zhaih commented on issue #12358:
URL: https://github.com/apache/lucene/issues/12358#issuecomment-1595284895
After roughly catching up with the threads, I would like to go back to
> +1 to update the Weight#count API contract or introduce a new API to
better optimize counting.
So we
almogtavor commented on issue #12318:
URL: https://github.com/apache/lucene/issues/12318#issuecomment-1595104039
@romseygeek Oh so that sounds even better than what I thought. In the case
of `ByteBuffersDirectory` & MMap, I can treat the match operation like a total
sync operation and use i
jainankitk commented on issue #12297:
URL: https://github.com/apache/lucene/issues/12297#issuecomment-1595083334
> I agree with Robert that 1kB per segment doesn't sound like a crazy amount
of allocations, which suggests that you are searching many segments. Does the
memory allocation profi
zhaih commented on PR #12371:
URL: https://github.com/apache/lucene/pull/12371#issuecomment-1595003976
Thank you @Jackyrie2 for working on it, I think @benwtrent 's concern about
memory makes sense but seems to me we should be able to reduce the memory usage
later on (as this is still a dra
Jackyrie2 commented on PR #12371:
URL: https://github.com/apache/lucene/pull/12371#issuecomment-1594985039
@benwtrent thanks for your suggestion, I will take some time this weekend to
try some benchmarking. I have some benchmark ideas in mind that might show some
improvements.
--
This is
zhaih commented on PR #12372:
URL: https://github.com/apache/lucene/pull/12372#issuecomment-1594967308
> Not sure how that's related to this code or how to fix it
I think this is due to newly cut 9.7 branch so we probably just need to wait
a bit more. I see all the auto testing is com
jbellis commented on PR #12372:
URL: https://github.com/apache/lucene/pull/12372#issuecomment-1594923701
CI says this test is failing
```org.apache.lucene.backward_index.TestBackwardsCompatibility > test
suite's output saved to
/home/runner/work/lucene/lucene/lucene/backward-codecs/b
jbellis commented on code in PR #12372:
URL: https://github.com/apache/lucene/pull/12372#discussion_r1232415072
##
lucene/core/src/java/org/apache/lucene/util/hnsw/HnswGraphSearcher.java:
##
@@ -262,7 +274,6 @@ private NeighborQueue searchLevel(
int visitedLimit)
t
sohami commented on PR #12374:
URL: https://github.com/apache/lucene/pull/12374#issuecomment-1594871500
@javanna I am also thinking with regards to too many constructor can we see
if deprecating the ones with IndexReader as inputs and keeping the ones with
IndexReaderContext over the time w
benwtrent commented on code in PR #12372:
URL: https://github.com/apache/lucene/pull/12372#discussion_r1232396894
##
lucene/core/src/java/org/apache/lucene/util/hnsw/HnswGraphSearcher.java:
##
@@ -262,7 +274,6 @@ private NeighborQueue searchLevel(
int visitedLimit)
jbellis commented on PR #12372:
URL: https://github.com/apache/lucene/pull/12372#issuecomment-1594763703
I'm using the million-row sift dataset via this harness
https://github.com/jbellis/hnswdemo/tree/benchmarking
I believe what is happening is that allocation is basically free and t
benwtrent commented on PR #12371:
URL: https://github.com/apache/lucene/pull/12371#issuecomment-1594666777
I ran this change with https://github.com/mikemccand/luceneutil with
knnPerfTest with over 1 vectors from `enwiki-20120502-lines-1k-300d.vec`
I tried maxConn 32 & 96 and bea
benwtrent commented on issue #12342:
URL: https://github.com/apache/lucene/issues/12342#issuecomment-1594648943
> And then switch the score to scale in 10 as a breaking change. Or is
condoning negative scores under any circumstances a non-starter?
If you are utilizing hybrid search, n
benwtrent commented on PR #12372:
URL: https://github.com/apache/lucene/pull/12372#issuecomment-1594642822
Hey @jbellis the change looks nice to me. But, I ran
https://github.com/mikemccand/luceneutil `knnPerTest` and saw no change at all
in indexing time.
Am I missing something? Cou
jbellis commented on PR #12372:
URL: https://github.com/apache/lucene/pull/12372#issuecomment-1594627724
cc @msokolov @benwtrent @zhaih
--
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
uschindler merged PR #12376:
URL: https://github.com/apache/lucene/pull/12376
--
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.
16 matches
Mail list logo
