless.
I spent some time to reduc
rmuir commented on PR #12737:
URL: https://github.com/apache/lucene/pull/12737#issuecomment-1789905123
OK. sounds good to me @uschindler . I agree the user should be in control.
And I'd like to avoid us adding tons of flags for this stuff unless we need it
for testing. that's just my opinio
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378882326
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,132 +214,99 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
dungba88 closed pull request #12735: Clean up UnCompiledNode.inputCount
URL: https://github.com/apache/lucene/pull/12735
--
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 unsu
jpountz commented on PR #12526:
URL: https://github.com/apache/lucene/pull/12526#issuecomment-1789662266
@mikemccand FYI I gave a try at adding some interesting boolean queries to
nightly benchmarks at https://github.com/mikemccand/luceneutil/pull/240.
--
This is an automated message fro
jpountz closed issue #12746: TestDeletionPolicy.testOpenPriorSnapshot failing
URL: https://github.com/apache/lucene/issues/12746
--
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.
jpountz commented on issue #12746:
URL: https://github.com/apache/lucene/issues/12746#issuecomment-1789651692
Thanks for reporting, I pushed a fix at
https://github.com/apache/lucene/commit/66324f763fc7fb0d8e7cd6f334e5438f0171c84e.
--
This is an automated message from the Apache Git Servi
benwtrent commented on issue #12746:
URL: https://github.com/apache/lucene/issues/12746#issuecomment-1789639897
I verified this doesn't happen in 9.8.
--
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 opened a new issue, #12746:
URL: https://github.com/apache/lucene/issues/12746
### Description
TestDeletionPolicy.testOpenPriorSnapshot fails. Assertion fails on assuming
we are on a previous commit with more than one leaf.
Fails on main and 9.x.
stack
jpountz commented on code in PR #12729:
URL: https://github.com/apache/lucene/pull/12729#discussion_r1379282330
##
lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99HnswVectorsReader.java:
##
@@ -399,41 +281,30 @@ private HnswGraph getGraph(FieldEntry entry) throws
dweiss commented on issue #12742:
URL: https://github.com/apache/lucene/issues/12742#issuecomment-1789587275
I've provided a PR that fixes this. It is a corner case of us providing
custom javac parameters (modular classpath). I am surprised this took so long
to be discovered and I apologize
dweiss opened a new pull request, #12745:
URL: https://github.com/apache/lucene/pull/12745
Fixes #12742.,
--
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-
dweiss commented on PR #12427:
URL: https://github.com/apache/lucene/pull/12427#issuecomment-1789519097
See #12742 - it's a problem with out javac task settings. Running a clean is
a workaround. I will provide a fix for this when I get a chance, it's not a
critical issue (but it is a bug in
uschindler commented on PR #12737:
URL: https://github.com/apache/lucene/pull/12737#issuecomment-1789483634
Hi @rmuir,
I am fine with both approaches. Let me just create a new PR with some
changes for the Constants.java class and a separate pkg private utility class
for looking up JVM ar
asubbu90 commented on issue #12449:
URL: https://github.com/apache/lucene/issues/12449#issuecomment-1789370514
Hi @robro612 , you can see I have already opened a PR #12507 on this issue.
Do you want to have more context on this?
--
This is an automated message from the Apache Git Service.
gsmiller commented on PR #12427:
URL: https://github.com/apache/lucene/pull/12427#issuecomment-1789348091
@benwtrent glad it was a simple fix. Sorry it created churn!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
benwtrent commented on PR #12427:
URL: https://github.com/apache/lucene/pull/12427#issuecomment-1789266991
@gsmiller you are 100% correct 🤦
--
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 sp
gsmiller commented on PR #12427:
URL: https://github.com/apache/lucene/pull/12427#issuecomment-1789264349
@benwtrent is this an issue of the code only being partially rebuilt? The
signature for `Automata#makeStringUnion` was changed to accept a more general
`Iterable` as opposed to a `Colle
rmuir commented on PR #12743:
URL: https://github.com/apache/lucene/pull/12743#issuecomment-1789229323
thanks @ChrisHegarty ! I will try to dig in more to this. Again i don't see
speedups on my x86 so something has changed (maybe architecture) thats allowing
more to happen in parallel on yo
benwtrent commented on PR #12427:
URL: https://github.com/apache/lucene/pull/12427#issuecomment-1789218896
Something weird is happening. This commit is causing the following failure:
```
./gradlew test --tests
TestUnifiedHighlighterTermIntervals.testCustomFieldValueSource
-Dtests.seed
easyice commented on issue #12717:
URL: https://github.com/apache/lucene/issues/12717#issuecomment-1789106681
Oh.. Group-varint is a interesting encoder, I'd love to try it later in
the week
--
This is an automated message from the Apache Git Service.
To respond to the message, please
dungba88 opened a new issue, #12744:
URL: https://github.com/apache/lucene/issues/12744
### Description
Spawn from https://github.com/apache/lucene/pull/12738
It seems on
[rehash](https://github.com/apache/lucene/blob/main/lucene/core/src/java/org/apache/lucene/util/fst/NodeHas
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378882326
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,132 +214,99 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378841828
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -328,7 +323,128 @@ private void rehash(long lastNodeAddress) throws
IOException {
}
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378848941
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -107,28 +121,43 @@ public long add(FSTCompiler.UnCompiledNode nodeIn)
throws IOException {
ChrisHegarty commented on PR #12743:
URL: https://github.com/apache/lucene/pull/12743#issuecomment-1789014829
jmh prof output from my Linux x64 rocket lake -
https://gist.github.com/ChrisHegarty/508bb1857cb50df0d757f711c81fd740
--
This is an automated message from the Apache Git Service.
ChrisHegarty commented on PR #12743:
URL: https://github.com/apache/lucene/pull/12743#issuecomment-1789007292
> Your x64 is still i5-11400 ? I will look into this more.
Yes.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on
mikemccand commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378809307
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -107,28 +121,43 @@ public long add(FSTCompiler.UnCompiledNode nodeIn)
throws IOException {
rmuir commented on PR #12743:
URL: https://github.com/apache/lucene/pull/12743#issuecomment-1788919960
and what is the compiler's excuse this time? :) no floating point here. I
should just be able to write a simple loop and get good performance!
--
This is an automated message from the Ap
rmuir commented on PR #12743:
URL: https://github.com/apache/lucene/pull/12743#issuecomment-1788910222
thank you for the explanation, so now it just adds another mystery for me.
Your x64 is still i5-11400 ? I will look into this more.
Yes on the ARM, the assembly didn't look like wha
ChrisHegarty commented on PR #12743:
URL: https://github.com/apache/lucene/pull/12743#issuecomment-179191
I see reasonable speedups on both x64 and ARM, but sadly see no
vectorization in the disassembly. The speed seems to come from the 2x
instructions/pipelining (?) per strip mined lo
mikemccand commented on issue #12399:
URL: https://github.com/apache/lucene/issues/12399#issuecomment-1788845861
Some exciting updates here ... Intel continues to improve this "super fast
sorting using SIMD instructions" library:
https://www.phoronix.com/news/Intel-x86-simd-sort-4.0
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378704913
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,119 +197,85 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378703915
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,119 +197,85 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
benwtrent commented on issue #12615:
URL: https://github.com/apache/lucene/issues/12615#issuecomment-1788822945
So, I replicated the jvector benchmark (the lucene part) using the new int8
quantization.
Note, this is with `0` fan out or extra top-k gathered. Since the benchmark
on JV
gf2121 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378670256
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,119 +193,85 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
-
slow-J commented on PR #12741:
URL: https://github.com/apache/lucene/pull/12741#issuecomment-1788786005
> Thanks @slow-J ! I left some minor comments about additional `90` -> `99`
refactoring.
Thanks @gf2121 , committed all the suggestions.
--
This is an automated message from the
mikemccand commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378650643
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,119 +197,85 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
dungba88 commented on PR #12738:
URL: https://github.com/apache/lucene/pull/12738#issuecomment-1788758864
> Eventually it's moved to fallback, and, maybe it never gets promoted back
(single copy), or, maybe it does (+1 copy)
There is actually already one copy before this, which is whe
mikemccand commented on PR #12738:
URL: https://github.com/apache/lucene/pull/12738#issuecomment-1788744551
> But I ended up using a `List` where each item is a node instead of
ByteBlockPool due to the following reasons:
Hmm -- this is sizable added RAM overhead per entry. Added arra
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378601593
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,119 +193,85 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
mikemccand commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378599798
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -214,7 +222,13 @@ private long hash(long node) throws IOException {
* Compares an unfroz
dungba88 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378577965
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -110,25 +110,34 @@ public long add(FSTCompiler.UnCompiledNode nodeIn)
throws IOException {
gf2121 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378573421
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -186,119 +193,85 @@ private long hash(FSTCompiler.UnCompiledNode node) {
return h;
}
-
gf2121 commented on code in PR #12738:
URL: https://github.com/apache/lucene/pull/12738#discussion_r1378571480
##
lucene/core/src/java/org/apache/lucene/util/fst/NodeHash.java:
##
@@ -110,25 +110,34 @@ public long add(FSTCompiler.UnCompiledNode nodeIn)
throws IOException {
gf2121 commented on code in PR #12741:
URL: https://github.com/apache/lucene/pull/12741#discussion_r1378509674
##
lucene/core/src/java/org/apache/lucene/codecs/lucene99/Lucene99PostingsFormat.java:
##
@@ -0,0 +1,518 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) und
49 matches
Mail list logo
