[104] in cryptography@c2.net mail archive
Re: RC5-12/32/5 contest solved
daemon@ATHENA.MIT.EDU (jim bell)
Thu Jan 30 10:52:26 1997
Date: Wed, 29 Jan 1997 20:42:20 -0800
To: stewarts@ix.netcom.com, Ian Goldberg <iang@cs.berkeley.edu>,
Germano Caronni <caronni@tik.ee.ethz.ch>
From: jim bell <jimbell@pacifier.com>
Cc: cryptography@c2.net, cypherpunks@toad.com
At 09:37 PM 1/28/97 -0800, stewarts@ix.netcom.com wrote:
>Any bets on whether the $5000 RC5-12/32/6 contest will be solved
>before the www.rsa.com contest status web page is updated? :-)
>
>Or how long before someone in the government starts talking about
>how 56 bits takes 65,000 times as long to solve as 40 bits,
>which is 26 years for a whole building full of computers,
>and even 48 bits ought to take a month and a half for a whole
>building full of computers (or supercomputers, if they hype it up....)?
This, as I pointed out long ago, is why I didn't think a "crack the DES key"
contest is necessarily a good idea, at least if it's ordinary
Von-Neumann-type computers doing the searching. It makes DES look
artificially good.
Assuming it's possible to build a chip which tests solutions in a
massively-pipelined mode, the 400,000 or so solutions per second tried (for
what is probably a $2000 machine) would probably increase to 100 million per
second per chip (at a cost of maybe $100 per chip, if implemented in
parallel). That's 5000 times more economical, which would translate to a
find in 2-3 days if the same dollars in hardware were invested.
_THAT_ is the break we should hope the media publicizes, not the one that
will eventually happen when accomplished by PCs or Suns, etc.
Jim Bell
jimbell@pacifier.com
