Podcasts

 

SC15: ACM Gordon Bell Prize Winners Supercompute Deep Earth

Published on November 24, 2015 by Jorge Salazar

Gordon Bell Prize co-winners Johann Rudi and Omar Ghattas of the Institute for Computational Engineering and Sciences, University of Texas at Austin.

The 2015 ACM Gordon Bell Prize, given in recognition of outstanding achievement in high-performance computing, was awarded to researchers Johann Rudi and Omar Ghattas of the Institute for Computational Engineering and Sciences at the University of Texas at Austin. Dr. Ghattas is also a professor in the Jackson School of Geosciences and the Department of Mechanical Engineering, UT Austin.

They share the award with their study co-authors, who utilized the Stampede supercomputer of the Texas Advanced Computing Center and the IBM Sequoia supercomputer at Lawrence Livermore National Laboratory.

The award-winning study modeled the flow thousands of kilometers deep in the mantle, which moves Earth's plates and triggers unpredictable events like volcanic eruptions and massive earthquakes.

The SC15 supercomputing conference took place in Austin, November 15-20, 2015. SC showcases the latest in high performance computing, networking, storage and analysis to advance scientific discovery, research, education and commerce.

The study, "An Extreme-Scale Implicit Solver for Complex PDEs: Highly Heterogeneous Flow in Earth's Mantle," was funded in part by the National Science Foundation and the Department of Energy.

Co-authors include Johann Rudi and Omar Ghattas of ICES; A. Cristiano Malossi, Peter Staar, Yves Ineichen, Costas Bekas, and Alessandro Curioni at the Foundations of Cognitive Solutions, IBM Research – Zurich, Switzerland; Tobin Isaac of ICES; Georg Stadler of the Courant Institute of Mathematical Sciences, New York; and Michael Gurnis of the Seismological Laboratory at CalTech.

Omar Ghattas: The absolute large number of cores and the big scaling result was done on the IBM system at Livermore. We couldn't have done it without the IBM guys.

But the actual science runs, a lot of the day-to-day stuff, testing of the solvers - Johann was using TACC's Stampede system.

Johann Rudi: Mainly, my research was done on the Stampede supercomputer at TACC. One part of the work is developing algorithms. That was wholly supported by TACC machines.

And also, the help that I got from TACC a couple of times was very valuable to me. There were certain small issues that I couldn't even see from where I was working with the machine. But people from the internal status, running the systems, they could see when something was going wrong. They actually helped a lot.

I was happy to work with TACC. Especially Bill Barth. I remember him helping me a lot. I was glad.

The development of the solvers was done on TACC. Also, everything in the paper that shows the scientific results, the visualizations - these were also done on TACC machines. The science part was also supported by TACC.



Related Link(s)

SC15 Award Presentation, "An Extreme-Scale Implicit Solver for Complex PDEs: Highly Heterogeneous Flow in Earth's Mantle."

ICES: Ghattas team wins Gordon Bell Prize, Biros students take top supercomputing awards