it possible to update the performance test results? and my latest performance test results #17
Description
Hello ceki ! the official performance report of Logback should be based on test data from 4 years ago. Do you have the latest test data based on the Master branch?
I cloned the Master branch code locally and changed the reloadj version from 1.2.20-SNAPSHOT snapshot version to 1.2.26. The pressure test results based on JDK 21.05 are as follows:
| type | version | sync score（ops/ms） | async score（ops/ms） |
|---|---|---|---|
| Logback | 1.5.12 | 1,699.46 | 1,844.13 |
| Log4j2 | 2.17.2 | 1,943.24 | 1,048.81 |
1. Operating environment and configuration
- Laptop configuration: I5-1350P CPU, 32GB DDR5 5200 RAM, and Samsung MZVL4512HBLU-00BLL 512GB SSD.
- The two running parameters of JMH are:
-jvmArgs "-Xmx2g -Xms2g" -f 2 -t 16 -w 10 -wi 2 -r 30 -i 2 -to 300 -prof gc -rf jsonand the parameters for JMH to run once is:-f 1 -tu ms -wi 2 -i 4 -to 3 -t 16. But the results of the three runs are similar. - Perform three performance tests and calculate the average value.
2. Detail log
3. Result
Logback's asynchronous performance is higher than Log4j2, by about 43.1%; The synchronization performance of Logback is lower than Log4j2, by about 14.3%.
After the software version update in the Master branch code, the performance gap is not as significant as advertised in the documentation, and Log4j2 synchronization performance is even better than Logback.
4. Suggestion:
Could you retest and update the documentation according to the latest software version?