News for hackers. Stuff that matters.
Re:Optimizing Open Addressing by sporadicity on Sunday April 02, 2023 @02:40PM (#35413749)
Have you tried absl::flat_map? It uses simd in a different way than described in this article, and Google claims that it saves them a lot of memory because it still works pretty well at 90-95% occupancy.
Re:Optimizing Open Addressing by anonymoushn on Sunday April 02, 2023 @06:06PM (#35415773)
I've benchmarked swiss tables and found that (for hit-heavy workloads) a minimum of 2 loads per lookup is expensive compared to 1.
Re:Optimizing Open Addressing by adgjlsfhk1 on Sunday April 02, 2023 @08:34PM (#35417043)
I've really wanted to try making a hash table that works like a Swiss table but that stores the metadata interleaved with the rest of the data (16 metadata, 16 key, 16 value, repeat). doing so would keep your memory accesses closer together, but be a pain to program