ExaWind
Abstract not provided.
Abstract not provided.
Abstract not provided.
Abstract not provided.
Abstract not provided.
Abstract not provided.
A method for providing non-diffuse transport of material quantities in arbitrary Lagrangian- Eulerian (ALE) dynamic solid mechanics computations is presented. ALE computations are highly desirable for simulating dynamic problems that incorporate multiple materials and large deformations. Despite the advantages of using ALE for such problems, the method is associ- ated with diffusion of material quantities due to the advection transport step of the computa- tional cycle. This drawback poses great difficulty for applications of material failure for which discrete features are important, but are smeared out as a result of the diffusive advection op- eration. The focus of this work is an ALE method that incorporates transport of variables on discrete, massless points that move with the velocity field, referred to as Lagrangian material tracers (LMT), and consequently prevents diffusion of certain material quantities of interest. A detailed description of the algorithm is provided along with discussion of its computational aspects. Simulation results include a simple proof of concept, verification using a manufac- tured solution, and fragmentation of a uniformly loaded thin ring that clearly demonstrates the improvement offered by the ALE LMT method.
Weak scaling studies were performed for the explicit solid dynamics component of the ALEGRA code on two Cray supercomputer platforms during the period 2012-2015, involving a production-oriented hypervelocity impact problem. Results from these studies are presented, with analysis of the performance, scaling, and throughput of the code on these machines. The analysis demonstrates logarithmic scaling of the average CPU time per cycle up to core counts on the order of 10,000. At higher core counts, variable performance is observed, with significant upward excursions in compute time from the logarithmic trend. However, for core counts less than 10,000, the results show a 3 × improvement in simulation throughput, and a 2 × improvement in logarithmic scaling. This improvement is linked to improved memory performance on the Cray platforms, and to significant improvements made over this period to the data layout used by ALEGRA.