The meeting of the seminar
in science, education and industry»
|Chairman of the organizing committee|
academician, rector of Lomonosov Moscow State University
Co-Chairs of the organizing committee
Corr. RAS, Director of Research Computing Center of Moscow State University
academician, dean of the faculty of Computational Mathematics and Cybernetics of Lomonosov Moscow State University
Doctor of Physical and Mathematical Sciences, Professor, Research Computing Center of Moscow State University
Scientific Secretary of the seminar
Ph.D., Research Computing Center of Moscow State University
The seminar was organized on the basis of the Scientific and Educational Center «Supercomputer Technologies». The reports cover all aspects of the use of supercomputers, parallel computing systems and distributed data processing methods for solving large computational problems. The seminar has a pronounced interdisciplinary nature, however, the various nuances of using supercomputer technologies are of interest to researchers from various fields.
The topics of the seminar cover the following main areas (but are not limited to them):
- Models, methods and technologies for applying high-performance and distributed computing to solve scientific and technical problems.
- Theory and practice of solving computational fluid and gas dynamics problems on supercomputers and in distributed computing environments.
- Application of supercomputer technologies for the analysis of highly nonlinear and fast processes in problems of solid and fluid mechanics.
- Supercomputer simulation of the behavior of complex building structures.
- High performance computing in bioengineering and bioinformatics.
- High performance computing and nanotechnology, design of new materials and drugs.
- Building technologies and experience in using GRID systems and services to solve scientific and technical problems.
The seminar is supported by the Moscow Center of Fundamental and Applied Mathematics.
|V.P. Nikolsky, D.G. Pavlov, A.S. Semyonov, V.V. Stegailov, A.V. Timofeev||JIHT RAS, National Research University Higher School of Economics, Moscow Institute of Physics and Technology (NRU), JSC NICEVT|
JIHT RAS Supercomputer Center: introduction of new types of hardware and new parallel programming technologies for mathematical modeling of energy problems
The continuous growth of the needs of the scientific community for high-performance computing systems stimulates the search for the optimal architecture and components of supercomputers in the face of a growing variety of post-Moore technologies in the development of microelectronics. The current shortage of imported element base increases interest in the components of domestic production. At the Supercomputer Center of the JIHT RAS, computing systems based on the domestic Angara interconnect with Intel and AMD EPYC processors and Nvidia and AMD graphics accelerators have been created and are being developed. Immersion oil cooling technologies are used. The report will present the achievements of recent years:
- experience with the Angara network and its latest enhancements regarding GPUDirect technology for AMD accelerators and the use of the Ethernet-over-Angara protocol for I/O tasks;
- experience with AMD EPYC processors and graphics accelerators with AMD Vega20 and Nvidia Ampere architectures;
- experience with immersion cooling versus air cooling;
- experience in using GPU to accelerate the solution of various problems of mathematical modeling (porting the LAMMPS molecular dynamics code to HIP, GPU acceleration of new interatomic potentials, using OpenMM to solve problems of plasma physics and turbulent flows, transfer of plasma calculations by the PIC method to graphics accelerators;
- experience in evaluating ARM and Elbrus processors.
This work was supported by the RSF grant 20-71-10127.
The seminar will be held in the form of a webinar on the Zoom platform.
To participate in the webinar, we ask you to enter your data in the Google spreadsheet:
A link to the Zoom conference will subsequently be sent to your email address.