Infopaket

Methods of parallel calculations

ДВС.1.08

Вибіркова дисципліна для ОП

First

2021/2022

8 Semester

4

LO 1. Demonstrate knowledge and understanding of basic concepts, principles, and theories of fundamental and applied mathematics and use them in practice. LO 5. Be able to develop and use in practice algorithms related to the approximation of functional dependencies, numerical differentiation and integration, solving systems of algebraic, differential and integral equations, solving boundary value problems, finding optimal solutions. LO 9. Build efficient algorithms regarding the accuracy of calculations, stability, speed and cost of system resources for numerical study of mathematical models and solution of practical problems. LO 11. Be able to apply modern technologies of programming and software development, software implementation of numerical and symbolic algorithms. LO 13. Use specialized software products and software systems of computer mathematics in research. ..

1. Know: basic problem statements and methods of solving the main problems of computing mathematicians 2. Be able to: develop, analyze and apply knowledge of numerical analysis for solving basic problems of computational mathematics (systems of linear algebra equations, algebraic problems on eigenvalues, systems of ordinary differentials equations) and applied mathematics using modern computing resources.

1 Principles of building parallel computing systems. 2 Problems of mathematical modeling and methods of solving them based on the development and use of high-performance calculations on the latest computer systems. 3 Evaluation of computational complexity of parallel variants of algorithms. 4 Methods and computer algorithms of parallel data processing for multi-core computers with shared memory. 5 Highlighting the main characteristics of efficiency and acceleration of parallel algorithms. The main stages of the technological scheme of their construction. 6 Definition of tools and strategies for solving mathematical modeling problems for science and engineering problems. 7 Software and Hardware for multi-core systems with shared memory. 8 Large-scale distributed computing systems. Problems of implementation of distributed computing. 9 Basic types and characteristics of communication systems for multi-node computer architectures. 10 Algorithms for research and solving basic problems of computational mathematics based on distributed computing. 11 Principles of algorithm development, efficiency research and algorithm acceleration. 12 Supercomputer systems, modern solutions and trends. 13 Application of information technologies based on distributed computing for science and engineering problems. 14 Principles and tools for creating parallel programs based on distributed computing. 15 Methods and information technologies for mathematical modeling based on technology grids, as a variant of distributed computing. Mathematical and software tools of computing grids. 16 Methods and information technologies for mathematical modeling based on cloud technologies, as a variant of distributed computing. Mathematical and software toolkit of cloud computing. 17 Application of grids and cloud computing for mathematical modeling of physical and technical processes. Mathematics and software.

1. Voevodin V.V., Voevodin Vl.V. Parallel-nye vychisleniia. – SPb.: BKhV-Peterburg, 2002. – 608 s. 2. Khimich A.N., Molchanov I.N., Popov A.V., Chistiakova T.V., Iakovlev M.F. Parallel-nye algoritmy resheniia zadach vychislitel-noi matematiki. – Kiїv: Nauk. dumka, 2010. – 198 s. 3. Gergel- V.P. Teoriia i praktika parallel-nykh vychislenii. BINOM. Labolatoriia znanii, 2007,423 s. 4. Quinn M.J. Parallel Computing. Theory and practice. – New York. – 1994. – 446 p. 5. Ortega D. M. Vvedenie v parallel-nye i vektornye metody resheniia lineinykh sistem / Moskva: Mir, 1991. 367 s. 6. http://www.netlib.org/scalapack ..

Lectures, consultations, independent work

The maximum number of points that can be obtained by a student: 100/60 points. - semester assessment: 1. Modular control work 1: 30 points/18 points. 2. Project defense: 30 points/18 points. - final evaluation (in the form of an exam): - the maximum number of points that can be obtained by a student: 40 points; - learning outcomes that will be evaluated: PH1.1, PH1.2, PH2.1; - form of implementation and types of tasks: written. Types of tasks: 4 written tasks.

Ukrainian