github-actions[bot] commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1907129767
This PR has not had activity in the past 2 weeks, labeling it as stale. If
the PR is waiting for review, notify the d...@lucene.apache.org list. Thank you
for your contributi
github-actions[bot] commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1880899961
This PR has not had activity in the past 2 weeks, labeling it as stale. If
the PR is waiting for review, notify the d...@lucene.apache.org list. Thank you
for your contributi
gf2121 commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1820687175
Thanks for feedback @mikemccand !
> Hmm it looks like random got a bit slower in candidate? Flush time ~550
ish ms in baseline and maybe ~650 ish ms in candidate?
Ohhh! I rec
mikemccand commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1820626033
> I also run the index script to see flush time with this new approach,
result in ~15% faster for random data and no regression on asc/desc :)
Hmm it looks like random got a bit
gf2121 commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1820564110
I also run the index script to see flush time with this new approach, result
in ~15% faster for random data and no regression on asc/desc :)
Benchmark Detail
**Baseline**
gf2121 commented on code in PR #12800:
URL: https://github.com/apache/lucene/pull/12800#discussion_r1400210595
##
lucene/core/src/java/org/apache/lucene/util/BaseLSBRadixSorter.java:
##
@@ -0,0 +1,122 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
gf2121 commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1820461507
I did some more work to find out the balance between memory / performance in
various data distribution. The way i'm thinking now is that we keep the
timsorter here, but make the run lengt
gf2121 commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1818846591
Thanks for feedback @jpountz !
> but this seems to come with greater heap requirements as well?
Yes, +1 for the concern. The original approach requires at most
`ArrayUtil.ove
jpountz commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1814761757
I like the idea, but this seems to come with greater heap requirements as
well?
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to
gf2121 commented on code in PR #12800:
URL: https://github.com/apache/lucene/pull/12800#discussion_r1392571640
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/DocSorterBenchmark.java:
##
@@ -0,0 +1,241 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF)
gf2121 commented on code in PR #12800:
URL: https://github.com/apache/lucene/pull/12800#discussion_r1391150384
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/DocSorterBenchmark.java:
##
@@ -0,0 +1,241 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF)
mikemccand commented on PR #12800:
URL: https://github.com/apache/lucene/pull/12800#issuecomment-1808206830
I wonder whether `Arrays.sort` might be a good choice instead of making our
own powerful sorting classes? [OpenJDK is (gradually?) taking advantage of
fast SIMD sorting](https://gith
mikemccand commented on code in PR #12800:
URL: https://github.com/apache/lucene/pull/12800#discussion_r1391138011
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/DocSorterBenchmark.java:
##
@@ -0,0 +1,241 @@
+/*
+ * Licensed to the Apache Software Foundation (A
mikemccand commented on code in PR #12800:
URL: https://github.com/apache/lucene/pull/12800#discussion_r1391137486
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/DocSorterBenchmark.java:
##
@@ -0,0 +1,241 @@
+/*
+ * Licensed to the Apache Software Foundation (A
gf2121 opened a new pull request, #12800:
URL: https://github.com/apache/lucene/pull/12800
**Description**
In https://github.com/apache/lucene/pull/12114, we had great numbers for LSB
radix sorter when sorting random docs in `SortingDocsEnum` . But we can not
take advantage of the LS
15 matches
Mail list logo
