In processors that do not support floating-point instructions, using fixed-point arithmetic instead of floating-point emulation trades off computation accuracy for execution spe...
Abstract. Conventional performance environments are based on pro ling and event instrumentation. It becomes problematic as parallel systems scale to hundreds of nodes and beyond. A...
Xian-He Sun, Mario Pantano, Thomas Fahringer, Zhao...
Abstract. The successful deploymentof next-generation distributedsystems is signi cantly dependent on the e cient management support that improves the performance and reliability o...
The backup of large data sets is preferably performed automatically outside of regular working hours. In highly structured computer networks, however, faults and exceptions may rel...
Supercomputer performance is highly dependent on its interconnection subsystem design. In this paper we study how di erent architectural approaches for router design impact into s...