On Tue, Jan 22, 2008 at 04:11:31PM -0800, SJS wrote:
It'll fit fine in something like a Xilinx Virtex 5. Of course, if you're
spending $3,000 on just the FPGA, you're probably better off just buying a
really fast computer.
And running a simulator?
No, just running whatever problem you were trying to solve.
Are FPGAs really that slow?
Simulating a HW design will be much slower than an FPGA, often by many
orders of magnitude. Simulating a CPU isn't that bad.
The FPGAs really shine when the application is carefully tailored to them.
I would think that an open-source hardware geek might like to play with
open-sourced hardware designs.
And $3k isn't that much. I thought it would be $30k.
It ends up being quite a bit by the time you get a board to put the chip
on. The free-ish version of Xilinx software doesn't support the huge
chips, and you'll want a fast jtag programmer to avoid waiting long periods
of time to program the chip. They do make some decent starter boards that
range from $100 to $5000. The nicest ones, though, are the Digilent ones.
Most of the larger Xilinx chips have a couple of PPC cores thrown in, since
a dedicated PPC block takes a lot less space (and is faster) than one
implemented in the gate logic.
I've found that Xilinx and Agilent don't really get along all that well
with the concepts behind open source. They both have large libraries of
blocks available, but under fairly restrictive licenses. There are a good
number of open cores, but the quality isn't always that great.
Hardware design is hard.
Dave
--
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-lpsg
