I just noticed my thunderbird client BOLDING my filter for this group.
Which was shocking. As, this was one of my favorite lists to lurk on,
and being now that it is in the summer and I find myself working on
entirely different areas (Computational Genomics) I was frightened it
would startup again and I wouldn't get any work done.
:-)
But having been a lurker for so long I didn't expect this project to go
anywhere for the following reasons:
1) The open source community wanted 3D. I didn't see this project
producing that. After discussing it with the MESA people, they became
largely uninterested as a result. You are producing a product that very
few people have an interest in in my opinion.
However, you did catch the interest of AMD and Nvidia, and they didn't
like what you were doing. :-) So I think you had a minor influence on
both of them, particularly AMD.
2) I didn't see number one happening because in order to produce such
parts, you are talking about fairly large entry level hurdles to
electronics problems producing high frequency parts. Especially board
integration and testing. I mean if I wanted to make a decent 3D card
that was open and worked with MESA, I would need, Oh, about 150K in
benchtop toolsets to do design verification. That would be minimal I
figure to produce a design that would work on any desktop/laptop you
would buy.
That would be for PCI Xpress support and different form factors. I
would want to be able to sell the cards in a MXM 3.0 form factor as well.
So I think you should drop the idea of producing a video _card_,
Tim/community. Build a GPU that is MESA community friendly.
3) I think you tried to chew too much at once with this project,
personally Tim. If I was the projects director, I would start by
producing a single chip, with a base set of specifications. That is
all. There are lots of digital simulators and logic design validation
tools in the open source world to create the logic for the chip which
would simplify 2D and 3D computation tasks.
In particular, you would work very closely with the MESA community to
create the GPU design.
This would keep the project in compute space, as I would describe it,
enabling it to accept more contributors. Once the community has a chip
for 3D/2D that is comparable to fulfilling the MESA communities desire
for decent performance I think you would have a great deal of interest
in the project.
Eventually I think the biggest challenge would be fabrication and board
testing. That takes big bucks for high freq parts. I think we would
have to hold a Open Graphics Telethon. :-) But I think it could be
done because it you would have the best interests of not just yourself
and others building the GPU, but a much larger fraction of people
working on X/Windows 3D/OpenGL support for LINUX.
But I think at this _time_ it is a mistake to try and produce a card.
-gc
PS: Post Thoughts:
One thing I believe is AMD and Nvidia will not like any kind of open
graphics hardware and will do everything above board and under the table
to prevent it from happening.
Including pulling a SCO move. (i.e. Giving a third party millions like
Microsoft did with SCO to sue on their behalf.).
Oh, and one last thing. Has nothing to do with the list topic but....
Daryl McBride, you suck it. :-)
-gc
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
