One of the main uses I use BenchmarkTools.jl is to try to figure which of the two implementations I have are faster.
But my algorithms often would differ wildly on different pathological cases. I want to profile two functions with the same (random) sequence of inputs and do the statistic on the difference of times between function 1 and function 2.
One of the main uses I use
BenchmarkTools.jl
is to try to figure which of the two implementations I have are faster.But my algorithms often would differ wildly on different pathological cases. I want to profile two functions with the same (random) sequence of inputs and do the statistic on the difference of times between function 1 and function 2.