A performance estimation model for Ls-Dyna computation workload

摘要

Given the rapid change in processor architecture in the past years, there is a driving necessity to assess processor performance for a high performance computation workload. Assessing performance for a given workload is important to understand the correlation between architecture parameters and the workload performance. A given workload can be categorized as memory bound, compute bound, or in between. In this paper we present performance characterization for a high performance computation workload using LS-DYNA/car2car. We also propose a performance estimation model in which we can estimate performance for LS-DYNA workload by changing specific processor architecture parameters. The model is verified to estimate performance for different processor architecture parameters with error margin <;5%.