Accelerating the FlowSimulator: Performance Analysis & Optimization of DLR High-Performance Solvers using Score-P

Abstract

With rapidly growing interest in high-fidelity simulations in the field of aircraft design, high performance computing (HPC) becomes indispensable. Besides, with the end of Moore’s law, processors evolve towards an ever-increasing core count which raises requirements on the scalability of scientific software. Consequently, the coordinated interaction of algorithms, hardware and software is required to efficiently use the available resources. Therefore, this work deals with an in-depth scalability analysis of the DLR flow solvers CODA [Huismann 2021 HyperCODA] and Musubi [Hasert 2014 Musubi].The scalability analysis is refined by tracing the applications using Score-P [Knuepfer 2012 Score-P], leading to the identification of critical aspects of both codes. The presented analysis reveals potentially sub-optimal communication patterns, proposes possible solutions, and shows first improvements of the codes.