Challenges and Opportunities in the Convergence of Extreme-Scale HPC, Big Data and Deep Learning
Dr. Yifeng Cui
San Diego Supercomputer Center
Host: Dr. Kim Bak Olsen
Wednesday, October 3rd at 1pm
CSL 422
Abstract
The highly nonlinear, multiscale dynamics of large earthquakes is a difficult physics problem that challenges HPC systems at extreme computational scales. This presentation will summarize how increased computational and data resources in recent years have contributed to accelerated earthquake simulations at increasing levels of scale and sophistication. The convergence of extreme-scale HPC, Big Data and Deep Learning is becoming the next game-changing opportunity, the future exascale systems will exhibit a new level of complexity, requiring substantially different balance among processor speed, system interconnect and memory bandwidth. To realize the full potential of future machines, there is a time-urgent need to develop a collaborative software ecosystem, encompassing seismic applications, system software, hardware technologies and architectures. The presentation will conclude with a discussion on the future opportunities and technical strategies toward Exascale computing.
Bio
Yifeng cui is the founding director of the High Performance GeoComputing Laboratory at SDSC. He’s also director of Intel Parallel Computing Center for Seismic Simulation at UCSD. Cui is the recipients of NVIDIA Global Impact Award in 2015, the IDC HPC Innovation Excellence Award in 2013, and an ACM Gordon Bell finalist in 2010 for parallel computing research. Cui received a Ph.D. in hydrology from University of Freiburg, Germany.