jpountz commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3072341580
Nightly benchmarks caught up with the change. No slowdowns and a few good
speedups! https://benchmarks.mikemccandless.com/OrStopWords.html I'll add an
annotation.
--
This is an automa
gf2121 merged PR #14935:
URL: https://github.com/apache/lucene/pull/14935
--
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
gf2121 commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3065989475
The luceneutil result sounds too good to be true for me, so i rechecked and
found that this new branch contains https://github.com/apache/lucene/pull/14906
while baseline not, so the prev
gf2121 commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2202681751
##
lucene/core/src/java/org/apache/lucene/codecs/lucene103/BitSetUtil.java:
##
@@ -0,0 +1,90 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or mo
gf2121 commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2202681751
##
lucene/core/src/java/org/apache/lucene/codecs/lucene103/BitSetUtil.java:
##
@@ -0,0 +1,90 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or mo
jpountz commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2202551786
##
lucene/core/src/java/org/apache/lucene/codecs/lucene103/BitSetUtil.java:
##
@@ -0,0 +1,90 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or m
jpountz commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2202357715
##
lucene/core/src/java/org/apache/lucene/codecs/lucene103/BitSetUtil.java:
##
@@ -0,0 +1,90 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or m
gf2121 commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3064581110
I did some iteration locally and find out a way providing similar
performance without `scratch` array (new code in `denseBranchLessParallel`).
```
Benchmark
github-actions[bot] commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3064579312
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
jpountz commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2201711227
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/BitsetToArrayBenchmark.java:
##
@@ -0,0 +1,379 @@
+/*
+ * Licensed to the Apache Software Foundation (
github-actions[bot] commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3063052843
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
github-actions[bot] commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3062971915
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
gf2121 commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2201126320
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/BitsetToArrayBenchmark.java:
##
@@ -0,0 +1,379 @@
+/*
+ * Licensed to the Apache Software Foundation (A
gf2121 commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2200932004
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/BitsetToArrayBenchmark.java:
##
@@ -0,0 +1,379 @@
+/*
+ * Licensed to the Apache Software Foundation (A
gf2121 commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2200926950
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/BitsetToArrayBenchmark.java:
##
@@ -0,0 +1,379 @@
+/*
+ * Licensed to the Apache Software Foundation (A
jpountz commented on code in PR #14935:
URL: https://github.com/apache/lucene/pull/14935#discussion_r2200809196
##
lucene/benchmark-jmh/src/java/org/apache/lucene/benchmark/jmh/BitsetToArrayBenchmark.java:
##
@@ -0,0 +1,379 @@
+/*
+ * Licensed to the Apache Software Foundation (
github-actions[bot] commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3062244403
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
gf2121 commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3062225468
FWIW jmh result of the vectorized version (code in #14910)
```
Benchmark (bitCount) Mode Cnt Score Error
Units
BitsetToArrayBenchmark.vectoriz
github-actions[bot] commented on PR #14935:
URL: https://github.com/apache/lucene/pull/14935#issuecomment-3061678159
This PR does not have an entry in lucene/CHANGES.txt. Consider adding one.
If the PR doesn't need a changelog entry, then add the skip-changelog label to
it and you will stop
gf2121 opened a new pull request, #14935:
URL: https://github.com/apache/lucene/pull/14935
This comes from
https://github.com/apache/lucene/pull/14910#issuecomment-3059176116.
JMH shows:
1. `whileLoop`, `forloop`, `forLoopManualUnrolling` are having even
performance, `whileLoo
20 matches
Mail list logo
