Speedup / on how much cores should I start my job?

The performance of LS-DYNA is optimized for actual parallel computer architectures. With the High Performance Computing platform available on LS-DYNAcloud satisfying speedups can be achieved up to parallel usage of 1000s of cores. The speedup varies depending on the type of application and the geometry of the model. In many cases (e.g. small models) speedup on high core numbers cannot be guaranteed. Therefore, we recommend to do a performance study to determine the scaling factor for your individual case in order to optimize valuable computing time.

Leave a Reply

You must be logged in to post a comment.