ABSTRACT
Since the advent of computer in the 1950s, weather forecasting has been a hugely challenging computational problem. Right since the beginning, weather models ran on a single supercomputer that could ll a gymnasium and contained a couple of fast (for the 1970s) CPUs with very expensive memory. Software in the 1970s was primitive, so most of the performance at that time was in clever hardware engineering. By the 1990s, software had improved to the point where a large program running on monolithic supercomputers could be broken into a hundred smaller programs working simultaneously on a hundred workstations. When all the programs nished running, their results were stitched together to form a weeklong weather simulation. What used to take fteen days to compute and simulate seven days of weather even in the 90’s, today the parallel simulations corresponding to a weeklong forecast can be accomplished in a matter of hours.
