Al writes "While the speed at which supercomputers can process data is growing rapidly, the same can't be said of the speed at which they can input and output that data. The result, is that some research teams are doing away with graphics-processing clusters altogether, and using the regular hardware for visualization instead. Researchers from Argonne National Laboratory (ANL) and the University of California at Davis created software for Intrepid, an IBM Blue Gene/P supercomputer at ANL, that allows it to bypasses the graphics-processing cluster entirely, while producing spectacular supernova visualizations. "It allows us to [visualize experiments] in a place that's closer to where data reside--on the same machine," says lead researcher Tom Peterka. His team's solution obviates the need to take the time-consuming step of moving the data from where it was generated to a secondary computer cluster. This may be part of a broader trend, that is blurring the line between the CUP and the GPU. Only last week, AMD recently released software that allows GPU code to run on X86 CPUs."
Link to Original Source