On 6/20/06, Dieter <[EMAIL PROTECTED]> wrote:
> and if we can cram a good 3D design into it,
There's that "3D" again.
But see a possible solution below.
This discussion was resolved. When people realized that our "3D"
design is a fixed function pipeline, minimized to support primarily
the stuff that desktop users need, they realized that it was what they
were asking for.
-------------
> that saves us on die area
> and therefore cost on the ASIC.
If you use the larger FPGA for OGD1, and then find a way to shrink
the design, is there a way to make a smaller ASIC?
Definitely. We'll be resynthesizing basically the same Verilog code
for an entirely different device.
-------------
> We want two independent video heads.
IIRC, 2 dual-link DVI plus s-video.
The way OGD1 is hooked up:
Head 1 is connected to 330MHz analog and dual-link DVI.
Head 2 is connected to TV/s-video and dual-link DVI.
There are only two video controllers, and lots of pins are shared.
Would 1 dual-link DVI plus
1 single-link DVI plus s-video be enough? Would it save significant
space? Single-link is supposed to be good for up to 1920x1200. How
many people will need to drive two displays both larger than 1920x1200?
I forget the exact pin arrangement, and I don't want to look it up.
There is probably too much signal dual-use for this to happen.
Besides, asking for three video controllers is just too much.
-------------
> An option we have is to produce an ECP2-50 board now and then switch
> to the ECP2-70 in 2007.
If you think that the ECP2-50 is too small, why make a ECP2-50 board?
We didn't think it was too small until just now. But our issue is
that we don't want to delay the project any longer than necessary.
How long do YOU want to wait for this stuff to happen?
Even if the -50 is too small, we can still fit MOST of the design in
it and do tons of incremental work.
-------------
I assume you have already "outsourced" any pieces available in an
existing sufficiently documented IC.
Yeah, if I understand what you mean.
-------------
There is that FPGA that plugs into an AMD64 socket. Expensive, and
requires a mainboard with at least two CPU sockets, therefore not practical
for a mass-production low-cost video solution, but perhaps useful for
development?
OGD1 has to be practical both for development and for mass sale to
people who need prototyping boards.
Perhaps this could be a way to split out some optional 3D stuff, as a couple
of us have suggested. Build a basic ECP2-50 board, and put the fancy 3D stuff
in the AMD64 socket. The AMD64 socket has direct access to the hypertransport
bus, thus very very high bandwidth available. This could turn out to be a
better solution for high-end 3D than the SLI type kludge.
Interesting idea, but this would require so much rework on the board
that we might as well wait until 2007.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
