I had some fun the other day and ported a parallel raytracer benchmark from F# to multicore OCaml: https://github.com/athas/raytracers
It doesn’t perform well, but the worst cases are on rendering, where I just naively spawned a Domain per pixel (: Chunking per row of pixels worsened performance even more though.
If you know how to show better performance with multicore, then make a PR - he will gladly accept