[rc5] Modified client assurance?

Remi Guyomarch rguyom at mail.dotcom.fr
Tue Jul 8 02:12:34 EDT 1997

RSA Crack Team wrote:
> > Sorry but at that time, you couldn't have any assurance that my client
> > is performing calculations right. If this ASM is integrated in a future
> > (2.01?) official client, you will have to trust official RC5 coders.
> > In the midtime, you will have to trust me, to check the code, to simply
> > ignore it, or to wait for somebody to review my code.
> >
> Don't you believe in the results of using
> rc5 -a rc5proxy.distributed.net -p 3056 email ?
> It's working fine with your clients, and I think it's a good way to test
> the clients.

At the time I wrote this, I wasn't aware of this behaviour. Later I
checked my client with the proxy test port and it worked fine.

I played a bit more on ASM with 6x86 and K5 optimizations in mind, and
put some test clients on my web site
(http://wwwperso.hol.fr/~guyom001/). Since I can't test them on real
6x86 or K5, I can't say exactly if they perform better than v2 clients.
I've just downloaded some docs from Cyrix and AMD, and played a bit with
ASM... :-)

Someone has already tested the Cyrix client, and it seems to be sligthly
faster than v2 (246 vs 243 kkeys/s on a 166+). Anyone to test the K5
version ?

I'm currently working on a two-keys-at-once crypt routine, that should
perform better on Pentiums, PPro, 6x86, k5 and superscalar risc machines
(Alpha UltraSparc or R10000, anyone ?).
As always, I will release full source code and binaries. Stay tunned !


To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.

More information about the rc5 mailing list