The funds from the Division of Materials Research support the University of Chicago to acquire a modern, Graphics Processor Unit (GPU) based computer that will allow researchers from a wide variety of disciplines to pursue new materials research directions that are based on the use of machine learning strategies and that could not be investigated without access to the necessary computational resources. Four focus areas are envisaged, namely (1) new materials discovery, (2) development of new molecular models, (3) machine-learning enabled interpretation of experiments, and (4) development of new simulation algorithms that rely on concepts from machine learning. A series of workshops and other outreach activities will be carried out to engage the broader materials research community in advanced computational research on GPU-accelerated computers.
The Division of Materials Research supports the University of Chicago with the project "Acquisition of a high-performance GPU-based computer for advanced multiscale materials modeling" . The instrument is a high-performance GPU (Graphics Processing Unit) cluster tailored for fast and efficient simulations, including molecular dynamics (MD) simulations, hybrid particle-continuum simulations, mesoscale simulations and continuum simulations. This cluster will thrust forward a wide array of multidisciplinary research projects anchored in molecular engineering, physics, chemistry, and biology, and builds upon the investigators' involvement in developing the fastest, most powerful suite of simulation software for particle-based simulations on GPUs, including molecular dynamics simulations, evolutionary optimization algorithms, and powerful deep learning-assisted advanced sampling algorithms. That software is now distributed freely, and it is used throughout the world. The acquisition of a large GPU cluster at The University of Chicago will provide a unique computational resource and drive new collaborative efforts in algorithm and software development at the interface between molecular engineering, physics, chemistry, biology, computer science and materials science. The system will operate as a facility to serve four interrelated objectives, namely (i) support the local research community and their collaborators to achieve significant scientific advances on an array of collaborative research projects touching on proteins, membranes, cytoskeleton, fluids, colloids, nanoparticles, polymers, and mechanical metamaterials; (ii) create a vibrant interdisciplinary intellectual research community of theoreticians and experimentalists around the facility, and promote the creative development of novel and more effective advanced sampling algorithms, centered on the creation of a unified set of software tools supporting state-of-the-art methodologies; (iii) organize workshops to train undergraduate, graduate, and postgraduate students from a wide array of disciplines to disseminate knowledge about the utilization of distributed computational environments; (iv) contribute to diversifying the scientific workforce by organizing training and educational programs for undergraduate students from institutions serving under-represented minorities.
This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.
