ABSTRACT

Karl Fu¨rlinger University of Tennessee Innovative Computing Laboratory, 1122 Volunteer Blvd, Knoxville, Tennessee 37996-3450 karl@cs.utk.edu

Performance analysis of applications is and certainly will be of even higher importance on future petascale systems. The reason is that these architectures will exploit multiple levels of parallelism and deep memory hierarchies to reach several petaflops (PFLOPS) of peak performance without having automatic programming tools available.