vlsi edited a comment on pull request #603:
URL: https://github.com/apache/jmeter/pull/603#issuecomment-657057936
I've tried the following:
1. I've installed a local `nginx` that serves `Apache JMeter` website
1. Then I created a test case that opens the home page, then it extracts
"getting started" URL and opens it, then it extracts "user manual" URL and
opens it. All the actions are run in a transaction controller.
1. I launched the benchmark as follows
../bin/jmeter -n -t jmeter_home_page.jmx
-Jjmeter.save.queue.size=524288 -Jthreads=2000 -Jduration=120 -Jrampup=1 -l
results.csv
The results are:
Before patch:
```
Starting standalone test @ Sat Jul 11 15:32:08 MSK 2020 (1594470728843)
summary + 76492 in 00:00:21 = 3650.8/s Avg: 506 Min: 1 Max: 6071
Err: 1101 (1.44%) Active: 2000 Started: 2000 Finished: 0
summary + 147463 in 00:00:30 = 4926.1/s Avg: 405 Min: 301 Max: 727
Err: 0 (0.00%) Active: 2000 Started: 2000 Finished: 0
summary = 223955 in 00:00:51 = 4400.9/s Avg: 439 Min: 1 Max: 6071
Err: 1101 (0.49%)
summary + 145472 in 00:00:30 = 4852.1/s Avg: 409 Min: 3 Max: 15673
Err: 1838 (1.26%) Active: 2000 Started: 2000 Finished: 0
summary = 369427 in 00:01:21 = 4568.2/s Avg: 427 Min: 1 Max: 15673
Err: 2939 (0.80%)
summary + 135559 in 00:00:30 = 4518.6/s Avg: 445 Min: 310 Max: 11305
Err: 0 (0.00%) Active: 2000 Started: 2000 Finished: 0
summary = 504986 in 00:01:51 = 4554.8/s Avg: 432 Min: 1 Max: 15673
Err: 2939 (0.58%)
summary + 58676 in 00:00:12 = 4981.4/s Avg: 366 Min: 0 Max: 733
Err: 0 (0.00%) Active: 0 Started: 2000 Finished: 2000
summary = 563662 in 00:02:03 = 4595.8/s Avg: 425 Min: 0 Max: 15673
Err: 2939 (0.52%)
Tidying up ... @ Sat Jul 11 15:34:11 MSK 2020 (1594470851780)
```
After patch:
```
Starting standalone test @ Sat Jul 11 15:40:24 MSK 2020 (1594471224569)
summary + 9043 in 00:00:05 = 1761.7/s Avg: 679 Min: 1 Max: 3400
Err: 193 (2.13%) Active: 2000 Started: 2000 Finished: 0
summary + 146023 in 00:00:30 = 4865.3/s Avg: 416 Min: 127 Max: 4879
Err: 2 (0.00%) Active: 2000 Started: 2000 Finished: 0
summary = 155066 in 00:00:35 = 4412.1/s Avg: 431 Min: 1 Max: 4879
Err: 195 (0.13%)
summary + 137837 in 00:00:30 = 4601.0/s Avg: 416 Min: 28 Max: 4500
Err: 758 (0.55%) Active: 2000 Started: 2000 Finished: 0
summary = 292903 in 00:01:05 = 4499.0/s Avg: 424 Min: 1 Max: 4879
Err: 953 (0.33%)
summary + 130667 in 00:00:30 = 4355.9/s Avg: 478 Min: 2 Max: 24626
Err: 1996 (1.53%) Active: 2000 Started: 2000 Finished: 0
summary = 423570 in 00:01:35 = 4453.8/s Avg: 441 Min: 1 Max: 24626
Err: 2949 (0.70%)
summary + 104537 in 00:00:28 = 3793.1/s Avg: 508 Min: 0 Max: 11232
Err: 0 (0.00%) Active: 0 Started: 2000 Finished: 2000
summary = 528107 in 00:02:03 = 4305.4/s Avg: 454 Min: 0 Max: 24626
Err: 2949 (0.56%)
Tidying up ... @ Sat Jul 11 15:42:27 MSK 2020 (1594471347561)
```
It does not look like the patch improves the performance.
[jmeter_home_page.jmx.txt](https://github.com/apache/jmeter/files/4906775/jmeter_home_page.jmx.txt)
<img width="872" alt="test plan overview"
src="https://user-images.githubusercontent.com/213894/87224400-c1f3b200-c38d-11ea-880a-66d0605008d8.png">
---
HW: MacBook late 2013, 4 core, 8 threads
Java: 11
`nginx.conf`
```
events {
worker_connections 2024;
}
http {
include mime.types;
default_type application/octet-stream;
#access_log logs/access.log main;
sendfile on;
#tcp_nopush on;
#keepalive_timeout 0;
keepalive_requests 5000;
keepalive_timeout 65;
#gzip on;
server {
listen 32768 backlog=16384;
server_name localhost;
#charset koi8-r;
#access_log logs/host.access.log main;
location / {
#root share/nginx/html;
root www/jmeter;
index index.html index.htm;
}
```
----------------------------------------------------------------
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]