Performance Modeling, Benchmarking and Simulation of High Performance Computer Systems

08:30 - 09:00
ML-based Performance Portability for Time-Dependent Density Functional Theory in HPC Environments

Adrian Perez Dieguez, Khaled Ibrahim
Lawrence Berkeley National Laboratory, CA

Min Choi, Bryan Wong
University of California, Riverside, CA

Xinran Zhu
Cornell University, NY

Time-Dependent Density Functional Theory (TDDFT) workloads are an example of high-impact computational methods that require leveraging the performance of HPC architectures. However, finding the optimal values of their performance-critical parameters raises performance portability challenges that must be addressed. In this work, we propose an ML-based tuning methodology based on Bayesian optimization and transfer learning to tackle the performance portability for TDDFT codes in HPC systems. Our results demonstrate the effectiveness of our transfer-learning proposal for TDDFT workloads, which reduced the number of executed evaluations by up to 86% compared to an exhaustive search for the global optimal performance parameters on the Cori and Perlmutter supercomputers. Compared to a Bayesian-optimization search, our proposal reduces the required evaluations by up to 46.7% to find the same optimal runtime configuration. Overall, this methodology can be applied to other scientific workloads for current and emerging high-performance architectures.