aocsa commented on pull request #11210:
URL: https://github.com/apache/arrow/pull/11210#issuecomment-938690728
Thanks @lidavidm, I update the benchmark code.
Here the last numbers
```
-------------------------------------------------------------------------------------------------------------------------------------------
Benchmark
Time CPU Iterations UserCounters...
-------------------------------------------------------------------------------------------------------------------------------------------
MinimalEndToEndBench/num_batches:100/batch_size:10/use_executor:1/real_time
3035047 ns 1055366 ns 233 bytes_per_second=1.5711M/s
items_per_second=32.9484k/s
MinimalEndToEndBench/num_batches:100/batch_size:10/use_executor:0/real_time
2431543 ns 2430187 ns 289 bytes_per_second=1.96105M/s
items_per_second=41.1262k/s
MinimalEndToEndBench/num_batches:100/batch_size:100/use_executor:1/real_time
3057831 ns 1067603 ns 225 bytes_per_second=15.594M/s
items_per_second=32.7029k/s
MinimalEndToEndBench/num_batches:100/batch_size:100/use_executor:0/real_time
2551493 ns 2550704 ns 272 bytes_per_second=18.6886M/s
items_per_second=39.1927k/s
MinimalEndToEndBench/num_batches:100/batch_size:1000/use_executor:1/real_time
2963222 ns 1082264 ns 233 bytes_per_second=160.918M/s
items_per_second=33.747k/s
MinimalEndToEndBench/num_batches:100/batch_size:1000/use_executor:0/real_time
3571509 ns 3570340 ns 202 bytes_per_second=133.511M/s
items_per_second=27.9994k/s
MinimalEndToEndBench/num_batches:1000/batch_size:10/use_executor:1/real_time
24303552 ns 7684245 ns 30 bytes_per_second=1.96201M/s
items_per_second=41.1462k/s
MinimalEndToEndBench/num_batches:1000/batch_size:10/use_executor:0/real_time
25323049 ns 25313647 ns 28 bytes_per_second=1.88302M/s
items_per_second=39.4897k/s
MinimalEndToEndBench/num_batches:1000/batch_size:100/use_executor:1/real_time
24240980 ns 7592179 ns 28 bytes_per_second=19.6707M/s
items_per_second=41.2525k/s
MinimalEndToEndBench/num_batches:1000/batch_size:100/use_executor:0/real_time
26440865 ns 26432060 ns 26 bytes_per_second=18.0341M/s
items_per_second=37.8202k/s
MinimalEndToEndBench/num_batches:1000/batch_size:1000/use_executor:1/real_time
24367160 ns 7565095 ns 29 bytes_per_second=195.688M/s
items_per_second=41.0388k/s
MinimalEndToEndBench/num_batches:1000/batch_size:1000/use_executor:0/real_time
40216462 ns 40199735 ns 17 bytes_per_second=118.568M/s
items_per_second=24.8654k/s
MinimalEndToEndBench/num_batches:10000/batch_size:10/use_executor:1/real_time
227825184 ns 68188690 ns 3 bytes_per_second=2.093M/s
items_per_second=43.8933k/s
MinimalEndToEndBench/num_batches:10000/batch_size:10/use_executor:0/real_time
275693035 ns 274119580 ns 3 bytes_per_second=1.72959M/s
items_per_second=36.2722k/s
MinimalEndToEndBench/num_batches:10000/batch_size:100/use_executor:1/real_time
226319606 ns 74475842 ns 3 bytes_per_second=21.0692M/s
items_per_second=44.1853k/s
MinimalEndToEndBench/num_batches:10000/batch_size:100/use_executor:0/real_time
289676886 ns 289590532 ns 2 bytes_per_second=16.461M/s
items_per_second=34.5212k/s
MinimalEndToEndBench/num_batches:10000/batch_size:1000/use_executor:1/real_time
236291693 ns 69614625 ns 3 bytes_per_second=201.8M/s
items_per_second=42.3206k/s
MinimalEndToEndBench/num_batches:10000/batch_size:1000/use_executor:0/real_time
396367253 ns 396224612 ns 2 bytes_per_second=120.302M/s
items_per_second=25.2291k/s
MinimalEndToEndBench/num_batches:100000/batch_size:10/use_executor:1/real_time
2060422883 ns 584906294 ns 1 bytes_per_second=2.31427M/s
items_per_second=48.5337k/s
MinimalEndToEndBench/num_batches:100000/batch_size:10/use_executor:0/real_time
2297644557 ns 2296919513 ns 1 bytes_per_second=2.07533M/s
items_per_second=43.5228k/s
MinimalEndToEndBench/num_batches:100000/batch_size:100/use_executor:1/real_time
2303122407 ns 645052654 ns 1 bytes_per_second=20.7039M/s
items_per_second=43.4193k/s
MinimalEndToEndBench/num_batches:100000/batch_size:100/use_executor:0/real_time
2475248324 ns 2474333854 ns 1 bytes_per_second=19.2642M/s
items_per_second=40.4k/s
MinimalEndToEndBench/num_batches:100000/batch_size:1000/use_executor:1/real_time
2235447040 ns 650730679 ns 1 bytes_per_second=213.307M/s
items_per_second=44.7338k/s
MinimalEndToEndBench/num_batches:100000/batch_size:1000/use_executor:0/real_time
3766306440 ns 3764680381 ns 1 bytes_per_second=126.606M/s
items_per_second=26.5512k/s
```
As you can see again performance in async mode when the workload is small
is not good but when the workload is huge the performance tends to be better.
--
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]
