vocal mist
I intend to deploy a headless bevy server. I have a choice of many instance types and instance sizes (across multiple providers). I have a pretty good idea which ones should perform better, but I need to back it up with empirical evidence by running a benchmark across the different CPU types. I might even compare ARM and x86.
I can see there's a benches directory. Is it be sufficient to run cargo bench there?
light egret
That might provide a good baseline... but performance testing not tailored for your app will be likely to NOT give you the correct result (I imagine RAM speed vs CPU speed (vs disk ; network ; etc etc...) could drastically impact performance, depending on what your app is bottlenecked on).
But in the end I guess it doesn't matter much, run the benches and go from there, keep a note to maybe adjust when you have more realistic benchmarks 🙂