Re: Good hash for pointers

Am 09.06.2024 um 13:40 schrieb Richard Harnden:

On 09/06/2024 12:35, Bonita Montero wrote:

With the code I've shown I proved that the hash AES hashfunction
is totally stochastically and thereby a good hash function. But
I was still discontent with the speed of the code.
Usually you're not able to update the buckets of a hashtable with
multiple threads without any locking. But with my code the buckets
are only simple counters which could be updated atomically.
So I could reduce the code and partitioned the input range to the
hash function with the available number of cores. Here's the code:
>
#include <iostream>

 This is not C

Its about the general principle of Michael's hash and how to prove that
it has a distibution of the bucket loads which is similar to a random
number generator. I did it in C++ because that's only half the code.