The alternative suggestion is that we use a powerful general-purpose
CPU or DSP as the primary processing element of the graphics chip
design.
Ah, ok. Now I understand. I third idea that was floating around was
an array
of floating point processors, not unlike the G80, that combined into
a card.
I suggest joining the gEDA-user mailing list. Before doing that, do
as much reading as you can about gEDA.
gEDA appears to be much more mature than the last time I looked at it,
when I decided that Eagle would suit me better.
I just looked at the gEDA website, and look what was linked to on the
main page:
http://dlharmon.com/dspcard/index.html
our work is done. time to write the software :-)
Is that what you really want? A video decoder? Not a graphics card?
I would like a graphics card, but isn't that what OGA is for? ;-)
No, I think it is an interesting idea to make a graphics card based
on a DSP,
but I am not convinced that it is possible to get anywhere close to
reasonable
performance from it (I would say that ATI Rage era is "reasonable"
sans games).
As a proof of concept and a learning tool, it is great. The same
learning can be
gained from a DSP based board that decodes video, but that has the
distinct
advantage of being useful.
If there is evidence that a single DSP is able to produce reasonably
accelerated
3D graphics, I am willing to do that instead.
If you want a graphics card, I think what you'll need is some
combination of a processing element (some processor) and a small FPGA
to send the video data. Most likely, you'll want a DSP with a bus on
it for memory, not a built-in memory controller, because we'd most
likely want to do the memory through the FPGA so that we have enough
bandwidth for video.
ok, that makes sense.
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
