stefanvodita commented on PR #12966:
URL: https://github.com/apache/lucene/pull/12966#issuecomment-1890420916
I found a fun HeisenBug in one of the tests. When we iterate cursors from
`IntFloatHashMap`, the order is not deterministic. Float summation is not
commutative, so the result we get
stefanvodita commented on PR #12966:
URL: https://github.com/apache/lucene/pull/12966#issuecomment-1890420978
I've also run the benchmarks (`python3 src/python/localrun.py -source
wikimediumall`). There is measurable regression in the
`BrowseRandomLabelTaxoFacets` task, but not in other tax
mikemccand commented on issue #13011:
URL: https://github.com/apache/lucene/issues/13011#issuecomment-1890454889
Neat -- I had never heard of Kahan summation. Here is its [Wikipedia
page](https://en.wikipedia.org/wiki/Kahan_summation_algorithm).
--
This is an automated message from the A
mikemccand commented on PR #12966:
URL: https://github.com/apache/lucene/pull/12966#issuecomment-1890455264
> I found a fun HeisenBug in one of the tests.
Oh the joys of floating point math.
> For those who want to follow along, here are the exact numbers we are
adding in the t
uschindler commented on PR #12706:
URL: https://github.com/apache/lucene/pull/12706#issuecomment-1890455325
I also committed supprot for incubator SIMD vectorization in Java 22.
According to Java's change logs there were no changes to API at all, so code
runs as is.
@ChrisHegarty jus
mikemccand commented on PR #12966:
URL: https://github.com/apache/lucene/pull/12966#issuecomment-1890455536
> The regression in the taxo task is explained in the profiler. Boxing is
not cheap:
> `11.24% 10402M java.lang.Integer#valueOf()`
Hmm this is sort of spooky -- should we aim
stefanvodita commented on code in PR #12995:
URL: https://github.com/apache/lucene/pull/12995#discussion_r1451664416
##
lucene/facet/src/java/org/apache/lucene/facet/taxonomy/directory/TaxonomyIndexArrays.java:
##
@@ -68,25 +90,49 @@ public TaxonomyIndexArrays(IndexReader reader
stefanvodita commented on PR #12966:
URL: https://github.com/apache/lucene/pull/12966#issuecomment-1890862937
What I've done is I've only taken advantage of the boxing for genericity
when collecting results `getTop...` and not use it while performing the
aggregations themselves. Most of the
