wesm commented on pull request #7542:
URL: https://github.com/apache/arrow/pull/7542#issuecomment-649601444
Here's my benchmarks on i9-9960X
```
$ archery benchmark diff --cc=gcc-8 --cxx=g++-8 cyb70289/sort master
--suite-filter=vector-sort
benchmark baseline
contender change % counters
12 SortToIndicesInt64Count/32768/10/min_time:1.000 1.233 GiB/sec
2.062 GiB/sec 67.239 {'iterations': 56417, 'null_percent': 10.0}
1 SortToIndicesInt64Count/32768/100/min_time:1.000 1.605 GiB/sec
2.033 GiB/sec 26.628 {'iterations': 73247, 'null_percent': 1.0}
3 SortToIndicesInt64Count/32768/0/min_time:1.000 2.438 GiB/sec
2.850 GiB/sec 16.911 {'iterations': 111533, 'null_percent': 0.0}
7 SortToIndicesInt64Count/32768/2/min_time:1.000 805.132 MiB/sec
854.275 MiB/sec 6.104 {'iterations': 36039, 'null_percent': 50.0}
11 SortToIndicesInt64Compare/32768/100/min_time:1.000 147.729 MiB/sec
149.201 MiB/sec 0.996 {'iterations': 6588, 'null_percent': 1.0}
2 SortToIndicesInt64Count/1048576/1/min_time:1.000 4.559 GiB/sec
4.580 GiB/sec 0.458 {'iterations': 6503, 'null_percent': 100.0}
5 SortToIndicesInt64Compare/1048576/1/min_time:1.000 4.554 GiB/sec
4.560 GiB/sec 0.136 {'iterations': 6577, 'null_percent': 100.0}
0 SortToIndicesInt64Compare/32768/10000/min_time:1.000 148.181 MiB/sec
148.368 MiB/sec 0.126 {'iterations': 6659, 'null_percent': 0.01}
9 SortToIndicesInt64Compare/32768/2/min_time:1.000 245.628 MiB/sec
245.887 MiB/sec 0.106 {'iterations': 11171, 'null_percent': 50.0}
6 SortToIndicesInt64Compare/32768/10/min_time:1.000 156.791 MiB/sec
156.841 MiB/sec 0.032 {'iterations': 7073, 'null_percent': 10.0}
4 SortToIndicesInt64Count/8388608/1/min_time:1.000 4.276 GiB/sec
4.269 GiB/sec -0.155 {'iterations': 772, 'null_percent': 100.0}
10 SortToIndicesInt64Compare/8388608/1/min_time:1.000 4.286 GiB/sec
4.273 GiB/sec -0.305 {'iterations': 769, 'null_percent': 100.0}
8 SortToIndicesInt64Compare/32768/1/min_time:1.000 4.677 GiB/sec
4.648 GiB/sec -0.634 {'iterations': 209517, 'null_percent': 100.0}
14 SortToIndicesInt64Count/32768/1/min_time:1.000 4.661 GiB/sec
4.617 GiB/sec -0.945 {'iterations': 214256, 'null_percent': 100.0}
15 SortToIndicesInt64Compare/32768/0/min_time:1.000 151.334 MiB/sec
147.025 MiB/sec -2.847 {'iterations': 6660, 'null_percent': 0.0}
13 SortToIndicesInt64Count/32768/10000/min_time:1.000 2.312 GiB/sec
2.131 GiB/sec -7.844 {'iterations': 91492, 'null_percent': 0.01}
```
and clang-11
```
benchmark baseline
contender change % counters
15 SortToIndicesInt64Count/32768/0/min_time:1.000 2.069 GiB/sec
3.490 GiB/sec 68.665 {'iterations': 95577, 'null_percent': 0.0}
13 SortToIndicesInt64Count/32768/10/min_time:1.000 1.288 GiB/sec
1.895 GiB/sec 47.099 {'iterations': 59504, 'null_percent': 10.0}
10 SortToIndicesInt64Count/32768/100/min_time:1.000 1.551 GiB/sec
1.993 GiB/sec 28.456 {'iterations': 71755, 'null_percent': 1.0}
2 SortToIndicesInt64Count/32768/10000/min_time:1.000 1.975 GiB/sec
2.069 GiB/sec 4.771 {'iterations': 87376, 'null_percent': 0.01}
3 SortToIndicesInt64Compare/32768/0/min_time:1.000 155.058 MiB/sec
158.211 MiB/sec 2.033 {'iterations': 6935, 'null_percent': 0.0}
9 SortToIndicesInt64Compare/32768/2/min_time:1.000 248.515 MiB/sec
252.036 MiB/sec 1.417 {'iterations': 11676, 'null_percent': 50.0}
0 SortToIndicesInt64Compare/32768/1/min_time:1.000 4.678 GiB/sec
4.735 GiB/sec 1.222 {'iterations': 214204, 'null_percent': 100.0}
8 SortToIndicesInt64Count/32768/1/min_time:1.000 4.677 GiB/sec
4.729 GiB/sec 1.110 {'iterations': 213435, 'null_percent': 100.0}
5 SortToIndicesInt64Compare/32768/10000/min_time:1.000 151.573 MiB/sec
151.744 MiB/sec 0.113 {'iterations': 6769, 'null_percent': 0.01}
1 SortToIndicesInt64Compare/32768/10/min_time:1.000 162.339 MiB/sec
162.226 MiB/sec -0.069 {'iterations': 7284, 'null_percent': 10.0}
14 SortToIndicesInt64Compare/32768/100/min_time:1.000 151.879 MiB/sec
151.589 MiB/sec -0.191 {'iterations': 6837, 'null_percent': 1.0}
7 SortToIndicesInt64Compare/1048576/1/min_time:1.000 4.570 GiB/sec
4.553 GiB/sec -0.383 {'iterations': 6579, 'null_percent': 100.0}
6 SortToIndicesInt64Count/1048576/1/min_time:1.000 4.569 GiB/sec
4.550 GiB/sec -0.407 {'iterations': 6459, 'null_percent': 100.0}
11 SortToIndicesInt64Count/8388608/1/min_time:1.000 4.280 GiB/sec
4.240 GiB/sec -0.939 {'iterations': 771, 'null_percent': 100.0}
12 SortToIndicesInt64Compare/8388608/1/min_time:1.000 4.281 GiB/sec
4.222 GiB/sec -1.369 {'iterations': 758, 'null_percent': 100.0}
4 SortToIndicesInt64Count/32768/2/min_time:1.000 835.964 MiB/sec
752.317 MiB/sec -10.006 {'iterations': 37122, 'null_percent': 50.0}
```
+1, thanks @cyb70289!
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
