Getting bigger slower

Print Friendly, PDF & Email

Ian Foster points to an interesting article in IEEE Distributed Systems Online in which Matei Ripeanu notes that the top end of the list is growing slower than the bottom end:

Thus, for example, the bottom 25 machines in the Top500, if aggregated together, would match only the #30 machine in 1993, but match the #5 machine in 2005.

Ripeanu attributes this to increasing ease of aggregation growing the low end. I would add that at the high end acquisitions are often (approximately) fixed dollar exercises and have always been tightly aggregated so that the only growth up there is from Moore’s law.