An Extreme-Scale Implicit Solver for Complex PDEs: Highly Heterogeneous Flow in Earth's MantleScroll Up
For "11 PFLOP/s Simulations of Cloud Cavitation Collapse."Scroll Up
Austin, Texas, November 20, 2015 – A 10-member team led by Johann Rudi of the University of Texas at Austin are the recipients of the 2015 ACM Gordon Bell Prize for their entry entitled An Extreme-Scale Implicit Solver for Complex PDEs: Highly Heterogeneous Flow in Earth's Mantle. The winning team includes representatives from the University of Texas at Austin, IBM Corporation, California Institute of Technology and the Courant Institute of Mathematical Sciences at New York University. The ACM Gordon Bell Prize tracks the progress of parallel computing and rewards innovation in applying high performance computing to challenges in science, engineering, and large-scale data analytics. The award was bestowed during SC15 (sc15.supercomputing.org) in Austin, Texas.
The team presented a solver which can process difficult partial differential equations (PDEs) at an extreme scale to predict activity in the earth's mantle and that scales up to half a million cores. By effectively modeling these processes, scientists can better understand the dynamics that produce earthquakes and related natural disasters. Mantle convection is just one application in the physical sciences wherein processing difficult PDEs at an extreme scale would be useful.
Team members include Costas Bekas (IBM), Alessandro Curioni (IBM), Omar Ghattas (University of Texas at Austin), Michael Gurnis (California Institute of Technology), Yves Ineichen (IBM), Tobin Isaac (University of Texas at Austin), Cristiano Malossi (IBM), Johann Rudi (University of Texas at Austin), Georg Stadler (Courant Institute of Mathematical Sciences), and Peter W.J. Staar (IBM).
Innovations from advanced scientific computing have far-reaching impact in many areas of science and society, from accurately predicting storms and other weather phenomena, to economic forecasts and developing new pharmaceuticals. The annual SC conference brings together scientists, engineers and researchers from around the world for an outstanding week of technical papers, timely research posters, tutorials and Birds-of-a-Feather (BOF) sessions.
Denver, Colorado, November 22, 2013 – Scientists from Switzerland, Germany and the U.S have set a new supercomputing simulation record in fluid dynamics by reaching 14.4 Petaflops of sustained performance to win the 2013 ACM Gordon Bell Prize awards.acm.org/bell.
The simulation, which represents a 150- fold improvement over current state-of-the-art performance levels for this type of application, has potential utility for improving the design of high pressure fuel injectors and propellers, shattering kidney stones, and therapeutic approaches for cancer treatment. The research was conducted by scientists at ETH Zurich and IBM Research, in collaboration with the Technical University of Munich and the Lawrence Livermore National Laboratory (LLNL). The results were presented by the team at SC13 sc13.supercomputing.org in Denver, where the winner of the ACM Gordon Bell Prize was announced on November 21. The simulation conducted by the team resolved unique phenomena associated with clouds of collapsing bubbles. This condition occurs when vapor bubbles formed in a liquid collapse due to changes in pressure. The successful effort employed 13 trillion cells and 6.4 million threads on LLNL’s “Sequoia” IBM BlueGene/Q, one of the fastest supercomputers in the world. The simulation resolved 15,000 bubbles and a 20-fold reduction in time to solution over previous research. The paper describing this achievement was one of six papers chosen as finalists for the 2013 Gordon Bell Prize awarded by ACM acm.org.
Members of the team included Diego Rossinelli, Babak Hejazialhosseini, Panagiotis Hadjidoukas, and Petros Koumoutsakos from ETH Zurich; Costas Bekas and Alessandro Curioni from IBM Zurich Research Laboratory; Adam Bertsch and Scott Futral from Lawrence Livermore National Laboratory; and Steffen Schmidt and Nikolaus Adams from Technical University Munich.