Michael Abbott

> jagged linear/log scale of rounding to 2 digits; have you generated synthetic benchmark results and see if you're happy with the centration of the histogram in all cases? Good...

~~Tests are missing but~~ it does have a special case for 0 & 1 samples. Experimenting with `◔ ◑ ◕` to mark quartiles, in a way obvious enough to avoid...

With 2/3 columns this will happily adjust down. With a huge screen, you won't have to turn your head. Sorry have not got around to typing a reply to all...

https://github.com/JuliaCI/BenchmarkTools.jl/pull/260#discussion_r720306154 is the explicit demonstration, in 55 columns. Care has been taken that line wrapping will be minimally disruptive on very narrow screens, like pasting into slack, even if initially...

My complaint remains that looking at one screenshot is pretty misleading. I think you need to use the hammer to find out what's good about it, not look at the...

> What needs to be done to resolve this? Is anyone waiting on a review or does it just lack for some free time to finish it off? Sorry I...

I think the logic of printing 3 digits is that the prefixes are separated by 3, so if you have some μs and some ms, it's easier to compare. Not...

Whole numbers for `ns` doesn't seem crazy. Could investigate something like `2.479 μs`, `82.47- μs`, `282.4-- μs`, maybe too weird? Just rounding to a fixed number of significant figures might...

I think you just want to specify `evals=1`, otherwise this is evaluated several times per setup. ```julia julia> function f!(x::AbstractVector) # length(x) == 2 || error("setup not correctly executed") sleep(length(x)/10)...

It could also say `min 1 allocation, 7.94 KiB` since that's what it is, but maybe it's usually stable enough for this not to matter. With everything now, the line...