Tony-X commented on PR #12688: URL: https://github.com/apache/lucene/pull/12688#issuecomment-1857380213
## Non-trivial amount of allocations for? .... building IndexInput slice descriptions !? `jdk.internal.misc.Unsafe#allocateUninitializedArray()`. This was not trivial to find out why. But again with the raw JFR report, we can analyze the call tree. It turn out that in the [`buildSlice()` implementation](https://github.com/apache/lucene/blob/main/lucene/core/src/java21/org/apache/lucene/store/MemorySegmentIndexInput.java#L451) of `MemorySegmentIndexInput` we call [`IndexInput#getFullSliceDescription()`](https://github.com/apache/lucene/blob/main/lucene/core/src/java/org/apache/lucene/store/IndexInput.java#L128) which creates new String. And `allocateUninitializedArray` is called to allocate the bytes for the String. AFAIK, the description is only used for debugging and tracking purposes. I didn't expect it'd cause that much of allocation. So I made a change to pass null when building the description so those allocations are gone. ``` Before PERCENT HEAP SAMPLES STACK 10.39% 12103M java.lang.Long#valueOf() 9.91% 11543M org.apache.lucene.facet.taxonomy.IntTaxonomyFacets#initializeValueCounters() 8.91% 10383M jdk.internal.misc.Unsafe#allocateUninitializedArray() After PERCENT HEAP SAMPLES STACK [37/1812] 25.97% 32791M java.lang.Long#valueOf() 7.58% 9571M org.apache.lucene.facet.taxonomy.IntTaxonomyFacets#initializeValueCounters() 5.13% 6482M org.apache.lucene.util.FixedBitSet#<init>() ``` -- 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.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
