In other ways, GigaPixel seems to emulate Imagination Technologies. GigaPixel uses a tile based architecture. GigaPixel breaks up a scene into tiles, checks each tile for visibility, and only renders the visible polygons in each frame. This should, in theory, save a great deal of bandwidth. Traditional 3D engines setup and shade a polygon before writing the visible ones to the frame buffer.

You can get more information on GigaPixel at their site. I just want to point out one big difference between 3D IP companies, and 3D chip companies, asides from the fact that one of them doesn't sell hardware: 3D IP companies tend to have more exotic solutions a la PowerVR and GigaPixel. There really isn't much room for anything out of the ordinary in the PC graphics arena. The design process is tied to the manufacturing process. You shrink dies, and add more transistors, and just chew up as much of the 3D pipeline as Intel and Microsoft allow.

3D IP companies have to be more concerned about addressing the needs of manufacturers that want to squeeze the 3D into either an existing design, or as part of some strategic component design. I don't honestly know if that is good or bad from a technology standpoint. Certainly, PowerVR has proven that the best intentions in the world don't add up to the best products. However, my contention is that in the desktop graphics market there is room for 3 big players, and everyone else is just not going to have the investment funds needed to compete in 3D. Therefore, 3D IP may be the best hope that's available to anyone who isn't ATI, Nvidia, or S3.

It's worth keeping an eye on companies such as GigaPixel. If GigaPixel can get a few big, committed customers, it will be able to stretch its architecture into the desktop arena, perhaps through a company like Matrox , or ELSA . I am pretty certain that there is still some life left in the graphics business, and that eventually, even the PC OEMs are going to want to buy in 3D technology that is exclusive to them. As for 3D in other devices, it would be a great thing for 3D IP companies, but unless a few Sega's spring up in the next five years, I don't see much to build a business on. Maybe Sony will scare up a few DVD and console competitors by virtue of its market superiority. Maybe that's Sony's the best thing that could happen to GigaPixel. We'll see.