1. What are the contributions mentioned in the paper "Predicting application performance using supervised learning on communication features" ?
These metrics make simplified assumptions about the cause of network congestion, and do not provide accurate correlation with execution time.. In this paper, the authors attempt to model the performance of an application using communication data, such as the communication graph and network hardware counters.. The authors use supervised learning algorithms, such as randomized decision trees, to correlate performance with prior and new metrics.. The authors propose new hybrid metrics that provide high correlation with application performance, and may be useful for accurate performance prediction.. For three different communication patterns and a production application, the authors demonstrate a very strong correlation between the proposed metrics and the execution time of these codes.
read more
2. What are the future works in "Predicting application performance using supervised learning on communication features" ?
This may provide a scalable method for performance prediction at large scales and for future machines without having to perform detailed network simulations.. Finally, the authors have demonstrated that supervised learning and ensemble methods can be used to predict performance not only for simple communication kernels but also for complex production applications with several diverse communication phases.
read more
