Thanks that’s super helpful!
I did a very short research and there is a rust lib called criterion which does outlier detection. However, they explicitly states that even though outliers are detected, they are not dropped from the data. They also describe how they detect outliers here.
EDIT: Looks like you described in the issue almost the same method for detecting outliers as in the criterion 
yeah it’s the same (barring the “extreme outliers” described there), which makes sense when one bases it on established methods 
Just published benchee 1.2.0 - no huge changes, and that’s a good thing 
Also did some house keeping upgrading CI for all ~5 sibling projects. No release worthy changes though.
PSA for benchee users: Ran a benchmark on elixir 1.14.0 - 1.16.0-rc.0? The results may have been missing some optimizations, check this post for what could have been affected and remedies:
Not 100% benchee, but it originated in Benchee and details some work I’ve been doing on the upcoming 1.3.0 version to reduce memory usage and total run time when dealing with big data structures as inputs!
And now as a follow up to 
benchee 1.3.0 is out! Post includes some details about the memory savings made in this release with an example - down to 12% & others 
© Copyright Elixir Forum | Terms | Privacy & Cookies
