[icsa]

I don't think so. Many of K's primitives operate on arrays using stride 1 access. Being cache-friendly accounts for much of the performance gains.

In addition, K is written in ANSI C to maintain portability.