4 ms·
Neat, the way I benchmarked was through BenchmarkTools.jl, which runs the functions sufficiently many times to get a statistically relevant result. It will not
by stabbles 6y ago
Neat, the way I benchmarked was through BenchmarkTools.jl, which runs the functions sufficiently many times to get a statistically relevant result. It will not measure compilation time, but I did not include that for gcc / clang either.
- hpcjoe 6y agoI kept that in simply to keep the "playing field level", as I run Julia code as "scripts", knowing full well that it is far more powerful than that. I do use BenchmarkTools.jl as well, though I've found TimerOutputs.jl to be quite useful in my optimization/debugging.