In this video from SC17, Jack Dongarra, Erich Strohmaier, and Horst Simon give an Invited Talk on the history of the TOP500 list.
“The TOP500 list has served as the defining yardstick for supercomputing performance since 1993. Published twice a year, it compiles the world’s 500 largest installations and some of their main characteristics. Systems are ranked according to their performance of the Linpack benchmark, which solves a dense system of linear equations. Over time, the data collected for the list has enabled the early identification and quantification of many important technological and architectural trends related to high-performance computing. We briefly describe the project’s origins, the principles guiding data collection, and what has made the list so successful during the 25 year long transition from giga- to tera- to petascale computing. We also examine the list’s limitations. The TOP500’s simplicity has invited many criticisms, and we consider several complementary or competing projects that have tried to address these concerns. Finally, we explore several emerging trends and reflect on the list’s potential usefulness for guiding large-scale HPC into the exascale era.”
