“Using Artificial Intelligence and Transprecision Computing for Accelerating Finite-Element Urban Earthquake Simulation”

The 1st R-CCS International Symposium 18 Feb,2019
"Using Artificial Intelligence and Transprecision Computing for Accelerating Finite-Element Urban Earthquake Simulation"

To address problems that occur due to earthquakes in urban areas, we propose a method that utilizes artificial intelligence (AI) and FP16-FP21-FP32-FP64 transprecision computing to accelerate a nonlinear dynamic low-order unstructured finite-element solver. The AI is used to improve the convergence of iterative solver leading to 5.56-fold reduction in arithmetic count from a standard solver. The developed solver demonstrated high scalability and performance up to 4096 nodes of Summit, 4608 nodes of Piz Daint and 49152 nodes of K computer. The proposed approach utilizing AI and transprecision computing has implications for accelerating other implicit solvers used for earthquake city simulations as well as various fields.

Kohei Fujita
(The University of Tokyo)