Replies: 1 comment 1 reply
-
|
This is mainly because Lighttpd is an I/O intensive workload that involves extensive OCALLs, see https://gramine.readthedocs.io/en/latest/performance.html#effects-of-system-calls-ocalls for details. I believe there is room for optimization, but this would require more detailed profiling. Can you maybe share more details w/ us? |
Beta Was this translation helpful? Give feedback.
1 reply
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
I directly tested the performance of lighttpd on Linux and Gramine, but the difference was 10 times. Why is this?
wrk -c 30 -d 30 -t 1 http://127.0.0.1:8003/random/10K.1.html
Running 30s test @ http://127.0.0.1:8003/random/10K.1.html
1 threads and 30 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 169.27ms 327.21ms 1.94s 87.21%
Req/Sec 10.48k 54.20 10.61k 56.81%
313825 requests in 30.10s, 3.06GB read
Requests/sec: 10426.02
Transfer/sec: 103.98MB
make start-native-server
wrk -c 30 -d 30 -t 1 http://127.0.0.1:8003/random/10K.1.html
Running 30s test @ http://127.0.0.1:8003/random/10K.1.html
1 threads and 30 connections
Thread Stats Avg Stdev Max +/- Stdev
Latency 268.50us 57.31us 768.00us 92.79%
Req/Sec 111.55k 1.14k 115.05k 66.78%
3340805 requests in 30.10s, 32.54GB read
Requests/sec: 110990.21
Transfer/sec: 1.08GB
Beta Was this translation helpful? Give feedback.
All reactions