Comment by sva_

Good article, but I think the multiply_naive should've used numpy instead for a more fair benchmark comparison.