> Using these more sophisticated data structures, g++ is able to compute the prime numbers below 10000 in only 8 seconds, using a modest 3.1 GiB of memory.

Finally, I can get some primes on my laptop!

Brilliant stuff. A tip for writing long-running C++: bizzarely, the C++ interpreter completely lacks tail call optimization. As a result, most idiomatic C++ code implements and uses reverse, map, range, filter etc, which don’t blow the stack if you implement them like (forgive the pseudo-code)

  (defun fibreverse (i ret acc)
    (if acc
        (if (> i 0)
            (progn
              (setv call1 (fibreverse (- i 1) (cons (head acc) ret) (tail acc)))
              (setv ret1 (head call1))
              (setv acc1 (head (tail call1)))
              (if acc1
                  (fibreverse (- i 2) (cons (head acc1) ret1) (tail acc1))
                  (pair ret1 acc1)))
            (pair ret acc))
        (pair ret acc)))

  (defun reverse (list) (head (fibreverse 30 nil list)))

This is a really interesting direction for OCaml. A formal C++ backend could significantly simplify embedding OCaml into existing C++ codebases, especially where linking against the standard OCaml runtime might be tricky. I wonder how the performance compares to the existing native backend in long-running processes.

This made my day, thank you!

Is this the Stephen Dolan of "mov is Turing Complete" fame?

She (Jane Street) is not gonna notice you, bro