jbewing commented on PR #5576:
URL: https://github.com/apache/hbase/pull/5576#issuecomment-1862076425

   > But for 512(2 or 3 bytes?), the microbench result shows that there are 
some performance lost?
   
   It does look like there's a bit of a regression in the 512 case (3 bytes). 
I'm not as worried about that given that this codepath is only used in the 
HBase codebase to read memstore timestamps. I can add a comment to the method 
signature to warn to use `WritableUtils#readVLong` in the case that smaller 
vLongs are being read. The common use case for this codepath is going to be the 
6 byte long/7 byte vLong decoding path which we're seeing the 40% performance 
increase on.


-- 
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]

Reply via email to