jmeter-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From GitBox <...@apache.org>
Subject [GitHub] [jmeter] vlsi commented on pull request #603: Bug 64558 - Improve performances and throughput of Sample Results by lifting contention on writing SampleResults in CSV/XML
Date Sat, 11 Jul 2020 13:16:10 GMT

vlsi commented on pull request #603:
URL: https://github.com/apache/jmeter/pull/603#issuecomment-657060836


   I've added 10ms sleep test as well: `-Jjmeter.save.queue.size=524288 -Jthreads=2000 -Jduration=120
-Jrampup=1`
   
   before:
   
   ```
   summary + 191792 in 00:00:21 = 9219.0/s Avg:   156 Min:    10 Max:   991 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary + 317236 in 00:00:30 = 10574.5/s Avg:   152 Min:    10 Max:   548 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary = 509028 in 00:00:51 = 10019.4/s Avg:   153 Min:    10 Max:   991 Err:     0 (0.00%)
   summary + 314744 in 00:00:30 = 10491.5/s Avg:   155 Min:    10 Max:  1218 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary = 823772 in 00:01:21 = 10194.7/s Avg:   154 Min:    10 Max:  1218 Err:     0 (0.00%)
   summary + 347093 in 00:00:30 = 11569.8/s Avg:   153 Min:    10 Max:   785 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary = 1170865 in 00:01:51 = 10567.0/s Avg:   154 Min:    10 Max:  1218 Err:     0 (0.00%)
   summary +  96680 in 00:00:10 = 9289.9/s Avg:   160 Min:    10 Max:  1339 Err:     0 (0.00%)
Active: 0 Started: 2000 Finished: 2000
   summary = 1267545 in 00:02:01 = 10457.3/s Avg:   154 Min:    10 Max:  1339 Err:     0 (0.00%)
   ```
   
   after:
   
   ```
   summary +      1 in 00:00:00 =    9.1/s Avg:    13 Min:    13 Max:    13 Err:     0 (0.00%)
Active: 90 Started: 90 Finished: 0
   summary + 320825 in 00:00:27 = 11825.9/s Avg:   157 Min:    10 Max:   884 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary = 320826 in 00:00:27 = 11778.2/s Avg:   157 Min:    10 Max:   884 Err:     0 (0.00%)
   summary + 371391 in 00:00:30 = 12342.3/s Avg:   157 Min:    10 Max:   946 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary = 692217 in 00:00:57 = 12074.3/s Avg:   157 Min:    10 Max:   946 Err:     0 (0.00%)
   summary + 375868 in 00:00:30 = 12513.5/s Avg:   157 Min:    10 Max:   954 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary = 1068085 in 00:01:27 = 12225.3/s Avg:   157 Min:    10 Max:   954 Err:     0 (0.00%)
   summary + 364263 in 00:00:30 = 12148.2/s Avg:   159 Min:    10 Max:  1309 Err:     0 (0.00%)
Active: 2000 Started: 2000 Finished: 0
   summary = 1432348 in 00:01:57 = 12205.6/s Avg:   157 Min:    10 Max:  1309 Err:     0 (0.00%)
   summary +  38372 in 00:00:04 = 10097.9/s Avg:   167 Min:    11 Max:   566 Err:     0 (0.00%)
Active: 0 Started: 2000 Finished: 2000
   summary = 1470720 in 00:02:01 = 12139.5/s Avg:   158 Min:    10 Max:  1309 Err:     0 (0.00%)
   ```
   
   That is roughly 10K -> 12K which is 20% improvement for the very limited case when Test
plan is doing nothing.
   
   
   
   I still insist that `sleep` is not a realistic test plan because it has no business logic.
The web plan would perform HTTP request send/receive which consumes CPU, and it makes the
contention on the result saver much less to be an issue.


----------------------------------------------------------------
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:
users@infra.apache.org



Mime
View raw message