Re: AES problem (was: Good hash for pointers)

Michael S <already5chosen@yahoo.com> writes:

On Wed, 05 Jun 2024 08:58:46 -0700
Tim Rentsch <tr.17687@z991.linuxsc.com> wrote:
>

Michael S <already5chosen@yahoo.com> writes:
>
>
I couldn't get the AES-based hash function to compile.  There
was a complaint about not being able to inline an 'always-inline'
function.  I played around with it for a while but didn't get
anywhere.

>
Not that using my reference hash is particularly important (I needed it
as a quick way to get something I could rely on;  you already have
something else with similar properties) but it's still worth
investigation.
>
There are two possibilities:
1. gcc compiler's -march=native logic incorrectly detects CPU type on
your computer.
2. you CPU really does not support AES instructions.
>
If the former, you can force compiler's hand with -maes flag.
If the later, then you can compile it, but wouldn't be able to run my
"reference" hash function on this particular computer

I finally got it to work.  I tried -maes before, and it still
failed, but maybe I was using some other option that messed
things up.  In any case it works now, and the AES hash gives
great results AFAICT.