[GitHub] [lucene] rmuir commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

rmuir commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r785220678 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReader r

[GitHub] [lucene] rmuir commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

rmuir commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r785197508 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReader r

[GitHub] [lucene] gsmiller commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

gsmiller commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r785195850 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReade

[GitHub] [lucene] rmuir commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

rmuir commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r785162808 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReader r

[GitHub] [lucene] rmuir commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

rmuir commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r785161786 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReader r

[GitHub] [lucene] gsmiller commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

gsmiller commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r785150523 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReade

[GitHub] [lucene] gsmiller commented on pull request #605: LUCENE-10379: Count directly into the dense values array in FastTaxonomyFacetCounts#countAll

gsmiller commented on pull request #605: URL: https://github.com/apache/lucene/pull/605#issuecomment-1013393369 @gf2121 glad it worked out. Now I just wish we could understand what exactly was causing the regression with the other part of the change. -- This is an automated message from

[GitHub] [lucene] gsmiller commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

gsmiller commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r785091711 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReade

[GitHub] [lucene] gsmiller commented on pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

gsmiller commented on pull request #606: URL: https://github.com/apache/lucene/pull/606#issuecomment-1013386501 Benchmarking this change locally shows no impact at all. So I don't think it's actually worth pushing this change unless we just want to isolate where the nightly benchmark runs

[GitHub] [lucene] rmuir commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

rmuir commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r784962069 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReader r

[GitHub] [lucene] rmuir commented on a change in pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

rmuir commented on a change in pull request #606: URL: https://github.com/apache/lucene/pull/606#discussion_r784962069 ## File path: lucene/facet/src/java/org/apache/lucene/facet/taxonomy/FastTaxonomyFacetCounts.java ## @@ -126,23 +126,39 @@ private void countAll(IndexReader r

[GitHub] [lucene] gsmiller opened a new pull request #606: LUCENE-10380: Further optimize FastTaxonomyFacetCounts#countAll by moving the liveDocs null check outside the loops

gsmiller opened a new pull request #606: URL: https://github.com/apache/lucene/pull/606 This change attempts to bring in the other piece of the LUCENE-10350 change without the regressions. See LUCENE-10374 for more details. -- This is an automated message from the Apache Git Service. To

[jira] [Comment Edited] (LUCENE-10379) Count directly into the values array in FastTaxonomyFacetCounts#countAl

[ https://issues.apache.org/jira/browse/LUCENE-10379?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17476216#comment-17476216 ] Greg Miller edited comment on LUCENE-10379 at 1/14/22, 3:51 PM: -

[jira] [Commented] (LUCENE-10379) Count directly into the values array in FastTaxonomyFacetCounts#countAl

[ https://issues.apache.org/jira/browse/LUCENE-10379?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17476216#comment-17476216 ] Greg Miller commented on LUCENE-10379: -- Looks like this change worked! The perform

[jira] [Created] (LUCENE-10380) Move liveDocs null check outside the loops in FastTaxonomyFacetCounts#countAll

Greg Miller created LUCENE-10380: Summary: Move liveDocs null check outside the loops in FastTaxonomyFacetCounts#countAll Key: LUCENE-10380 URL: https://issues.apache.org/jira/browse/LUCENE-10380 Proj

[GitHub] [lucene] uschindler commented on a change in pull request #602: LUCENE-10376: Roll up the loop in vint/vlong in DataInput

uschindler commented on a change in pull request #602: URL: https://github.com/apache/lucene/pull/602#discussion_r784844870 ## File path: lucene/core/src/java/org/apache/lucene/store/ByteArrayDataInput.java ## @@ -109,55 +110,20 @@ public long readLong() { @Override pub

[GitHub] [lucene] uschindler commented on pull request #602: LUCENE-10376: Roll up the loop in vint/vlong in DataInput

uschindler commented on pull request #602: URL: https://github.com/apache/lucene/pull/602#issuecomment-1013117133 Thanks. So unrolling brings no difference and as the bug is gone, we can replace all of those implementations. About the ByteBufferIndexInput for MMapDirectory: This is r

[jira] [Commented] (LUCENE-10168) Only test N-2 codecs nightly

[ https://issues.apache.org/jira/browse/LUCENE-10168?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17476121#comment-17476121 ] ASF subversion and git services commented on LUCENE-10168: -- Co

[jira] [Commented] (LUCENE-10168) Only test N-2 codecs nightly

[ https://issues.apache.org/jira/browse/LUCENE-10168?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17476122#comment-17476122 ] ASF subversion and git services commented on LUCENE-10168: -- Co

[GitHub] [lucene] gf2121 edited a comment on pull request #602: LUCENE-10376: Roll up the loop in vint/vlong in DataInput

gf2121 edited a comment on pull request #602: URL: https://github.com/apache/lucene/pull/602#issuecomment-1012836713 FYI here are some of my thoughts based on all these benchmark reports: * Rolling up loops for vint/vlong seems not bring a significant speed up (neither significant re

[jira] [Commented] (LUCENE-10377) Replace 'sortPos' parameter in SortField.getComparator()

[ https://issues.apache.org/jira/browse/LUCENE-10377?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17476049#comment-17476049 ] Alan Woodward commented on LUCENE-10377: Given that this change is on an intern

[GitHub] [lucene] romseygeek commented on pull request #603: LUCENE-10377: Replace 'sortPos' with 'enableSkipping' in SortField.getComparator()

romseygeek commented on pull request #603: URL: https://github.com/apache/lucene/pull/603#issuecomment-1012964049 > I think you have already done this removal in this PR, haven't you? Yes, sorry I meant to comment to call it out - it turned out to be a really easy change so I folded

[GitHub] [lucene] gf2121 edited a comment on pull request #541: LUCENE-10315: Speed up BKD leaf block ids codec by a 512 ints ForUtil

gf2121 edited a comment on pull request #541: URL: https://github.com/apache/lucene/pull/541#issuecomment-1012163438 Thanks @iverase ! > I remember in the approach I tried I was batching the docIds by a number (128 or 256) but in general tI didn't see much better performance comparin

[GitHub] [lucene] gf2121 edited a comment on pull request #602: LUCENE-10376: Roll up the loop in vint/vlong in DataInput

gf2121 edited a comment on pull request #602: URL: https://github.com/apache/lucene/pull/602#issuecomment-1012836713 FYI here are some of my thoughts based on all these benchmark reports: * Rolling up loops for vint/vlong seems not bring a significant speed up (neither significant re

[GitHub] [lucene] gf2121 commented on pull request #605: LUCENE-10379: Count directly into the dense values array in FastTaxonomyFacetCounts#countAll

gf2121 commented on pull request #605: URL: https://github.com/apache/lucene/pull/605#issuecomment-1012923242 I'm seeing BrowseMonthTaxoFacets increased 30% without any regression in [last night benchmark](https://home.apache.org/~mikemccand/lucenebench/2022.01.13.18.03.24.html), Thanks @