You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
- LocalExec increase thread possibility
- Change minimal / maximal threads for retrieve
- Increase runlimit to 50 000
- Improve Client Thread / Server Thread
- Add Benchmark tests on Multiple Server
- Fix issue waarp#90 on Oracle JDBC driver from Maven Central inspired from Pull Request waarp#92
- Fix issue waarp#92 on profile option description in README.md, inspired from Pull Request waarp#93
Results:
- Loop Benchmarks gives on 2 cores (4 threads) with 2 Waarp Servers, 1 with PostgreSQL as Container, 1 with H2 Database on the same server
- TLS efficiency: up to 84 trasfers/second
- No TLS efficiency: up to 103 transfers/second
- Dual Waarp servers (using 1 core for Network) on the same 2 cores with the same PostgreSQL Database as Container and Filesystem with one Client on the same server
- 1 server: 42 transfers/second with 75% CPU usage and 90 Mbps
- 2 servers: 66 transfers/second, so 57% more with 100% CPU usage and 142 Mbps
- Note on benchmark:
- Numbers are lower than loop Benchmark since clients are in a lot of concurrent client threads, therefore having difficulties to have high concurrency on client side and much more CPU activities due to 2000 threads
- Difference between 1 and 2 servers is low since all are running on the same host, using in 2 servers case 100% CPU, while 75% in 1 server case, therefore the efficiency is not fully visible since there is not enough CPU available but still 57% more. One can expect at least 90% of linearity (almost doubling if we were using 2 different hosts for both servers in the same cluster).
Is there any commercial support available for this project
The text was updated successfully, but these errors were encountered: