Techniques of High-Performance Computing (PHAS0102)

Key information

Faculty: Faculty of Mathematical and Physical Sciences
Teaching department: Physics and Astronomy
Credit value: 15

Restrictions: In order to take this module, students will need to understand basic programming, which is not taught as part of the module. The main programming language in the course is Python. Students should have a good grasp of Python and its concepts (functions, objects, etc.). We will use some low-level programming in C. While strong prerequisites in C are not necessary, students will be required to engage with C programming. It is strongly recommended that students have a basic knowledge on working with a Linux/Unix command line. A basic understanding of simple PDEs (e.g. Laplace equation, wave equation, heat equation), and their numerical solution is required. Students also need a working knowledge of basic techniques from linear algebra.
Timetable

Alternative credit options

There are no alternative credit options available for this module.

Description

Outline and Aims:

The aim of this module is to give an introduction into modern techniques of High-Performance Computing. The emphasis is on a balanced mix between the necessary theoryand practical examples. At the end students will be able to apply modern HPC techniques to implement and solve a wide range of typical PDE problems on advanced computing platforms.

Intended Learning Outcomes:

On successful completion of the module the student should be able to:

Choose the right HPC technologies and parallel software libraries for a given problem;
Write fast algorithms for CPUs and GPUs;
Be able to use parallel sparse-matrix techniques, linear and iterative solvers;
Apply a range of preconditioning techniques;
Develop a computational simulation from the initial PDE up to the solver and post-processing step.

Teaching and Learning Methodology:

In addition to timetabled lecture hours, it is expected that students engage in self-study in order to master the material. This can take the form, for example, of practicing example questions and coding, and further reading in textbooks and online.

Indicative Syllabus:

Part 1: HPC Technologies听

Design of HPC Systems
Parallelization Technologies,
Python for High-Performance Computing

Part 2: An introduction to many-core technologies via OpenCL听

Basics of many-core processing
Differences between CPUs and GPUs
The OpenCL Programming Model
Design of OpenCL kernels

Part 3: Sparse matrices and linear algebra听

Data storage schemes for sparse matrices
Sparse Matrix Vector Products
Sparse LU Decomposition
Iterative solvers
Efficient implementation of sparse matrix operations in OpenCL

Part 4: Many particle simulations听

Basis of particle simulations
Acceleration of particle simulations
Implementation in OpenCL

Part 5: Spectral methods听

Basic ideas of spectral methods (Fourier and Chebychev type methods)
Spectral differentiation matrices
Implementing spectral methods to solve wave problems

Module deliveries for 2024/25 academic year

Intended teaching term: Term 1 听听听 Postgraduate (FHEQ Level 7)

Teaching and assessment

Mode of study: In person
Methods of assessment: 100% Coursework
Mark scheme: Numeric Marks

Other information

Number of students on module in previous year: 32

Intended teaching term: Term 1 听听听 Undergraduate (FHEQ Level 7)

Teaching and assessment

Mode of study: In person
Methods of assessment: 100% Coursework
Mark scheme: Numeric Marks

Other information

Number of students on module in previous year: 48

Last updated

This module description was last updated on 8th April 2024.

听

果冻影院