Infopaket

Methods of Parallel Computation

Вибіркова дисципліна для ОП

First

2021/2022

7 Semester

4

PL16. Perform parallel and distributed computation, apply numerical methods and algorithms for parallel structures, parallel programming languages in the development and operation of parallel and distributed software. PL18.2. Analyze, evaluate and select tools and computing tools, paradigms, technologies, algorithmic and software solutions in the design and development of software systems.

Full-time form

1. Know: basic concepts of mathematical logic and theory of algorithms, architecture of computer systems, mathematical analysis, formal models of algorithms and complexity of algorithms. 2. Be able to: perform structuring of programs into independent fragments of calculations taking into account the target architecture of the parallel system, as well as to assess the performance of parallel software. 3. Have the means: synchronization and interaction of parallel processes in modern programming languages (java or C #).

The aim of the discipline is to provide with competence, knowledge and skills at the level of using the latest advances in parallel programming of multiprocessor systems and high-performance computing required for modern computer and software technologies in accordance with the qualification. As a result of studying the discipline the student must: know: types and levels of parallelism of operations in computer systems; characteristics of multiprocessor systems architecture; paradigms of parallel programming; tools for programming parallel computing systems with shared memory; problems of deadlocks and mutual exclusion and their methods of solution; model of parallel computing with distributed memory; methods of developing parallel programs; be able to: choose the architecture of a multiprocessor system for the appropriate class of application tasks; use parallel programming tools appropriate to the target architecture; develop parallel algorithms for solving applied problems.

1. F.I. Andon, A.E. Doroshenko, K.A. Zhereb, R.S. Shevchenko, E.A. Yatsenko. Metody Algebraicheskogo Programmirovania. Formal’nyye Methody Razrabotki Parallel’nykh Program. Kiev, “Naukova Dumka”, 2017. 440 s. 2. Doroshenko A.Yu. Lectsii z Paralel’nykh Obchysliuval’nykh System. Metodychnyi posibnyk. Kyiv, Vydavnychyi Dim “KM Akademia”, 2003. 42 s. 3. A.E. Doroshenko. Matematicheskiye Modeli i Metody Organizatsii Vysokoproizvoditel’nykh Parallel’nykh Vychislenii. K., “Naukova Dumka”, 2000. 177 s. 4. A.Yu. Doroshenko, G.S.Finin, G.O. Tseitlin. Algebroalgorytmichni Osnovy Programuvannia. Objektna Orientatsia I Paralelizm. Kyiv, “Naukova Dumka”, 2004. 458 s. 5. Nemniugin S.A., Stesik O.L. Parallel’noye Programmirovanie dlia Mnogoprocessornykh Vychislitel’nykh Sistem. SPb.: BHV-Peteburg, 2002. 400 s. 6. P.I. Andon, A.Yu. Doroshenko, K.A. Zhereb, O.A. Yatsenko. Algebra-algorithmic models and methods of parallel programming. Kyiv, VD “Akademperiodyka”, 2018. 192 s. 7. http:/parallel.ru/.

Lectures, independent work, tests, homework, exam.

- Semester assessment: 1. Test 1: LO 1.1, LO 1.2, LO 1.3 — 30 points/18 points 2. Test 2: LO 2.1, LO 2.2, LO 2.3, LO 2.4 — 30 points/18 points - Final assessment (in the form of an exam): - maximum number of points: 40 points; - learning outcomes which shall be assessed: LO1.1, LO1.2, LO1.3, LO2.1., LO2.2.? LO2.3., LO2.4. - form of conducting: written work.

Ukrainian