On Wed, 23 Feb 2005 16:02:32 -0800, Jeff Carr <[EMAIL PROTECTED]> wrote:
> Timothy Miller wrote:
> 
> > Yes.  Now that the primary product will be in ASIC form, the choice of
> > FPGA is less important, although it's not entirely a non-issue,
> > because it will still affect the price of the prototypes.
> > Furthermore, die-area is still a factor in the cost of the ASIC.
> 
> OK.
> 
> I wonder if any of the FPGA parts are really capable of handling even 2D
> video anyway. Maybe an ASIC design is the only way to get that kind of
> performance? I'm interested in tinkering with a FPGA even if I thought
> it might be slower; it's technology I've heard about and never had a
> reason to use before.

The original FPGA design was not unrealistic.  It was just
unprofitable.  Yes, we may not have been able to run at 200Mhz, but I
have gotten a heck of a lot out of Spartan 3 chips before.  It was
challenging, but I've gotten internal logic to run at 400Mhz
before--that requires intimate familiarity with Xilinx architecture.

There is the open question about transistor budget, but we've already
been talking about ways to trim things without affecting accuracy too
much.  All this talk about switching to fixed-point after perspective
divide is very helpful.

The ability to put the design into an FPGA is something we are most
certainly going to productize.  We have identified markets where that
is necessary.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to