On Mon, 2006-04-03 at 21:16 -0500, Patrick McNamara wrote: > Vinicius Santos wrote: > > On 4/3/06, Timothy Miller <[EMAIL PROTECTED]> wrote: > > > >> [snip] > >> That all being said, some of all of the Linux community will shift > >> over time to wanting to have programmable shaders and hardware vertex > >> processing. With products like OGC1 funding us, we're going to want a > >> piece of the game market. In the shorter term, I can't help but > >> wonder if we couldn't produce a simpler design (probably which runs at > >> a much higher clock rate) that is fully programmable. I personally > >> will have to try not to get distracted by it, but there's no reason > >> why the community couldn't attempt to best the OGA1 design by spec'ing > >> out a programmable architecture. > >> > >> Until the next major OGD-related event, what do you say to spending > >> some time discussing some different ideas? > >> > > > > The first problem seems to be prototyping a full programmable gpu in > > fpga. It's pratically a superescalar generic processor with my past > > researching on fpga mailing lists showed to be not feasible, but I am > > not an expert. > > But I like the idea, specially because programmable vertex/pixel > > shaders aren't important only to games, but CAD(in a generic sense) > > as well. And let's not forget uni-verse ^_^ > > > > > So we don't simulate the entire thing in a single FPGA. > > I also think that we should not be afraid to "do it differently". > Perhaps we should design a graphics coprocessor board to hang off a > HyperTransport expansion connector. Break down the major functional > areas into seperate smaller chips connected by high speed busses through > a decent crosspoint switch and hang the whole thing behind a HTX > interface. Each functional area could be simulated in an FPGA (maybe > even smaller cheaper ones). In some cases, perhaps certain chunks could > be handled by existing off the shelf processors. Could you simulate 16 > programmable shaders by themselves in a FPGA? Maybe use a PPC with some > very tuned altivec code for geometry setup (perhaps the complete OGL > stack?). >
Hmm... A NUMA graphics card? ISTR ages back a company made a GFX card that was packed full of many really single processing engines... ABout as complex as a 6502 IIRC... But clocked quite fast, and with an instruction set optimised for graphics. They were building them into full sized games machines (Arcade type) that you climbed into to play. Damn, wish I could remeber what they were called. It would have been mid 1990's. > Trying to simulate a fully programmable, fully pipelined GPU as a whole > unit is something that takes million dollar pieces of hardware to do. > We can't do that. As someone famous one said, sometime necessity is the > mother of invention. > > Patrick M > Patrick M > > > _______________________________________________ > Open-graphics mailing list > [email protected] > http://lists.duskglow.com/mailman/listinfo/open-graphics > List service provided by Duskglow Consulting, LLC (www.duskglow.com) _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
