Section: New Results

Process Affinity, Metrics and Impact on Performance

Process placement, also called topology mapping, is a well-known strategy to improve parallel program execution by reducing the communication cost between processes. It requires two inputs: the topology of the target machine and a measure of the affinity between processes. In the literature, the dominant affinity measure is the communication matrix that describes the amount of communication between processes. The goal of this work is to study the accuracy of the communication matrix as a measure of affinity. We have done an extensive set of tests with two fat-tree machines and a 3d-torus machine to evaluate several hypotheses that are often made in the literature and to discuss their validity. First, we have checked the correlation between algorithmic metrics and the performance of the application. Then, we have checked whether a good generic process placement algorithm never degrades performance. And finally, we have seen whether the structure of the communication matrix can be used to predict gain [16].