Generators vs loops vs broadcasting: Calculate PI via Monte Carlo Sampling

General Usage Performance
2

You forgot to include the fastest solution which fully removes the allocations by sticking to scalar calculations:

function f(N)
    cnt = 0;
    for i in 1:N
        if rand()^2 + rand()^2 <= 1
            cnt += 1
        end
    end
    pi = cnt/N * 4
end

using BenchmarkTools
n = 100000
@btime $f($n) # 611.200 μs (0 allocations: 0 bytes)

https://github.com/JuliaSIMD/VectorizedRNG.jl could be a nice asset for optimizing this example as well.

8 Likes