Open kilork opened 3 years ago
We need also measure following characteristics for test dataset:
What else?
There should be 2 performance tests, large files (>1MB), and small files (<1KB). This should better gauge overhead and improvements from 0 copy body.
Created docker images for https://github.com/koute/memory-profiler.
Repositories:
Created repository with prepared static data for benchmark.
https://github.com/kilork/actix-web-static-files-bench
This has already some benefit - my tests on Apple M1 shows I cannot bundle more than 256 Mb.
Need to clarify, is there a possibility to pack more or it is not depending from us.
An additional find: change-detection
is really helpful and allows to solve issue with deleted data from static resources. Without change detection by default cargo does not trigger build script, this leads to compile failure.
Last outcome - rust fails to build executable with current approach somewhere between 256Mb and 512 Mb.
To check: https://github.com/fcsonline/drill for load testing
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
We are missing benchmarking suite in project. We need to have some numbers to provide this for comparision.
Related to #17 #10