yashmayya commented on PR #13758:
URL: https://github.com/apache/pinot/pull/13758#issuecomment-2276020558

   I've updated the mode aggregation benchmark and pushed the change to this 
branch itself. Here are the results:
   
   Without this optimization:
   ```
   Benchmark                      (_nullHandlingEnabledPerCent)  (_nullPeriod)  
 Mode  Cnt   Score   Error   Units
   BenchmarkModeAggregation.test                            100              1  
thrpt   50  46.303 ± 0.274  ops/ms
   BenchmarkModeAggregation.test                            100              2  
thrpt   50  36.063 ± 0.203  ops/ms
   BenchmarkModeAggregation.test                            100              4  
thrpt   50  31.343 ± 0.128  ops/ms
   BenchmarkModeAggregation.test                            100              8  
thrpt   50  30.965 ± 0.631  ops/ms
   BenchmarkModeAggregation.test                            100             16  
thrpt   50  31.089 ± 0.154  ops/ms
   BenchmarkModeAggregation.test                            100             32  
thrpt   50  32.219 ± 0.192  ops/ms
   BenchmarkModeAggregation.test                            100             64  
thrpt   50  32.074 ± 0.157  ops/ms
   BenchmarkModeAggregation.test                            100            128  
thrpt   50  32.154 ± 0.388  ops/ms
   BenchmarkModeAggregation.test                             50              1  
thrpt   50  47.861 ± 8.267  ops/ms
   BenchmarkModeAggregation.test                             50              2  
thrpt   50  34.433 ± 0.933  ops/ms
   BenchmarkModeAggregation.test                             50              4  
thrpt   50  31.789 ± 0.204  ops/ms
   BenchmarkModeAggregation.test                             50              8  
thrpt   50  32.006 ± 0.303  ops/ms
   BenchmarkModeAggregation.test                             50             16  
thrpt   50  32.274 ± 0.243  ops/ms
   BenchmarkModeAggregation.test                             50             32  
thrpt   50  32.425 ± 0.200  ops/ms
   BenchmarkModeAggregation.test                             50             64  
thrpt   50  32.301 ± 0.171  ops/ms
   BenchmarkModeAggregation.test                             50            128  
thrpt   50  32.473 ± 0.147  ops/ms
   BenchmarkModeAggregation.test                              0              1  
thrpt   50  28.122 ± 0.310  ops/ms
   BenchmarkModeAggregation.test                              0              2  
thrpt   50  27.857 ± 0.163  ops/ms
   BenchmarkModeAggregation.test                              0              4  
thrpt   50  27.667 ± 0.164  ops/ms
   BenchmarkModeAggregation.test                              0              8  
thrpt   50  27.933 ± 0.188  ops/ms
   BenchmarkModeAggregation.test                              0             16  
thrpt   50  27.772 ± 0.159  ops/ms
   BenchmarkModeAggregation.test                              0             32  
thrpt   50  28.018 ± 0.187  ops/ms
   BenchmarkModeAggregation.test                              0             64  
thrpt   50  28.060 ± 0.358  ops/ms
   BenchmarkModeAggregation.test                              0            128  
thrpt   50  28.159 ± 0.192  ops/ms
   ```
   
   With this optimization:
   ```
   Benchmark                      (_nullHandlingEnabledPerCent)  (_nullPeriod)  
 Mode  Cnt       Score       Error   Units
   BenchmarkModeAggregation.test                            100              1  
thrpt   50  154371.180 ±   592.954  ops/ms
   BenchmarkModeAggregation.test                            100              2  
thrpt   50      36.142 ±     0.301  ops/ms
   BenchmarkModeAggregation.test                            100              4  
thrpt   50      32.071 ±     0.403  ops/ms
   BenchmarkModeAggregation.test                            100              8  
thrpt   50      31.035 ±     0.396  ops/ms
   BenchmarkModeAggregation.test                            100             16  
thrpt   50      31.827 ±     0.201  ops/ms
   BenchmarkModeAggregation.test                            100             32  
thrpt   50      32.297 ±     0.138  ops/ms
   BenchmarkModeAggregation.test                            100             64  
thrpt   50      32.251 ±     0.158  ops/ms
   BenchmarkModeAggregation.test                            100            128  
thrpt   50      32.452 ±     0.161  ops/ms
   BenchmarkModeAggregation.test                             50              1  
thrpt   50   67106.709 ± 37843.797  ops/ms
   BenchmarkModeAggregation.test                             50              2  
thrpt   50      34.638 ±     0.855  ops/ms
   BenchmarkModeAggregation.test                             50              4  
thrpt   50      32.408 ±     0.210  ops/ms
   BenchmarkModeAggregation.test                             50              8  
thrpt   50      29.867 ±     0.635  ops/ms
   BenchmarkModeAggregation.test                             50             16  
thrpt   50      31.973 ±     0.408  ops/ms
   BenchmarkModeAggregation.test                             50             32  
thrpt   50      32.236 ±     0.293  ops/ms
   BenchmarkModeAggregation.test                             50             64  
thrpt   50      31.478 ±     0.494  ops/ms
   BenchmarkModeAggregation.test                             50            128  
thrpt   50      32.253 ±     0.276  ops/ms
   BenchmarkModeAggregation.test                              0              1  
thrpt   50      27.938 ±     0.179  ops/ms
   BenchmarkModeAggregation.test                              0              2  
thrpt   50      27.863 ±     0.204  ops/ms
   BenchmarkModeAggregation.test                              0              4  
thrpt   50      28.036 ±     0.202  ops/ms
   BenchmarkModeAggregation.test                              0              8  
thrpt   50      28.203 ±     0.372  ops/ms
   BenchmarkModeAggregation.test                              0             16  
thrpt   50      28.106 ±     0.146  ops/ms
   BenchmarkModeAggregation.test                              0             32  
thrpt   50      28.663 ±     0.352  ops/ms
   BenchmarkModeAggregation.test                              0             64  
thrpt   50      28.048 ±     0.145  ops/ms
   BenchmarkModeAggregation.test                              0            128  
thrpt   50      28.103 ±     0.164  ops/ms
   ```
   
   Also, I tried this optimization with a couple of other aggregation functions 
and benchmarks locally and saw similar result.


-- 
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