Leonid Sokolinsky: Exaflop Capacity Supercomputer Will Be Built by 2022

Vice-Rector for Informatization of South Ural State University Leonid Sokolinsky delivered a lecture on models of parallel computing at the Youth Scientific School “High-performance Computing Using Grid Systems”.

Present day brings radical changes to the paradigm of developing numerical algorithms for big tasks, first of all for solving of mathematical models in chemistry, physics, biology, and economics. A model of parallel computing is a system for developing parallel programs. Such model provides a formula using which we may predict the time of execution of a program even before it is started on a computer.

The expert adds that it is important to study the efficiency of the parallel program since not everyone has access to big supercomputers to be able to understand the degree of efficiency of the program running on such a computer.

“The progress in development of computing capacities is a very fast paced one. If we turn to the times of as far as the year 1994, single-processor systems were listed in the TOP500. Approximately by the year 2022 the TOP500 list should include an exaflop capacity supercomputer. What today is a supercomputer will become a common computer 10-15 years from now,” stresses Leonid Sokolinsky.

According to the TOP500 ranking, which comprises information on the computing systems available around the world, the Chinese Sunway TaihuLight is considered to be the most powerful supercomputer in the world. The features of this supercomputer are truly impressive: its number of processor cores amounts to 10,649,600, and the aggregated memory is 1.3 petabytes. At Linpack test it demonstrates the capacity of 93 petaflops.

Let us remind you that the 9th Youth Scientific and Practical School “High Performance Computing on Grid” is being held February 5th through 10th at the Higher School of Information Technologies and Automated Systems of M.V. Lomonosov Northern (Arctic) Federal University.

The School agenda features on-campus lectures, meetings, master classes, as well as teleconferences with Lulea University of Technology (Sweden) and Ural Federal University regarding modern technologies of machine learning.

 

Event date: 
Friday, 9 February, 2018 - 16:15
You are reporting a typo in the following text:
Simply click the "Send typo report" button to complete the report. You can also include a comment.