Lecturer: Dr Zoltán Juhász
The aim of the course is to study the most important parallel numerical algorithms used in scientific computating, to understand the role of parallelization, and to study the relationship between the application of supercomputers and the interaction of the hardware and the algorithm. The course requires independent learning, but the development of parallel implementations and the analysis of their performance is also an important part the course.
Syllabus
• parallel architectures and programming languages
• parallel interpolation, approximation and curve fitting
• parallel solution of systems of linear equations (dense and sparse solvers, different factorization methods)
• direct and iterative parallel partial differential equation solvers
• parallel boundary element methods; parallel solution of tri-diagonal equations
• parallel eigenvalue calculation and eigenvalue decomposition
• numerical integration
• independent component analysis
• fast Fourier transform, different parallel filtering methods
• parallel graph algorithms.
Literature:
John H. Mathews: Numerical Methods for Mathematics, Science, and Engineering, Prentice-Hall International (1992), p. 646
Dimitri P. Bertsekas, John N. Tsitsiklis: Parallel and Distributed Computation: Numerical Methods, Prentice-Hall International (1989), p. 715
Ian Goodfellow, Yoshua Bengio, Aaron Courville: Deep Learning, MIT Press (2016)
Field Cady: The Data Science Handbook, Wiley (2017), Chapter 8.
Selected research papers