I started the idea, and someone else followed up.

When we think of performance of OGC, we think it's not going to be
that great, because something like bitblt will be slower than what you
can get from, say, a Radeon.

But we already know that where we're going to shine is in the drivers.
 Take that a step further and consider just how much tweaking you can
really do.  Say you accelerate the top 20% of X11 calls for a Radeon,
but you find a way to accelerate 99% of them for OGC.  They'll beat us
at "x11perf -copywinwin500", and probably beat us even more on
"x11perf -rect500", but what about the full xmarks suite?  More
obscure functions don't count as much towards the benchmark, but as I
have exploited in the past, they do nevertheless add up.

With really good drivers, there are going to be a few things that they
perform better than us on, but with the number of things that they do
not accelerate, we may well be able to provide a better, more
responsive user experience.  This would be the result of a combination
of more accelerated functions and better DMA and who knows what other
clever things we can do in the driver.

Ok, so we're not going to beat then at OpenGL games.  But what about
alpha-composited windows on the desktop?  How about text-heavy apps
(aliased and antialiased)?

We need to get this thing DONE.

_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to