[Hardware] "success"

[Martin Klingensmith]

Well Kris,
There are a lot of devices that can do it, dare I say just about any 
FPGA. But to do it fast you need a LOT of registers one way or another - 
pipelined or paralleled.
The algorithm I have nearly complete requires a device slightly larger 
than a Spartan 3 - 1.6 million "gates" device. It would only run at 55 
MHz but would run one key per clock cycle so it would theoretically do 
55 million keys per second.
This is just an estimate of course. Flame suit on..
--
Martin K

Kris Amy wrote:
> Hi Folks,
>
> What is actually required in terms of hardware for this?
>
> I'm just browsing around on that website and unsure of what you need.
>
> Cheers,
> Kris
>
> -----Original Message-----
> From: gmeurice at dice.ucl.ac.be [mailto:gmeurice at dice.ucl.ac.be] 
> Sent: Friday, 20 October 2006 11:11 PM
> To: Hardware
> Subject: Re: [Hardware] "success"
>
> Sorry, wrong "from" adress
>
> This is a forwarded message
> From: Guerric Meurice de Dormale <guerric.meurice at uclouvain.be>
> To: Hardware <hardware at lists.distributed.net>
> Date: Thursday, October 19, 2006, 4:23:06 PM
> Subject: [Hardware] "success"
>
> ===8<==============Original message text===============
> Hello,
>
> Ok, I guess everything is clear between you and John now.
>
> Talking about devices, if the biggest device you "can" (afford) work
> with is a Spartan3E, there is maximum 14.5 kSlices and 36 bRAMS.
> >From my point of view, it is not possible to fit a fully unrolled RC5
> core with such amount of slices. (or I will be very happy to know
> how!).
> For such kind of devices, a partially unrolled core should probably be
> the best solution (with let's say 3 key schedule stages and 1 decrypt
> stage).
>
> I'm not only saying that as I'm working on fully unrolled core and
> with bigger FPGA: we worked on the problem during the first half of
> the year and I really can't figure out how to reduce further our
> ~17500 slices. I'm currently adding other optimizations, but only a
> couple of hundred slices are going to be saved.
>
> Does a partially unrolled core can fit with the schedule of your work
> ? (I'm not saying you to do that, just thinking about what's
> possible)
>
>