zacharymorn commented on pull request #56:
URL: https://github.com/apache/lucene-site/pull/56#issuecomment-824588404
The change looks good at staging
https://lucene.staged.apache.org/whoweare.html .
Hi @janhoy, I have a quick question. I see you had a few recent PRs to merge
from `
[
https://issues.apache.org/jira/browse/LUCENE-9335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17327129#comment-17327129
]
Zach Chen commented on LUCENE-9335:
---
Makes sense. I guess the general strategy then wo
[
https://issues.apache.org/jira/browse/LUCENE-9934?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17327083#comment-17327083
]
Thomas DuBuisson commented on LUCENE-9934:
--
Previously Muse was installed on ma
[
https://issues.apache.org/jira/browse/LUCENE-9934?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17327079#comment-17327079
]
David Smiley commented on LUCENE-9934:
--
+1 Yes definitely. The "INFRA" Jira projec
[
https://issues.apache.org/jira/browse/LUCENE-9936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17326884#comment-17326884
]
Dawid Weiss commented on LUCENE-9936:
-
Nope, you're correct. I just stated the fact.
[
https://issues.apache.org/jira/browse/LUCENE-9936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17326842#comment-17326842
]
Chris M. Hostetter commented on LUCENE-9936:
but that's just for the maven
[
https://issues.apache.org/jira/browse/LUCENE-9936?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17326837#comment-17326837
]
Dawid Weiss commented on LUCENE-9936:
-
For the record - the artifacts published to A
dweiss commented on pull request #100:
URL: https://github.com/apache/lucene/pull/100#issuecomment-824288978
Thanks @Jawnnypoo
--
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 comme
dweiss merged pull request #100:
URL: https://github.com/apache/lucene/pull/100
--
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.
For queries about this service, please con
[
https://issues.apache.org/jira/browse/LUCENE-9936?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Chris M. Hostetter updated LUCENE-9936:
---
Attachment: LUCENE-9936.patch
Assignee: Chris M. Hostetter
Status: Open
Chris M. Hostetter created LUCENE-9936:
--
Summary: update gradle build to support gpg signing of tgz/zip
distributions
Key: LUCENE-9936
URL: https://issues.apache.org/jira/browse/LUCENE-9936
Proje
jpountz commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-824192328
Very cool, thanks for testing what happens with small integers!
Do the numbers get better if you use `TimSorter` as a fallback sort?
`InPlaceMergeSorter` is convenient because
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617698010
##
File path:
lucene/core/src/java/org/apache/lucene/util/bkd/MutablePointsReaderUtils.java
##
@@ -39,13 +41,23 @@
public static void sort(
BKDCon
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617697446
##
File path: lucene/core/src/java/org/apache/lucene/util/StableMSBRadixSorter.java
##
@@ -0,0 +1,91 @@
+/*
+ * Licensed to the Apache Software Foundation (A
neoremind edited a comment on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-824164245
> +1 to always use the stable version of the algorithm. This would only use
transient memory and in reasonable amounts, so I'm not concerned with the
memory usage.
Pe
neoremind edited a comment on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-824164245
> +1 to always use the stable version of the algorithm. This would only use
transient memory and in reasonable amounts, so I'm not concerned with the
memory usage.
Pe
neoremind edited a comment on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-824172567
--
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.
For que
neoremind commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-824172567
> For instance I'd expect users who index integers (4 bytes) between 0 and
2^24 to notice speedups that are closer to the one that you computed for
bytesPerDim=3 than for bytesPerD
Jawnnypoo commented on pull request #100:
URL: https://github.com/apache/lucene/pull/100#issuecomment-824170487
Yep, palantir's consistency check is on the latest version.
Hopefully we will get the 7.0 check soon🤞
--
This is an automated message from the Apache Git Service.
To res
neoremind commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-824164245
> +1 to always use the stable version of the algorithm. This would only use
transient memory and in reasonable amounts, so I'm not concerned with the
memory usage.
Per comme
jpountz commented on pull request #92:
URL: https://github.com/apache/lucene/pull/92#issuecomment-824151870
This PR is mixing a force-merge-specific setting with natural merges.
Can you give more context about the problem that you are trying to solve? Is
setting `deletesPctAllowed` t
[
https://issues.apache.org/jira/browse/LUCENE-9932?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
neoremind updated LUCENE-9932:
--
Attachment: refined-code-benchmark2.png
> Performance improvement for BKD index building
> ---
Adrien Grand created LUCENE-9935:
Summary: Bulk merges for stored fields when index sorting is
enabled
Key: LUCENE-9935
URL: https://issues.apache.org/jira/browse/LUCENE-9935
Project: Lucene - Core
mikemccand edited a comment on pull request #56:
URL: https://github.com/apache/lucene-site/pull/56#issuecomment-824056348
Woot! First commit for @zacharymorn, congrats!
Er, rather, first *push*!
--
This is an automated message from the Apache Git Service.
To respond to the messag
mikemccand commented on pull request #56:
URL: https://github.com/apache/lucene-site/pull/56#issuecomment-824056348
Woot! First commit for @zacharymorn, congrats!
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
U
[
https://issues.apache.org/jira/browse/LUCENE-9335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17326487#comment-17326487
]
Adrien Grand commented on LUCENE-9335:
--
I'd be interested in seeing how the results
jpountz commented on a change in pull request #84:
URL: https://github.com/apache/lucene/pull/84#discussion_r615920337
##
File path:
lucene/analysis/common/src/java/org/apache/lucene/analysis/miscellaneous/ScandinavianNormalizer.java
##
@@ -0,0 +1,134 @@
+/*
+ * Licensed to th
jpountz commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-823962146
> If we are ok to take the memory penalty of using this extra array, then I
think it make sense to always use the stable version of the algorithm?
+1 to always use the stable v
jpountz commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617418482
##
File path:
lucene/core/src/java/org/apache/lucene/util/bkd/MutablePointsReaderUtils.java
##
@@ -39,13 +41,23 @@
public static void sort(
BKDConfi
neoremind edited a comment on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-823960282
> For instance I'd expect users who index integers (4 bytes) between 0 and
2^24 to notice speedups that are closer to the one that you computed for
bytesPerDim=3 than for by
neoremind commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-823960282
> For instance I'd expect users who index integers (4 bytes) between 0 and
2^24 to notice speedups that are closer to the one that you computed for
bytesPerDim=3 than for bytesPerD
iverase commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-823957855
What I read for this results is that the stable sorting is faster than the
non-stable. I think it makes sense because re-ordering using a helper array is
probably faster that in-plac
jpountz commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-823954946
I left a few comments before your previous comment, I left some suggestions
to make the code simpler but I think that it's getting close.
Thanks for the great performance analy
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617409172
##
File path:
lucene/core/src/java/org/apache/lucene/util/bkd/MutablePointsReaderUtils.java
##
@@ -39,13 +41,23 @@
public static void sort(
BKDCon
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617406830
##
File path: lucene/core/src/java/org/apache/lucene/util/StableMSBRadixSorter.java
##
@@ -0,0 +1,91 @@
+/*
+ * Licensed to the Apache Software Foundation (A
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617406830
##
File path: lucene/core/src/java/org/apache/lucene/util/StableMSBRadixSorter.java
##
@@ -0,0 +1,91 @@
+/*
+ * Licensed to the Apache Software Foundation (A
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617406029
##
File path: lucene/core/src/java/org/apache/lucene/util/StableMSBRadixSorter.java
##
@@ -0,0 +1,91 @@
+/*
+ * Licensed to the Apache Software Foundation (A
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617406029
##
File path: lucene/core/src/java/org/apache/lucene/util/StableMSBRadixSorter.java
##
@@ -0,0 +1,91 @@
+/*
+ * Licensed to the Apache Software Foundation (A
neoremind commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617403651
##
File path: lucene/core/src/java/org/apache/lucene/util/StableMSBRadixSorter.java
##
@@ -0,0 +1,91 @@
+/*
+ * Licensed to the Apache Software Foundation (A
neoremind commented on pull request #91:
URL: https://github.com/apache/lucene/pull/91#issuecomment-823949164
@jpountz Per your advice, I have updated the code.
In terms of performance, I refined `TestBKDDisableSortDocId`, to make it
re-runnable as a benchmark. I have made the follow
[
https://issues.apache.org/jira/browse/LUCENE-9932?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
neoremind updated LUCENE-9932:
--
Attachment: flame-graph.png
> Performance improvement for BKD index building
> ---
jpountz commented on a change in pull request #91:
URL: https://github.com/apache/lucene/pull/91#discussion_r617367359
##
File path: lucene/core/src/java/org/apache/lucene/util/StableMSBRadixSorter.java
##
@@ -0,0 +1,91 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF
[
https://issues.apache.org/jira/browse/LUCENE-9932?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
neoremind updated LUCENE-9932:
--
Attachment: refined-code-benchmark.png
> Performance improvement for BKD index building
>
[
https://issues.apache.org/jira/browse/LUCENE-9334?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
neoremind updated LUCENE-9334:
--
Attachment: (was: refined-code-benchmark.png)
> Require consistency between data-structures on a p
[
https://issues.apache.org/jira/browse/LUCENE-9334?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
neoremind updated LUCENE-9334:
--
Attachment: refined-code-benchmark.png
> Require consistency between data-structures on a per-field ba
dweiss commented on pull request #100:
URL: https://github.com/apache/lucene/pull/100#issuecomment-823851178
Please upgrade palantir's consistency check to the latest version too (if
it's not there) as it's known to cause incompatibility problems.
The build is not compatible with Gr
[
https://issues.apache.org/jira/browse/LUCENE-9335?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17326323#comment-17326323
]
Zach Chen commented on LUCENE-9335:
---
Hi [~jpountz], I took a stab at implementing BMM
zacharymorn commented on pull request #101:
URL: https://github.com/apache/lucene/pull/101#issuecomment-823838918
luceneutil benchmark result with wikimedium5m
```
TaskQPS baseline StdDevQPS my_modified_version
StdDevPct diff p-value
zacharymorn opened a new pull request #101:
URL: https://github.com/apache/lucene/pull/101
Implement BMM algorithm from "Optimizing Top-k Document Retrieval Strategies
for Block-Max Indexes" by Dimopoulos, Nepomnyachiy and Suel.
--
This is an automated message from the Apache Git Service
49 matches
Mail list logo
