Section: Overall Objectives
With the Inria GRAND-LARGE Project-Team, we are involevd in the G8 project entitled “Enabling Climate Simulation at Extreme Scale" (https://wiki.engr.illinois.edu/display/G8/G8+ESC++--+Enabling+Climate+Simulations+at+Extreme+Scale ) which also involves research groups from Europe, Japan and North America.
With University of Tennessee (ICL) and University of Colorado at Denver an associated team has been initiated, which name is MORSE (http://www.inria.fr/en/teams/morse ). The goal of Matrices Over Runtime Systems at Exascale (MORSE) project is to design dense and sparse linear algebra methods that achieve the fastest possible time to an accurate solution on large-scale multicore systems with GPU accelerators, using all the processing power that future high end systems can make available. To develop software that will perform well on petascale and exascale systems with thousands of nodes and millions of cores, several daunting challenges have to be overcome, both by the numerical linear algebra and the runtime system communities. By designing a research framework for describing linear algebra algorithms at a high level of abstraction, the MORSE team will enable the strong collaboration between research groups in linear algebra and run-time systems needed to develop methods and libraries that fully benefit from the potential of future large-scale machines. The first outcome of this associated team is the release of the MAGMA package (http://icl.cs.utk.edu/magma/ ).
The thesis of Mathieu Chanaud (in collaboration with CEA/CESTA) has led to the design and the parallel implementation of an hybrid solver combining a parallel sparse direct solver and full multigrid cycles. A 1.3 billion unknown sparse linear system, arising from the discretization of the 3D Maxwell equations on a fully unstructured mesh, has been solved very efficiently on the CEA/DAM TERA100 supercomputer.