hndrk
hndrk
Good point! I will quickly create a PR for it 👍
Subtracting the average no-op time from the average benchmark time sounds reasonable. We could also ship this via an experimental flag in the config, if its not set then we...
Should we compare this feature with all [build options](https://ziglang.org/documentation/0.12.0/#toc-Build-Mode)? I think you mentioned that you tested this in debug mode.
Sure I will test this on mac 👍 For the windows case, we could implement a generic experimental flag. This would allow us to enable the feature on linux and...
Here are the [mac baseline results](https://gist.github.com/hendriknielaender/5a23094f56d63eb0c7114720a6b2e08b) they were executed with an old 2016 macbook on macOS 12.7. I could also add another test with a newer macbook pro, but earliest...
You are right, I think we need to move the timer instantiation out of the `runImpl` function and move it into a higher scope where it is created once and...
@FObersteiner how we wanna continue with this PR? I really like the feature, to have a more precise benchmark.
> > @FObersteiner how we wanna continue with this PR? I really like the feature, to have a more precise benchmark. > > To be honest, I didn't reach a...
@bens thank you for your review 👍 i will try to address these points tomorrow.
> The thing that makes me uncertain about this overall change is that if two benchmarks have different numbers of iterations then they're not comparable. I've been reading up on...