Lourens Veen wrote:
On Monday 19 February 2007 22:17, James Richard Tyrer wrote:
luc wrote:
Maybe I'm wrong, but I've read on this mailinglist that the more
you have to populate a board, the more that cost.
Yes, that is one of the factors.
Other factors go the other way, including:
Larger packages tend to require more board layers.
Custom chips cost more than off the shelf.
Larger chips have lower yields and therefore higher cost per unit
area.
A single chip might not have as long a life cycle.
Smaller ASIC chips might not need to be full custom, structured
might work.
and some of these factors are dependent on volume. E. G.
structured chips have lower up front costs but cost more per unit.
It is an optimization problem with no simple answer.
Hmm. Perhaps we need some kind of coprocessor then, if it's too hard
for a normal CPU. Perhaps some kind of PCI board with an FPGA on it?
You would definitely want custom GPUs.
The point is that if they are separate that it is easier to stay ahead
of the curve.
--
JRT
_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)
