Look at 2.cl, though: the lisp solution is faster than everything except one c++ solution. (And, aside from the SIMD intrinsics, the lisp solution is fairly idiomatic)

I was referring to this with the lock-free hash maps: https://twitter.com/nodefunallowed/status/137196906733924761...

> I was referring to this with the lock-free hash maps…

Well thank you for providing an actual reference.

afaict from a twitter thread, "42nd At Threadmill" and "Luckless" are both Lisp re-implementations of the same Java hashtable code.

afaict the Rust sofware is not a re-implementation of that same Java hashtable code.

afaict that chart does not show any measurements of Java software, just Lisp and Rust.

So "… Java can match or exceed Rust …" seems to be based on nothing.

> Look at 2.cl, though…

So hand-coded AVX is hand-coded AVX in any language?