Tony-X opened a new issue, #15189: URL: https://github.com/apache/lucene/issues/15189
### Description At CoC I was presenting the idea in ForUtil inspired by Paul's (tantivy's creator) [post](https://fulmicoton.com/posts/bitpacking/) in 2019 that without the real SIMD instruction we can use long/int as a wide container to operate multiple smaller bit-width ints at the same time. The current [ForUtil](https://github.com/apache/lucene/blob/9d8685f26e9641aae59819cbec0ed1700d787ede/lucene/core/src/java/org/apache/lucene/codecs/lucene104/ForUtil.java#L26-L28) uses the 32bit int as the container. Natrually @mikemccand raised this question about why we are not using longs? In theory operating longs should give better performance on modern hardwares. Goal: * understand the int/long choice if there is any historical exploration. * explore implementations with long to see if we get more performance boost -- 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]
