Materials science research like the work performed at the Institute for Molecular Engineering (IME) involves running large numbers of simulation and modeling programs, often in complex patterns, over long periods of time, by many collaborators. This work may be done to simulate natural phenomena, reduce the design space for subsequent physical experiments, or interpret those experiments in terms of detailed molecular models. These complex simulation and data analysis processes require large parallel computing systems and distributed data storage servers. The difficulty and complexity of these processes demand a high degree of computing skills, and distract scientists from their core domain focus.
For December's Inside the Discovery Cloud event, Juan de Pablo, Liew Family Professor in Molecular Engineering at the IME, discussed how his laboratory uses computational approaches to study DNA mapping and self-assembly. Then, Michael Wilde, CI Senior Fellow, and software architect at Argonne National Laboratory discussed methods developed and applied at UChicago and Argonne that help researchers conquer the complexity of high performance computer modeling and better integrate it into the scientific knowledge discovery process.
You can watch videos of both talks below.