mikemccand commented on a change in pull request #181:
URL: https://github.com/apache/lucene/pull/181#discussion_r659245670
##
File path: lucene/core/src/java/org/apache/lucene/search/MultiCollector.java
##
@@ -78,14 +107,17 @@ public static Collector wrap(Iterable
mikemccand commented on a change in pull request #146:
URL: https://github.com/apache/lucene/pull/146#discussion_r659244839
##
File path:
lucene/analysis/common/src/test/org/apache/lucene/analysis/core/TestFlattenGraphFilter.java
##
@@ -314,5 +323,221 @@ public void
mikemccand commented on a change in pull request #146:
URL: https://github.com/apache/lucene/pull/146#discussion_r659244744
##
File path:
lucene/analysis/common/src/test/org/apache/lucene/analysis/core/TestFlattenGraphFilter.java
##
@@ -314,5 +314,116 @@ public void
mikemccand commented on pull request #177:
URL: https://github.com/apache/lucene/pull/177#issuecomment-869081788
> That's just an idea.
I like this idea! We could maybe identify "known bad hotspot compilation
patterns" and retry the benchmark.
But the problem is (as you
mikemccand commented on a change in pull request #179:
URL: https://github.com/apache/lucene/pull/179#discussion_r659242645
##
File path:
lucene/facet/src/java/org/apache/lucene/facet/taxonomy/directory/DirectoryTaxonomyReader.java
##
@@ -351,12 +349,139 @@ public FacetLabel
mikemccand commented on pull request #179:
URL: https://github.com/apache/lucene/pull/179#issuecomment-869081261
That is really odd that the `SSDVFacets` show ~4% gain! I wonder if there
is a bug in `luceneutil` mis-labeling which facets task is using which impl!?
You could run a
mikemccand commented on a change in pull request #191:
URL: https://github.com/apache/lucene/pull/191#discussion_r659242352
##
File path:
lucene/facet/src/java/org/apache/lucene/facet/LongValueFacetCounts.java
##
@@ -162,8 +164,14 @@ private void count(String field, List