Euclidian

[Kahn62]

What do you think off this Eugene. It has NO polygons.All of the graphics can be done without needing a big video card. It uses points, called "atoms" instead of polygons. Much less work for video card.

[Eugene Cojocar]

There must be a catch in it. All modern graphics card can render polygons only. So, if the atom consists of polygons then a graphics card must render billions of polygons per second. It's not possible. Maybe this video is a promotion video to find an investor or capital for this startup or something similar.

Btw, I heard about something similar revolution in computer graphics some years ago, but can't remember their name and technology.
They also promised revolution in CG. Instead of revolutions we can see a typical evolution.

[Kahn62]

It uses no polygons. It also does not require a modern video card. The "atoms" are eight pixel groups. Each object is made of hundreds of grains stuck together to form the objects. This was demoed three years ago but nothing was heard about until now. They have put this new demo together. Their world is made up of billions of these atoms joined together. Check this out, it explains it better than I can. http://en.wikipedia.org/wiki/Euclideon The way it would be used, would be for static objects and scenery to be done with this method and mobile or movable stuff to be done with polygons and laid over the background.

[Eugene Cojocar]

I looked at their another videos and tried to find some more information about it...frankly I don't know their secret.
At first they told that they used voxels http://en.wikipedia.org/wiki/Voxel then declined it.
Maybe it's a modified voxel rendering maybe something another. I don't know for sure.
It's possible to make a final decision when this technology is available for testing.
Otherwise it looks like E-Cat http://en.wikipedia.org/wiki/Energy_Catalyzer in CG.

[Kahn62]

Lol. I know what E-Cat is. How it works is by creating an index of a world that can be searched quickly and simply. then it determines what each pixel on the screen will be focused on and that gives it a dot of colour. therefore if you have a resolution of 1920 x 1080 then you only need look up two million points per frame. Something like that anyway. I do not expect you to incorperate this into XMR this year, I just wanted you to see what the future may look like.

[Kahn62]

Could you imagine not having to use polygons for scenery? It would stop a lot of the issues you have now with "busy" parts of the track slowing frame rates down. Like near the pits on Donnington Park for instance.

[Eugene Cojocar]

Could you imagine not having to use polygons for scenery? It would stop a lot of the issues you have now with "busy" parts of the track slowing frame rates down. Like near the pits on Donnington Park for instance.

I think that more advanced LOD system could help to solve the issue.
A polygon (triangle) is a base element of CG and the most fastest method of rendering. I have no ideas how to render a picture without polygons. Ray tracing is very slow. Voxels are not ideal also. We have only polygons.

[Kahn62]

The idea is that the CPU can do all of the immovable objects, leaving the GPU free to handle the moving objects as an overlay over the background. I don't know how many frames a CPU core could render per second using this process, but it is a cool thing to think about. I don't know where the line could be drawn between what is moving and what isn't, like is the inside of a car a moving object? They make some pretty big claims, but then so do most frauds. And yes, I've been following the E-Cat for a few years now, and it never seems to do anything, and this could be like that too, but just because we don't understand something is no reason to assume it can't work.

[Eugene Cojocar]

The bottleneck is lower performance of system RAM. VRAM is faster than RAM. Actually RAM is very slow without cache memory.
Even if a scene is static you have to read data from memory and render it each time if you move in it.
Theory is theory, practice is practice.