2016/17 Undergraduate Module Catalogue

COMP3920 Parallel Scientific Computing

10 creditsClass Size: 75

Module manager: Dr. David Head
Email: D.Head@leeds.ac.uk

Taught: Semester 1 (Sep to Jan) View Timetable

Year running 2016/17

Pre-requisite qualifications

COMP 2647 Numerical Computation and Visualization
Or
COMP2941 Algorithms II
or
MATH2600 Numerical Analysis

Pre-requisites

COMP2647	Numerical Computation and Visualization
COMP2941	Algorithms II
MATH2600	Numerical Analysis

This module is not approved as a discovery module

Objectives

On completion of this module, students should be able to:
- show awareness of the range of parallel computer architectures and programming paradigms;
- understand the role of parallel scientific computing and the importance of reliability, efficiency and accuracy;
- understand how to design algorithms to make efficient use of parallel architectures and how to predict and measure the efficiency and scalability of an implementation;
- write portable parallel programs using the message passing system MPI;
- understand what partial differential equations are and how they can be approximated by a discrete system which can be solved on a computer;
- implement simple parallel Scientific Computing algorithms in an efficient manner.

Syllabus

Basics of parallel computing: distributed and shared memory architectures; multicore processors; programming paradigms; MPI; threads.

Parallel programming with MPI: basic computational procedures; types of task and data partitioning; load-balancing; communications and synchronization.

Parallel efficiency: speed-up; efficiency; scalability; isoefficiency; performance analysis.

Introduction to parallel scientific computing: reliability, efficiency and accuracy of computational methods; the role of parallel computation; partial differential equations (PDEs) and their use in the modelling of physical systems.

Parallel Scientific Computing problems: boundary value problems; initial-boundary value problems; finite difference approximation; domain decomposition; block and strip partitioning; explicit and implicit time-stepping; red-black ordering.

Advanced parallel scientific computing: complex geometry; unstructured data; graph partitioning; sparse matrices and data structures; direct versus iterative solvers; multigrid and parallel implementations.

Teaching methods

Delivery type	Number	Length hours	Student hours
Lectures	22	1.00	22.00
Class tests, exams and assessment	1	2.00	2.00
Practical	10	1.00	10.00
Private study hours			66.00
Total Contact hours			34.00
Total hours (100hr per 10 credits)			100.00

Private study

Taught session preparation: 18 hours
Taught session follow-up: 18 hours
Self-directed study: 7 hours
assessment activities: 23 hours

Opportunities for Formative Feedback

Attendance and formative assessment

Methods of assessment

Coursework

Assessment type	Notes	% of formal assessment
Assignment	Departmental	20.00
Total percentage (Assessment Coursework)		20.00

This module is re-assessed by exam only.

Exams

Exam type	Exam duration	% of formal assessment
Standard exam (closed essays, MCQs etc)	2 hr 00 mins	80.00
Total percentage (Assessment Exams)		80.00

This module is re-assessed by exam only.

Reading list

The reading list is available from the Library website

Last updated: 07/09/2016

Disclaimer

Browse Other Catalogues

• Undergraduate module catalogue
• Taught Postgraduate module catalogue
• Undergraduate programme catalogue
• Taught Postgraduate programme catalogue

Errors, omissions, failed links etc should be notified to the Catalogue Team.PROD