[rc5] Re: Supercomputers & Bovine, Q&A
vorlon at dodds.net
Tue Aug 5 03:04:46 EDT 1997
On Tue, 5 Aug 1997, root wrote:
> > Prove that RC5 cracking is a useful benchmark for determining real world
> > performance with respect to mainstream applications, and I'll quote you
> > some numbers..
> You are here. The Rc5 is used to protect information every day. If that
> is not "real world" or mainstream performance and your own admission of
> it, I don't know what is.
Oy, some people. RC5 as a mainstream application? I'll grant you that
it's used, although not as commonly as some other encryption algoritms
like DES or PGP (The only code anywhere on my system that decrypts RC5 is
the Bovine client), but most people I know use their computers for
something other than encrypting and decrypting data. That is not, as they
To extend your car analogy (which seems, btw, a little backwards: it's NT,
not Unix, that comes equipped with flashy excesses designed to please the
user at the expense of an appalling amount of resources--that, and the
day Gates is quoted as saying 'Quality is Job 1,' I'm moving far FAR
away), choosing a computer solely on the basis of how well it does at
decrypting RC5 is like buying a new vehicle because it has the style of
drink holder you like. Sure, it may be nice to show your friends, but it
certainly doesn't guarantee it's going to get you where you need to go.
So no, it's not a very useful benchmark. Encryption makes up a negligible
portion of the number of instructions executed by a non-Bovine computer
each day, and unless you work for the NSA <peers suspiciously at the
headers>, it's no basis on which to choose an architecture.
(And lest this spark any more comments about who did or didn't
say/mean/imply what, regardless of who actually started it, the claim
*was* made that RC5 proves x86 is better than alpha, and this discussion
is a natural result of that claim. So.)
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.
More information about the rc5