Group leader for Statistical Sciences at Los Alamos National Laboratory
Exascale simulations, like those of interest to the Department of Energy, will produce more data than can be stored for post hoc statistical analysis. To address this, we are using Julia as an HPC-embedded language to develop workflows for statistical analysis inside simulations as they run. Our project, PRISM, is being developed to fit Bayesian hierarchical models to answer scientific questions using the high resolution data that will be produced on the next generation of supercomputers.
In this talk, we will present our work of coupling the Julia runtime with E3SM, an advanced earth system simulation application for supercomputers, and running E3SM with swappable in-situ Julia modules at large scale. The talk includes (1) the Julia runtime coupling with legacy High-Performance Computing (HPC) applications (i.e., E3SM), (2) the design of two in-situ data analysis modules in Julia, and (3) the communication design for E3SM and the in-situ Julia modules.