3 ms·Fooling the masses with performance results on parallel computers (2010)2 points by jsnell 10y ago