An Approach to Performance Prediction for Parallel Applications

• Published in: Euro-Par 2005 Parallel Processing pp. 196 - 205
• Main Authors: Ipek, Engin, de Supinski, Bronis R., Schulz, Martin, McKee, Sally A.
• Format: Book Chapter Conference Proceeding
• Language: English
• Published: Berlin, Heidelberg Springer Berlin Heidelberg 01.01.2005; Springer
• Series: Lecture Notes in Computer Science
• Subjects: Applied sciences; Computer science; control theory; systems; Computer systems and distributed systems. User interface; Exact sciences and technology; Feedforward Neural Network; Hide Unit; Parallel Application; Percentage Error; Performance Prediction; Software; Performance evaluation; High performance; Parallel algorithm; Software architecture; Multidimensional space; Neural network; Distributed system; Distributed computing; Modeling; Time variation; Multilayer network; Execution time
• ISBN: 3540287000; 9783540287001
• ISSN: 0302-9743; 1611-3349; 1611-3349
• DOI: 10.1007/11549468_24

Accurately modeling and predicting performance for large-scale applications becomes increasingly difficult as system complexity scales dramatically. Analytic predictive models are useful, but are difficult to construct, usually limited in scope, and often fail to capture subtle interactions between architecture and software. In contrast, we employ multilayer neural networks trained on input data from executions on the target platform. This approach is useful for predicting many aspects of performance, and it captures full system complexity. Our models are developed automatically from the training input set, avoiding the difficult and potentially error-prone process required to develop analytic models. This study focuses on the high-performance, parallel application SMG2000, a much studied code whose variations in execution times are still not well understood. Our model predicts performance on two large-scale parallel platforms within 5%-7% error across a large, multi-dimensional parameter space.