>> Wednesday, February 23, 2011
One of the reasons I first became interested in OpenCL was AMD's Fusion APU (Accelerated Processor Unit). Since Intel's Larrabee project collapsed, this seems to be the super-processor of the future.
Originally, I heard that the Fusion would make discrete graphics cards obsolete because it contains a CPU and a GPU on a single chip. But AMD, realizing that this would cut into their lucrative graphics card business, has forsworn this. In the past, they used statements like "a supercomputer in every home," but now it looks like they don't have a coherent selling strategy at all.
So I'm nervous. The marketing for the Fusion has been meager, and although AMD has provided a whitepaper, they still haven't come up with a compelling reason for people to buy the Fusion. Instead, it looks like AMD is waiting for enthusiasts/entrepreneurs to tell them why people should buy the Fusion.
AMD hasn't provided any simulators or training classes, as IBM did for the Cell processor. They haven't announced any app contests with prizes, as Google did with their Android. But they have set up the AMD Fusion Fund, calling for companies to submit their ideas. I submitted my idea for Fusion development this afternoon, so we'll see what happens.