6 May
2007
6 May
'07
3:54 p.m.
On Saturday 05 May 2007 21:02, Richard wrote:
In article <463C9ECC.16427.23507AFE at
cclist.sydex.com>,
"Chuck Guzis" <cclist at sydex.com> writes:
Sure, but no one (in their right mind) builds a general purpose CPU out
of them. Well, unless it's a research or "toy" project, a prototype
design, or something very unusual and high end[1], but even in that case
it's still not making a general purpose CPU out of it.
You normally take an FPGA board, program your specific algorithm into it
(that is, convert the algorithm into hardware baseically), and shove
bits through it.
Have you actually used a FPGA board? I've actually taken a workshop (a
few years ago now) on them. Everyone that I know who's using FPGAs in
their product (in any sort of manner that approximates a CPU) uses them
as some sort of co-processor to offload a specific processing task from
the main CPU.
I challenge you to come up with the name of a single product that someone
can purchase right now, which uses an FPGA-implemented CPU, which is
general-purpose, and reasonably widely available.
[1] I'm specifically referring to a NAS box we just got at work. MMmmm,
NFS in hardare... see http://bluearc.com/
Pat
--
Purdue University ITAP/RCAC --- http://www.rcac.purdue.edu/
The Computer Refuge --- http://computer-refuge.org
Actually, assembly was a step up. I started by
programming machine
language. The sad thing is that architectures are far less varied
than they were 30-40 years ago. Maybe that's a good thing; maybe
not.
Au contraire. "Architecture" these days is whatever you want it to be
cooked into an FPGA. Maybe you don't consider that architecture, but
the hardware possibilities available to you for low cost these days
are limited only by your imagination and gate count. Of course
there's a trade-off between speed and versatility but FPGA designs are
*very* popular for custom architectures these days.