Alexandrescu’s Dictum: "You must measure everything". (https://fasiha.github.io/post/replicating-matlab-columns/#you-must-measure-everything)
I thought I was being sooooooo clever here:
`varZ = var(inrange; mean=meanZ)`
giving #Julia `var` (sample variance) a pre-computed mean I had lying around. "It's gotta be faster right?"
Removed the `mean=meanZ`, letting `var` calculate the mean: runtime of my whole program is 8% faster, with *half* as many ephemeral allocations.
I am impressed with Julia's @code_warntype and @time macros
https://docs.julialang.org/en/release-0.6/manual/performance-tips/#man-code-warntype-1
@22
Early optimization is the root of all evils.
Every single time.