From March 27–29, 2017, approximately 70 scientists from the Russian and German scientific communites gathered at the High Performance Computing Center Stuttgart (HLRS) to discuss the state of the art in high-performance computing (HPC) and its applications for scientific and industrial research.
The German-Russian Conference was the second in an annual series of meetings that began in March 2016 in Moscow. Participants included key members of the mathematical and high-performance computing communities in Moscow, Novosibirsk, St. Petersburg, Rostov am Don, and Pokrowsk, Ukraine. In addition to local presenters based at HLRS and the University of Stuttgart, the German research community was also represented by scientists from Ludwig-Maximilians-University Munich, the Albert-Ludwigs University Freiburg, and the Technical University Dresden.
Michael Resch (Director, HLRS) and Boris Chetverushkin (Russian Academy of Sciences) provided welcoming remarks, followed by a packed program of more than 30 scientific presentations. In his welcome address Resch emphasized the importance of scientific collaboration in a world of economic and political competition: “The exchange of scientific ideas and the international collaboration of scientists in workshops like this helps to keep communication channels open in Europe and beyond.”
One area of particular expertise among the Russian delegates was numerical analysis, a mathematical discipline that uses numerical approximations to create representative models of complex phenomena such as weather or logistics. Because high-performance computers have become essential tools for running and analyzing such simulations, one focal point of the conference was how to optimize numerical algorithms for parallel computing systems. Speakers presented several strategies for improving the speed and efficiency of numerical algorithms on HPC systems, focusing on applications in fluid dynamics and in solving complex mathematical problems.
Talks also presented a variety of recent developments in the design and improvement of high-performance computing systems. These included approaches for reducing energy consumption by HPC systems, coping with the hardware and software failures that become more frequent as supercomputers grow in size, wirelessly networking supercomputer components, and forecasting algorithm performance in virtual HPC computing environments.
Several speakers presented recent academic and industrial uses of high-performance computing, providing a snapshot of the many ways in which HPC is facilitating advances in basic research and technology development. Related topics included the use of HPC for simulation in helicopter aerodynamics, storm and air pollution dynamics, ventilation systems in mines, molecular dynamics, materials interfaces, and medical applications, among others.
Also on the program were presentations focusing on the evolving industry surrounding HPC. This included a talk about the growing market for HPC systems in Russia and on the FORTISSIMO program at HLRS, which offers small to medium commercial enterprises access to high-performance computing resources. Additionally, the meeting provided attendees a chance to tour the HLRS facilities, including its computing room and the CAVE virtual reality environment, which turns simulation results into immersive visual representations that can be explored and analyzed more intuitively.
Considering the fruitful and collegial dialogue that grew around the meeting, planning is currently underway for a third annual German-Russian conference, which is scheduled for April 2018 in Kaliningrad. The budding community also hopes to identify specific opportunities for fruitful international collaborations when it returns to Moscow in 2019.
— Christopher Williams