CH-Abhinav opened a new issue, #16677: URL: https://github.com/apache/lucene/issues/16677
### Description ### Description Currently, when Lucene decodes byte-aligned DocValues (specifically 8-bit, 16-bit, and 32-bit formats) into 64-bit `long` arrays, it relies on standard scalar `for` loops. Using Panama Vector API , we can process these conversions in bulk using in built vector casts (such as `ZERO_EXTEND_B2L` and `ZERO_EXTEND_S2L`). ### Possible solution In `PanamaDocValuesBulkDecodeSupport` we should add vectorized decoding of Byte, Short, and Int formats directly into Long arrays. This may amortize the conversion. Also we should add new test suite to check code in `PanamaDocValuesBulkDecodeSupport`. (Note : -previous issue `#16451` by someone is pure AI hallucination with benchmarks data unrelated to DocValues. So I made a new Issue where JMH benchmarking can be done) -- 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: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
