On the Interpretation of Top500 Data
Dror G. Feitelson · The International Journal of High Performance Computing Applications · 1999
A novel representation of Top500 data is proposed in which the systems are plotted in rank and number-of-processors coordinates. This rendering brings to light various trends, such as complete Japanese domination of the range of high rank achieved with a small number of processors. In addition, an evolutionary trend whereby the rank of a given machine doubles every year is identified. This, in turn, leads to a characterization of the distribution of computing power across the whole list.