On Mon, 5 Dec 2005 15:17:54 -0500
Timothy Miller <[EMAIL PROTECTED]> wrote:
> > Interesting. I thought that the mask costs are so high, that
> > it wouldnt matter to have more functionality on the chip and a
> > bit more of peripherial stuff.
>
> Die area is a factor, but it's more than that. The more you pack in
> there, the more that can fail, and the lower your yield.
Good argument :)
> Plus, we're not going for full custom. The sorts of chips we'd be
> going after already have the silicon doped and some metal, and our
> circuitry is added by more layers of metal. Lower performance, lower
> cost, fewer masks. The dies come in fixed sizes, and if some
> low-priority feature pushes over from one size up to the next, we'll
> drop it in favor of the smaller die.
I think performance is not an issue here. The performance
of a sea of gates is still factor 1.5-3 better than the one
of an fpga. But yes, cost is an issue.
Attila Kinali
--
wer soviel schoggi isst, kann sowieso nicht dumm sein ;-)
-- Sandra
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
