[rc5] RC5 Chips?

Dan Sugalski sugalskd at osshe.edu
Fri Oct 31 15:27:31 EST 1997


At 02:14 PM 10/31/1997 -0800, Douglas Bitting wrote:
>OK, someone must have already thought of this, 'cuz I'm not terribly
>original...
>
>How hard would it be to fab a general RC5 chip which could chew through ~1
>key/clock? (imagine a chip that chews through 500Mkeys/sec).  Throw that
>chip (or a few) on a PCI card, write a driver, and hook that up to a client,
>and chew through something on order of N blocks per sec. per chip? (Where N
>is dependent upon clock speed and number of chips on a card).
>
>Obviously, this gets slightly beyond the "free, idle CPU cycles", but still
>would fall under the "distributed" effort.  This particular techno-geek
>finds the idea compelling... :-)
>
>Anyone thought this through beyond the wild fantasy idea stage?

Somebody probably has. There are already DES chips out there (you can burn
it into an ASIC), so RC5 ought to be reasonably simple. Probably not
key/cycle, but still pretty fast, 'specially if you pipeline the operations.

Which is not to say that I have any idea of the specifics--I don't. But
it's certainly easy enough. (Heck, compared to the x86 emulation built into
the current crop of Pentiums, this'd be a piece of cake)

					Dan

----------------------------------------"it's like this"-------------------
Dan Sugalski                            even samurai
sugalskd at osshe.edu                      have teddy bears
                                        and even the teddy bears
                                        get drunk
----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list