[RC5] Performance on P-IV
enojon at ATTGLOBAL.NET
Sat May 12 16:42:21 EDT 2001
No, but then again, I didn't read up on the RC5 alogrithm documentation.
Q= P/ SHR( !?!?!?
Apparently RC5 is vectorizable in that you can take 16 or 32 keys as
vectors and divide into the vector elements, shift the vector elements,
rotate the vector elements and voilla--$2,000 later!
Vector processors are very good for the mathematics of Rc5 "cracking".
These take several values and perform same operation simultaneously [in
clock cycle terms], but are costly in terms of requiring constant data
feed and bandwidth.
----- Original Message -----
From: "Dean Morrell" <dmorrell at auntvi.com>
To: <rc5 at lists.distributed.net>
Sent: Friday, May 11, 2001 9:37 AM
Subject: RE: [RC5] Performance on P-IV
> At 05:18 PM 5/10/2001 -0400, you wrote:
> >RC5 isn't truly vectorizable,
> >because of inter-loop dependencies. Does the dnet stuff break it down
> >into parallel tasks using the vector registers each as a separate
> Wow! Am I the only one that doesn't understand a word of this?!?
> "When a man assumes a public trust, he should consider himself as public
> -Thomas Jefferson
> To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
> rc5-digest subscribers replace rc5 with rc5-digest
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5