i see blake3
https://docs.rs/blake3/latest/blake3/struct.Hasher.html

Hasher implementsstd::io::Write, so it’s possible to usestd::io::copy to update aHasher from any reader. Unfortunately, this standard approach can limit performance, because copy currently uses an internal 8 KiB buffer that isn’t big enough to take advantage of all SIMD instruction sets. (In particular,AVX-512 needs a 16 KiB buffer.) update_reader avoids this performance problem and is slightly more convenient.

I want ask, how many bytes use write() each time that are most efficient when use gxhash?