kaivalnp commented on issue #15697:
URL: https://github.com/apache/lucene/issues/15697#issuecomment-3937386266

   > short->int conversion to be quite expensive
   
   Thanks @mccullocht, this was an interesting find!
   
   I reproduced JMH benchmarks from your PR on an AWS Graviton3 machine and the 
weird performance drop after a few warmup iterations no longer exists!
   
   I replaced all `convert` operations (including from other `int4` functions) 
with reinterpret casting + bit manipulation -- and I'm seeing even better 
performance -- I opened #15742 
   
   Would be great to see benchmarks for `x86` or other `aarch64` machines!


-- 
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]

Reply via email to