YangKeao
YangKeao
Maybe we could use quantile regression.
@Eh2406 As I know criterion don't throw any outliers. It will only warn you. I have a [forked branch of criterion](https://github.com/YangKeao/criterion.rs/tree/remove_outliers) which deletes outliers. (It can only work with `--noplot`...
@gz Thanks for sharing the paper! It inspired me a lot! @bheisler I have a modified version of [criterion.rs](https://github.com/YangKeao/criterion.rs/tree/quantile_regression) which used quantile_regression (I have also developed a crate [langbardo](https://github.com/YangKeao/langbardo/blob/master/src/lib.rs) to...
@gereeter Thanks for your proof. I realized that even with OLS, the unbiasness of slope can be also derived from the i.d.d., which means if all bias has the same...
@bheisler @hdevalence Maybe you are interested. I just implemented a [POSIX CPU time measurement](https://github.com/YangKeao/criterion-cpu-time) today (README and introduction will be added tomorrow). It's surprising that implementing a measurement takes only...
@bheisler Migrate from 0.2 to 0.3 is painful. One of the reason is adding generic parameter everywhere :cry: (though this is reasonable). I have done [this](https://github.com/tikv/tikv/pull/5336) for a quite big...
@bheisler I noticed that the only way to apply multiple input on a function is loop if I didn't miss some important functions. I have to change from: ```{rust} for...
@bheisler I have noticed that criterion provides a `Report` trait. Should we add method for adding report methods into reports list? Maybe simply: ```rust impl Criterion { fn add_report(&mut self,...