Sciweavers

SC
2005
ACM

How Well Can Simple Metrics Represent the Performance of HPC Applications?

14 years 6 months ago
How Well Can Simple Metrics Represent the Performance of HPC Applications?
In this paper, a systematic study of the effects of complexity of prediction methodology on its accuracy for a set of real applications on a variety of HPC systems is performed. Results indicate that the use of any single, simple synthetic metric to predict performance does an inadequate job, and the use of a linear combination of these simple metrics with optimized weights also performs poorly. Better, however, are methodologies that rely on the convolution of an application “transfer function” based on tracing information with system performance data measured by simple benchmarks. This latter methodology can predict performance with an average accuracy of 80%, based on the current work.
Laura Carrington, Michael Laurenzano, Allan Snavel
Added 26 Jun 2010
Updated 26 Jun 2010
Type Conference
Year 2005
Where SC
Authors Laura Carrington, Michael Laurenzano, Allan Snavely, Roy L. Campbell, Larry P. Davis
Comments (0)