Performance predictions on implicit communication systems

This paper presents a multiprocessor performance prediction methodology supported by experimental measurements, which predicts the execution time of large application programs on large parallel architectures based on a small set of sample data. We propose a graph model to describe application program behavior. Important and implicit architecture parameters are obtained by experiments. We focus on performance predictions of application programs on multiprocessors with implicit communications. A large scientific simulation program is implemented using the shared-memory model on the KSR-1 and using the data-parallel model on the CM-5 for performance measurements and prediction validation. We show that experimental measurements provide strong support for the performance prediction on multiprocessors with implicit communications and complex memory systems