Abstract
Even today, when microprocessor vendors announce breakthroughs every other week, performance evaluation is still one of the key issues in parallel computing. One of the observations is that on a single PE, many, if not most applications relevant to the technical field do not benefit adequately from clock rate im- provement. The reason for this is memory access: most data read and write operations access memory which is relatively slow compared to processor speed. With several levels of caches we now have complex system architectures which, in principal, provide plenty of options to keep the ata as close as possible to the processor. Nevertheless, compiler evelopment proceeds in slow progression, an the single PE still dominates the results achieve on large parallel machines.
Chapter PDF
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2000 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Fahringer, T., Nagel, W.E. (2000). Performance Evaluation and Prediction. In: Bode, A., Ludwig, T., Karl, W., Wismüller, R. (eds) Euro-Par 2000 Parallel Processing. Euro-Par 2000. Lecture Notes in Computer Science, vol 1900. Springer, Berlin, Heidelberg. https://doi.org/10.1007/3-540-44520-X_14
Download citation
DOI: https://doi.org/10.1007/3-540-44520-X_14
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-67956-1
Online ISBN: 978-3-540-44520-3
eBook Packages: Springer Book Archive