Patrick Elliott <sel### [at] rraznet> wrote:

> No it couldn't, because graphics cards ***do not*** raytrace, then
> scanline.

I think what's actually being referred to is using the GPU (in sufficiently
modern cards) as a fast vector processor that can be programmed to carry
out arbitrary (i.e. not triangular or scanliney) calculations. I would
actually read the headlines on that site (http://www.gpgpu.com) to see what
is being talked about. GPU pipelines are still single-precision though, and
POVRay uses double-precision maths throughout. It would be a massive job in
any case, and easier to buy more CPUs.

> That is the other thing you don't get. At some point you still have to
> do the math. You can do it the way POV-Ray does and produce "exact"
> mathematical versions, or you can use the math the generate
> "approximations" that the GPU can handle. And it will always be an
> approximation. You can't do true physical models of real world objects,
> which are "not" made up of bunches of triangles

I don't agree. Raytracing is absolutely an approximation (e.g. no forward
light without photon maps, ignores wave physics of light, etc). Also, the
use of triangles is a separate issue (you can clearly raytrace triangles).
I would say that real-world objects are not made up of any sort of
primitive that POVRay or other renderers use, triangles included.
Certainly, I've never seen a perfect box{} in the real world (checks
furniture).

Tom

Post a reply to this message