Yeah, this would be great! Currently only 128-bit SSE/NEON is working but AVX is coming very soon. There's also nothing blocking Windows, but it will require some work. (I added the SIMD support in OxCaml)
I was actually wondering about that - it appears a (i+i^2)/2 sequence makes insertions (and by extension erases with rehashing) 7-10% faster, which is pretty significant. Lookups and probe lengths are about the same, so I think the conclusions stand.
Agreed - I do mention in the post that open addressing is impractical when using intrusive linked lists, which are common in low level data structures.
I just benchmarked absl::flat_hash_map and got results comparable to Robin Hood with a load factor between 75% and 90%, which makes sense. It's also faster for looking up missing keys, so seems like a good option. I didn't benchmark the maximum probe lengths, though, so not sure on that front.