MSA-Seminar: SOMA: Scaling Soft Polymers to many GPUs via OpenACC With Applications to Battery Electrolytes

Modular Supercomputing Architectures Seminar

Referent: Ludwig Schneider, Institut für Theoretische Physik, University of Göttingen

SOMA, our massively-parallel and GPU accelerated implementation of a soft, coarse-grained polymer model, enables us to study systems with billions of particles and predict their nano-scale self-assembly on micrometer scales. OpenACC and MPI allow us to utilize modern supercomputers without sacrificing readability and flexibility even for non-GPU expert scientists. The talk addresses the general multi-level parallel approach to the problem as well as some features for very large simulations necessary for specific scientific questions.

Dieses Seminar wird auch über DFNconf angeboten:

Name des Meetingraums: MSA-Seminar
Nummer des Meetingraums: 97983314

Um an dem Meeting teilzunehmen, haben Sie mehrere Möglichkeiten, s. Anleitung.