Exascale Climate Emulators: Enhancing Earth System Model Outputs and Reducing Petabytes of Storage
Speaker
Sameh Abdulah
King Abdullah University of Science and Technology
Title:
Exascale Climate Emulators: Enhancing Earth System Model Outputs and Reducing Petabytes of Storage
Abstract:
We present an exascale climate emulator to tackle the rising computational and storage demands of high-resolution Earth System Model simulations. Using spherical harmonic transform, we model spatio-temporal climate variations stochastically, providing tunable resolution and significantly enhancing emulation fidelity. We extend linear solver software to multi-precision arithmetic GPUs, adapting to different correlation strengths. The PaRSEC runtime system optimizes parallel matrix operations by balancing computation, communication, and memory. Our BLAS3-rich code is optimized for diverse GPU systems, achieving 0.523 EFlops/s on 4,096 ORNL Frontier nodes (projected 1.04 EFlops/s on 8,192 nodes), 0.243 EFlops/s on 1,024 Cineca Leonardo nodes, and 0.375 EFlops/s on 3,072 ORNL Summit nodes.
Bio:
Sameh Abdulah obtained his MS and Ph.D. degrees from Ohio State University, Columbus, USA, in 2014 and 2016, respectively. Presently, he serves as a research scientist at the Extreme Computing Research Center (ECRC), King Abdullah University of Science and Technology, Saudi Arabia. His research focuses on various areas, including high-performance computing applications, big data, bitmap indexing, handling large spatial datasets, parallel spatial statistics applications, algorithm-based fault tolerance, and machine learning and data mining algorithms. Sameh was a part of the KAUST team nominated for the ACM Gordon Bell Prize in 2022 for their work on large-scale climate/weather modeling and prediction.
Exascale Climate Emulators: Enhancing Earth System Model Outputs and Reducing Petabytes of Storage
Abstract:
We present an exascale climate emulator to tackle the rising computational and storage demands of high-resolution Earth System Model simulations. Using spherical harmonic transform, we model spatio-temporal climate variations stochastically, providing tunable resolution and significantly enhancing emulation fidelity. We extend linear solver software to multi-precision arithmetic GPUs, adapting to different correlation strengths. The PaRSEC runtime system optimizes parallel matrix operations by balancing computation, communication, and memory. Our BLAS3-rich code is optimized for diverse GPU systems, achieving 0.523 EFlops/s on 4,096 ORNL Frontier nodes (projected 1.04 EFlops/s on 8,192 nodes), 0.243 EFlops/s on 1,024 Cineca Leonardo nodes, and 0.375 EFlops/s on 3,072 ORNL Summit nodes.
Bio:
Sameh Abdulah obtained his MS and Ph.D. degrees from Ohio State University, Columbus, USA, in 2014 and 2016, respectively. Presently, he serves as a research scientist at the Extreme Computing Research Center (ECRC), King Abdullah University of Science and Technology, Saudi Arabia. His research focuses on various areas, including high-performance computing applications, big data, bitmap indexing, handling large spatial datasets, parallel spatial statistics applications, algorithm-based fault tolerance, and machine learning and data mining algorithms. Sameh was a part of the KAUST team nominated for the ACM Gordon Bell Prize in 2022 for their work on large-scale climate/weather modeling and prediction.