[
https://issues.apache.org/jira/browse/PARQUET-2375?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17784687#comment-17784687
]
ASF GitHub Bot commented on PARQUET-2375:
-----------------------------------------
jatin-bhateja commented on PR #1186:
URL: https://github.com/apache/parquet-mr/pull/1186#issuecomment-1805161923
Following are the scores with modified benchmark
System Configuration: Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz (40C 2S
Icelake Server)
```
Benchmark (bitWidth) Mode Cnt
Score Error Units
ByteBitPackingVectorBenchmarks.testUnpack 1 thrpt 2
12.922 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 2 thrpt 2
11.555 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 3 thrpt 2
9.762 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 4 thrpt 2
12.181 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 5 thrpt 2
8.285 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 6 thrpt 2
7.843 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 7 thrpt 2
6.445 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 8 thrpt 2
14.482 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 9 thrpt 2
5.312 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 10 thrpt 2
5.291 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 11 thrpt 2
4.793 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 12 thrpt 2
6.291 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 13 thrpt 2
4.439 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 14 thrpt 2
4.416 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 15 thrpt 2
4.221 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 16 thrpt 2
6.722 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 17 thrpt 2
3.566 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 18 thrpt 2
3.513 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 19 thrpt 2
3.306 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 20 thrpt 2
3.559 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 21 thrpt 2
3.083 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 22 thrpt 2
2.951 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 23 thrpt 2
2.825 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 24 thrpt 2
3.927 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 25 thrpt 2
2.744 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 26 thrpt 2
2.678 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 27 thrpt 2
2.684 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 28 thrpt 2
2.995 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 29 thrpt 2
2.473 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 30 thrpt 2
2.591 ops/s
ByteBitPackingVectorBenchmarks.testUnpack 31 thrpt 2
2.468 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 1 thrpt 2
52.159 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 2 thrpt 2
56.189 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 3 thrpt 2
38.339 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 4 thrpt 2
55.079 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 5 thrpt 2
36.410 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 6 thrpt 2
36.316 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 7 thrpt 2
38.673 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 8 thrpt 2
47.676 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 9 thrpt 2
48.053 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 10 thrpt 2
51.113 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 11 thrpt 2
34.871 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 12 thrpt 2
53.676 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 13 thrpt 2
31.572 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 14 thrpt 2
36.683 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 15 thrpt 2
21.280 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 16 thrpt 2
23.885 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 17 thrpt 2
31.902 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 18 thrpt 2
28.049 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 19 thrpt 2
29.078 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 20 thrpt 2
26.854 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 21 thrpt 2
28.077 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 22 thrpt 2
26.200 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 23 thrpt 2
27.233 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 24 thrpt 2
39.025 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 25 thrpt 2
28.820 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 26 thrpt 2
31.965 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 27 thrpt 2
22.974 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 28 thrpt 2
21.621 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 29 thrpt 2
19.245 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 30 thrpt 2
20.863 ops/s
ByteBitPackingVectorBenchmarks.testUnpackVector 31 thrpt 2
22.689 ops/s
```
> Extend vectorized bit unpacking benchmark for various bit sizes.
> ----------------------------------------------------------------
>
> Key: PARQUET-2375
> URL: https://issues.apache.org/jira/browse/PARQUET-2375
> Project: Parquet
> Issue Type: Improvement
> Components: parquet-mr
> Reporter: JATIN BHATEJA
> Priority: Minor
>
> * Cleanup the benchmark and remove noise from main benchmarking kernels.
> * Extend the benchmark for various bit widths (1-31)
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
