You should be interested in it. It is a great idea, since GPUs are good at doing one thing, and one thing alone, calculating FPUs. They are a direct mathamatical pipe. The GPU version of folding works by filling up the vector matrixs and running the calculations on them. Very very very effecient way of churning out work units, there are a few down sides to the GPU client vs CPU, it still doesn't totally off loan the CPU, since it is a GPU it lacks alot of command sets that make CPUs more overal versital (and slower in comparison to math), such as the ability to OR/XOR data, so imagine trying to write a program that does something, but you can't have any conditions or variables outside of your math matrix's? So yous till need to use a decent (few % depending on CPU) to run the FAH client.