zhaih commented on PR #12660:
URL: https://github.com/apache/lucene/pull/12660#issuecomment-1783709669
Thanks Mike and Ben for reviewing! I'll try to merge and backport it
tomorrow. And create a issue for future refactoring
--
This is an automated message from the Apache Git Service.
To r
rmuir opened a new pull request, #12731:
URL: https://github.com/apache/lucene/pull/12731
The intel fma is nice, and its easier to reason about when looking at
assembly. We basically reduce the error for free where its available. Along
with another change (reducing the unrolling for cosine,
vigyasharma commented on PR #12724:
URL: https://github.com/apache/lucene/pull/12724#issuecomment-1783653077
I've backported this to `branch_9x`. Do we need it in any other branches?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to Gi
vigyasharma merged PR #12730:
URL: https://github.com/apache/lucene/pull/12730
--
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
vigyasharma commented on PR #12730:
URL: https://github.com/apache/lucene/pull/12730#issuecomment-1783652392
Backport change already approved in
https://github.com/apache/lucene/pull/12724. Merging.
--
This is an automated message from the Apache Git Service.
To respond to the message, pl
vigyasharma opened a new pull request, #12730:
URL: https://github.com/apache/lucene/pull/12730
Backport of fix in #12724
--
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.
vigyasharma merged PR #12724:
URL: https://github.com/apache/lucene/pull/12724
--
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
stefanvodita commented on PR #12506:
URL: https://github.com/apache/lucene/pull/12506#issuecomment-1783530943
I’ve integrated most of the suggestions. There’s just the matter of the name
for the slice pool class and the right package for it. I don’t have a strong
opinion on this. Maybe we m
stefanvodita commented on code in PR #12506:
URL: https://github.com/apache/lucene/pull/12506#discussion_r1375044479
##
lucene/core/src/java/org/apache/lucene/util/ByteBlockPool.java:
##
@@ -170,42 +191,42 @@ public void reset(boolean zeroFillBuffers, boolean
reuseFirst) {
}
benwtrent opened a new pull request, #12729:
URL: https://github.com/apache/lucene/pull/12729
Currently the HNSW codec does too many things, it not only indexes vectors,
but stores them and determines how to store them given the vector type.
This PR extracts out the vector storage int
shubhamvishu commented on PR #12726:
URL: https://github.com/apache/lucene/pull/12726#issuecomment-1783364698
Oh nice! @benwtrent I have added a CHANGES.txt entry under optimizations
now. Thanks!
--
This is an automated message from the Apache Git Service.
To respond to the message, pleas
benwtrent opened a new pull request, #12728:
URL: https://github.com/apache/lucene/pull/12728
follow up to https://github.com/apache/lucene/pull/12582
For user convenience, I added back the two parameter ctor for the HNSW codec.
--
This is an automated message from the Apache Git Se
benwtrent opened a new pull request, #12727:
URL: https://github.com/apache/lucene/pull/12727
We shouldn't ever return negative scores from vector similarity functions.
Given vector panama and nearly antipodal float[] vectors, it is possible that
cosine and (normalized) dot-product become s
benwtrent commented on PR #12726:
URL: https://github.com/apache/lucene/pull/12726#issuecomment-1783318072
@shubhamvishu could you add a "CHANGES.txt" entry under optimizations?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub
shubhamvishu opened a new pull request, #12726:
URL: https://github.com/apache/lucene/pull/12726
### Description
While going through
[VectorUtil](https://github.com/apache/lucene/blob/main/lucene/core/src/java/org/apache/lucene/util/VectorUtil.java)
class, I observed we don't have a
uschindler commented on PR #12724:
URL: https://github.com/apache/lucene/pull/12724#issuecomment-1783287980
Should I merge and backport it or do you want to do it?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
UR
luyuncheng commented on code in PR #12723:
URL: https://github.com/apache/lucene/pull/12723#discussion_r1374824046
##
lucene/core/src/java/org/apache/lucene/util/bkd/BKDReader.java:
##
@@ -216,7 +216,7 @@ private static class BKDPointTree implements PointTree {
scratchM
luyuncheng commented on code in PR #12723:
URL: https://github.com/apache/lucene/pull/12723#discussion_r1374824046
##
lucene/core/src/java/org/apache/lucene/util/bkd/BKDReader.java:
##
@@ -216,7 +216,7 @@ private static class BKDPointTree implements PointTree {
scratchM
luyuncheng commented on PR #12723:
URL: https://github.com/apache/lucene/pull/12723#issuecomment-1783206971
> this seems promising. I guess the only cost is the increased memory needed
because we create the FixedBitSet? Can you say how large this might get in the
worst case?
@msokolo
easyice commented on issue #12717:
URL: https://github.com/apache/lucene/issues/12717#issuecomment-1783202938
@jpountz Thanks for your explanation, i got some flame graph that shows the
`readVIntBlock` takes up a bit large proportion, I'll try to reproduce it with
some mocked data
--
Thi
ChrisHegarty commented on PR #12703:
URL: https://github.com/apache/lucene/pull/12703#issuecomment-1783135109
I've not been able to spend all that much time on this this week, but here's
my current thinking.
The abstractions in the PR are currently not great (as discussed above), but
jpountz commented on issue #12717:
URL: https://github.com/apache/lucene/issues/12717#issuecomment-1783020278
Does this actually matter for performance? My gut feeling is that either a
value has a long postings list, and then the vast majority of blocks will be
encoded with PFOR and should
msokolov commented on code in PR #12723:
URL: https://github.com/apache/lucene/pull/12723#discussion_r1374646933
##
lucene/core/src/java/org/apache/lucene/util/bkd/BKDReader.java:
##
@@ -216,7 +216,7 @@ private static class BKDPointTree implements PointTree {
scratchMin
msokolov commented on PR #12660:
URL: https://github.com/apache/lucene/pull/12660#issuecomment-1782983831
as for the renaming maybe we can look to do it as a followup PR?
IHnswGraphSearcher -> HnswGraphSearcher? Or maybe we can just use HnswSearcher
right now?
Re: controlling concurr
shubhamvishu commented on PR #12427:
URL: https://github.com/apache/lucene/pull/12427#issuecomment-1782977506
@gsmiller I finally got a chance to run the benchmarks for this change.
Below are the results(looks all good to me). Let me know what do you think?
Thanks!
```
jpountz commented on issue #12696:
URL: https://github.com/apache/lucene/issues/12696#issuecomment-1782814872
FWIW I could reproduce the speedup from disabling patching locally on
wikibigall:
```
TaskQPS baseline StdDevQPS
my_modified_version
shubhamvishu commented on PR #12716:
URL: https://github.com/apache/lucene/pull/12716#issuecomment-1782810001
@bruno-roustant Absolutely! I totally agree with your point.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and us
bruno-roustant commented on PR #12716:
URL: https://github.com/apache/lucene/pull/12716#issuecomment-1782623703
Thank you @shubhamvishu for these experiments. The table answers exactly to
the questions.
And it means actually there is no point to change the bit mixing, since it
does not b
dungba88 commented on issue #12543:
URL: https://github.com/apache/lucene/issues/12543#issuecomment-1782469977
I put a new revision with support for DataOutput and FileChannel.
When using DataOutput, if suffix sharing is enabled one also needs to pass a
RandomAccessInput for reading.
vigyasharma commented on issue #12708:
URL: https://github.com/apache/lucene/issues/12708#issuecomment-1782426261
Made a small change to assert on the exception type instead of checking the
exception message string.
--
This is an automated message from the Apache Git Service.
To respond
vigyasharma opened a new pull request, #12724:
URL: https://github.com/apache/lucene/pull/12724
Addresses #12708
`xml.TestCoreParser#testSpanNearQueryWithoutSlopXML` fails because of
changed exception message Java 22 EA. This change removes the test's dependency
on Java exception me
31 matches
Mail list logo
