Combining Aspen with Massively Parallel Simulation for Effective Exascale Co-Design

Our position is that new performance prediction methods are required to adequately find and mitigate potential performance bottlenecks in exascale system designs. To address this gap, we propose to leverage two complementary approaches – a structured performance modeling methodology called Aspen and massively parallel simulation – to enable the understanding of future exascale designs with unprecedented accuracy and greater ease than is available today. The overall benefit of this activity is that it will lead to a much higher return on the amount of exascale science produced for the machine investment made by the DOE.