That is the reason why my point rendering with illumination calculated in the GPU is faster than small polygons.
I calculate everything from scratch. Dec 4, 2013 17.
That's where I think GPU designers have missed the point. The Cell allows you to do that by hand, thus, solving this problem. A complete real-time Ray-tracing program is out of question unless you compromise many things.
The scenes are much more complex there than in polygon scenes rendered in GPUs. It's four generations out of date. The 's GPU essentially eliminates memory bottlenecks and allows for things like 4xAA with almost no penalty in performance because of this.
A Cell with 16 SPEs is not inconceivable in the near future. I am not saying one should completely abandon the GPU.
Unfortunately, that gave a technical edge to Sony while introducing a bottleneck for the console. Hmm, this discussion continues to be interesting. It's awesome. They should try and enable P for those of us with high end gaming rigs though.
However that is not possible!
Nowadays, we do it using Gouraud shading that interpolates colors, thus, avoiding the calculation. Dec 2, 2013 9. The answer is very simple: Let's say that they will have 32 pipelines soon.
You see the beauty? Since the days of GeForce 256 you have been able to do per pixel illumination even if only with simple illumination models.
The short answer: If you believe what I am saying without this test, then just answer me what do you think it is happening there???
If you don't believe just do a simple program that tessellates parametric surfaces take a simple torus, for example on the fly where you can increase the number of polygons by further subdividing the surface.
It is slower!