computer scientists at the U.S. Department of Energy’s Lawrence Berkeley National Laboratory evaluated the processor’s performance in running several scientific application kernels, then compared this performance against other processor architectures. The paper, “The Potential of the Cell Processor for Scientific Computing,” was written by Samuel Williams, Leonid Oliker, Parry Husbands, Shoaib Kamil and Katherine Yelick, of Berkeley Lab’s Future Technologies Group and by John Shalf from NERSC. On average, Cell is eight times faster and at least eight times more power efficient than current Opteron and Itanium processors, despite the fact that Cell’s peak double precision performance is fourteen times slower than its peak single precision performance. If Cell were to include at least one fully utilizable pipelined double precision floating point unit, as proposed in their Cell+ implementation, these speedups would easily double. Details on the cell processor are at wikipedia. It includes descriptions of moving from the current 90nm process to 65nm and below. A 65nm cell chip would be two or more times as fast as a 90nm. A 45 nm cell chip would be two or more times faster than a 65nm version.
If the equivalent of a Blue Gene/L computer were made with Cell+ chips would have over 7 petaflop performance. A Cell+ 45nm version of a Blue Gene/L would be 28 petaflops. Cell processors will be made in volume for the Playstation 3, so the costs of the chips should be very competitive.