Undergraduate summer research in high performance computing with engineering applications: An experience report

Abstract

High Performance Computing (HPC) stands at the forefront of engineering innovation. With affordable and advanced HPC resources more readily accessible than ever before, computational simulation of complex physical phenomena becomes an increasingly attractive strategy to predict the physical behavior of diverse engineered systems. Furthermore, novel applications of HPC in engineering are highly interdisciplinary, requiring advanced skills in mathematical modeling, algorithm development as well as programming skills for parallel, distributed and concurrent architectures and environments. Reasons such as this have created a shortage of qualified workforce to conduct the much-needed research and development in these areas. This paper describes our experience with mentoring a cohort of ten high achieving undergraduate students in Summer 2019 to conduct engineering HPC research for ten weeks in Clarkson University. Our mentoring activity was informed and motivated by an initial informal study with the goal to learn the roles and status of HPC in engineering research and what can be improved to make more effective use of it. Through a combination of email surveys, in-person interviews, and an analysis of faculty research profiles in Clarkson University, we learn several characteristics of their research. First, a large proportion of the engineering faculty conducts research that is highly mathematical and computational and driven by disciplinary sciences, where simulation and HPC are widely needed as solutions. Second, due to the lack of resources to provide the necessary training in software development to their students, the interviewed engineering groups are limited in their ability to fully leverage HPC capability in their research. Therefore, novel pathways for training and educating engineering researchers in HPC software development must be explored in order to further advance the engineering research capability in HPC. With a three-year support from NSF, our summer research mentoring activities were able to accommodate per year ten high-achieving undergraduate students recruited from across the USA and their faculty mentors on the theme of HPC applications in engineering research. In this experience report, we describe our first offering of the site in summer 2019, including the processes of students recruitment and selection, training and engagement, research mentoring, and professional development for the students. Best practices and lessons learned are identified and summarized based on our own observations and the evaluation conducted by an independent evaluator. In particular, improvements are being planned so as to deliver a more holistic and rigorous research experience for future cohorts.