high performance computing on graphics processing units: hgpu.org

The increasing computing power of GPUs can be used to improve the performance of CAE systems.[1]. Within LS-DYNA an improved direct equation solver can be used, which accelerates the performance of implicit applications by use of a CUDA-based solver [2], [3], [4]. In this paper the performance improvements for different customer input decks for metal forming application (implicit springback calculations) are being compared on a Cluster system with actual GPU hardware nodes. First an overview of the underlying hardware architecture of actual GPU systems is given. Different possibilities to use GPU for CAE computations are being introduced. The current implementation within LS-DYNA is explained and possible applications are decribed, which may benefit from the CUDA-version. For some customer applications the performance data of the LS-DYNA CUDA-version versus the LS-DYNA SMP-version is are compared.