Carmack on Parhelia

"An Optimal Algorithm for Expanding the Composition of Polynomials" sounds more like it's related to the calculation of the actual curve in HOS, not tied to any specific surface. That is, just from the title, it really looks like an algorithm for generating a curve from control points.
 
MfA said:
Pretty sure, OpenGL guy :) If you want all the technical info dig up their siggraph paper on the tesselator, this quote from the OpenGL specs says it all though :

Should we support triangular patches?
RESOLVED: Yes. Otherwise, applications will have to convert
them to rectangular patches themselves. We can do this more
efficiently.

NVIDIA uses a simple degenerate quad for a tri
I'll take your word for it :)
 
But it should be obvious if a degenerate quad was used, as the tessellation would sort of converge at one point...I haven't seen any evidence of this yet.

Update: Never mind...shouldn't be an issue if tessellation was controlled on a per-side basis.
 
My bad again, they use the 3-quad approach for fractional tesselation, a degenerate quad for integer tesselation.
 
I am a bit young on the 3d scene, so could someone please explain the difference between how a quad is drawn and a triangle? I wouldn't think it would be too different as the vertices seem to most important to me anyhow.
 
i dunno

i'm not the smartest here but isn't this like comparing
a duron to a top of the line p4 ? of course the p4 will win its much faster but that doesn't mean the duron isn't a better chip. Can someone mabye down clock the geforce 4 or mabye over clock the Matrox part and see which one is more effective. At some point in time nvidia wont be able to make a higher clocked geforce anymore. But anyway if they took all the features of this card like the 3 monitor plugs and what not and got it out the door at 200 bucks i think it sell well against the radeons and the geforces out there. but what does it matter a new power vr card will come out and ownz you all :)
 
Back
Top