Franco-American Collaborative Research Is Awarded American Prize
© Matthieu Dorier
Matthieu Dorier received the second prize at the ACM Student Research Competition, held in conjunction with the International Conference on Supercomputing, in Tucson (Arizona), from 1st to 3rd June, 2011. He was awarded this prize while doing a Masters internship at the Joint Laboratory for Petascale Computing-JLPC, a joint laboratory set up by Inria and the University of Illinois at Urbana-Champaign.
What was the focus of the work for which you received this award?
Matthieu Dorier: My research work focuses on the interaction between large-scale simulations and storage systems. My proposed approach limits the impact that writing data has on the computations performed by the simulation. In this case, I worked with a simulation of tornadoes, generating a colossal amount of data which must be stored efficiently. This approach could also be adapted to other areas, such as astrophysics or nuclear fusion simulations.
Can you tell us about your three-month internship at the Inria-Urbana-Champaign joint laboratory?
Matthieu Dorier: It was a very pleasant environment, offering optimal work conditions, and the people I worked with, including Franck Cappello, one of my internship supervisors, were all very supportive. In my PhD thesis which I will start in September, I intend to further develop my research within the framework of this collaborative effort, so I will be completing a number of further internships at UIUC. This will allow me to have access to American HPC platforms, and frequent contact with the research world in the United States, without having to worry about the time difference...
How do you see this collaborative effort between Inria and the University of Illinois at Urbana-Champaign?
Matthieu Dorier: One of the objectives of the joint lab is to provide solutions for the efficient operation of the next supercomputer from the NCSA (National Center for Supercomputing Applications), Blue Waters. France has a good candidate supercomputer (ranked among the top ten worldwide), but the United States has numerous very powerful supercomputers. Our papers will always have a lesser impact than those papers presenting experiments using hundreds of thousands of computing cores. This is a shame for France, and is why we need to collaborate with U.S. teams as we do via this "joint lab."
Frank Cappello - © Inria Photo Jim Wallace
Productive partnership for high performance computing
A successful collaboration