Re: sha-512 efficiency

[Greg Troxel <gdt%lexort.com@localhost>]

Sad Clouds <cryintothebluesky%gmail.com@localhost> writes:

> On Sat, 4 Jan 2020 09:53:40 -0800
> Jordan Geoghegan <jordan%geoghegan.ca@localhost> wrote:
>
>> You can test your machines hashing speed using the command: "openssl 
>> speed sha256 sha512"
>
> I wrote a small C program where I call OpenSSL SHA512 function in a
> loop, on a string buffer of 1 KiB in size.

It seems that hash functions have some fixed time and then time per kB
(or something) for longer inputs.  So you may want to test various
sizes, and assess which sizes get used by ZFS for your filesystem
workload.

> But the speed test aside, how do you test the randomness/uniqueness of
> hash values it generates?  It is a 64-byte binary string, so can take
> quite a long time before you come across collisions.

You read the papers from the crypto community :-)