Surprisingly replacing
INLINE uint64_t PRFX(NH_PFX,,Hash_)(uint64_t key) {
key ^= key >> 33;
key *= 0xff51afd7ed558ccd;
key ^= key >> 33;
key *= 0xc4ceb9fe1a85ec53;
key ^= key >> 33;
return key;
}
with
INLINE uint64_t PRFX(NH_PFX,,Hash_)(uint64_t key) {
return (key ^ (key >> 33)) * 0xff51afd7ed558ccd;
}
Makes the code setting 2**32 bits 20% faster.
Counter intuitive, since every one recommends stronger hash functions.
Guess that is another "security expert" thingie.
Because some hacker can spam your hashmap with collisions.
Which never happens in reality, because nobody cares about my apps.
Lesson learned: pay no attention to retards, especially renowned PhD retards.
Current Mood: amused