Re: [PR] [WIP] Introduce bpv24 vectorized decoding for DocIdsWriter [lucene]

[via GitHub]

gf2121 commented on PR #14203:
URL: https://github.com/apache/lucene/pull/14203#issuecomment-2642377091

   E2E result is disappointing:
   ```
                               TaskQPS baseline      StdDevQPS 
my_modified_version      StdDev                Pct diff p-value
                             IntSet      839.86      (3.1%)      835.78      
(3.7%)   -0.5% (  -7% -    6%) 0.677
                         TermDTSort      203.80      (5.1%)      206.38      
(5.0%)    1.3% (  -8% -   11%) 0.464
                CountFilteredIntNRQ       91.06      (2.3%)       92.48      
(3.5%)    1.6% (  -4% -    7%) 0.123
                  TermDayOfYearSort      203.77      (3.7%)      208.42      
(3.7%)    2.3% (  -4% -   10%) 0.075
                             IntNRQ      148.29      (1.9%)      151.84      
(3.2%)    2.4% (  -2% -    7%) 0.008
                     FilteredIntNRQ      146.00      (2.6%)      149.63      
(3.2%)    2.5% (  -3% -    8%) 0.013
   ```
   
   Profile suggests bottleneck is `FixedBitset#set` rather than decoding
   
![image](https://github.com/user-attachments/assets/3a68c6be-05d8-45bc-978d-44abcd99fcab)
   


-- 
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