stanford is optimal for integers (i.e. literals divisible by 256),
greg7mdp/parallel-hashmap for strings and structs.
Converting swiss table or greg7mdp/parallel-hashmap from C++/SIMD might be too much work.
stanford_hash is similar and good enough for now.
EDIT: see LIMachi/swiss-table on github.
Try the o1hash from wyhash as default, but add security measures (collision counting or tree conversion, no sorted_vector yet)
stanford is optimal for integers (i.e. literals divisible by 256),
greg7mdp/parallel-hashmap for strings and structs.
Converting swiss table or greg7mdp/parallel-hashmap from C++/SIMD might be too much work.
stanford_hash is similar and good enough for now.
EDIT: see LIMachi/swiss-table on github.
Try the o1hash from wyhash as default, but add security measures (collision counting or tree conversion, no sorted_vector yet)