ARCHER, the UK's national supercomputing service offers training in software development and high-performance computing to scientists and researchers across the UK. As part of our training service we are running a series of four half-day Scientific Computing workshops in conjunction with Informatics at University of Edinburgh.
These workshops will be open to Informatics students, but also anyone with an interest is warmly invited to attend via the screencasts. All registered participants will be given access to HPC facilities in order to take part in the practical work and this is therefore an ideal opportunity to get first-time, hands-on experience of using HPC.
ARCHER courses are offered free of charge to all academics.
Lecture sessions will take place in Appleton Tower Lecture Theatre 2 between 14:00 and 17:00
The lectures will take place on the Wednesdays 31st May, 7th, 14th and 21st June 2017.
The lectures will be live-streamed and also recorded and published via the ARCHER YouTube Channel.
Links to the Slides and exercise material for this course will be available here.
The lectures and practical sessions will cover the following topics:
Session 1 - Introduction + HPC
- motivation for large parallel systems such as ARCHER
- parallel architectures and programming models
- methodology of computer simulation
- write serial code for simple traffic model (1D cellular automaton)
- run sharpen exercise in parallel on ARCHER
Session 2 - Computational Science
- parallel decomposition
- quantifying performance
- propose parallelisation approaches using shared and distributed-memory models
- investigate performance of Mandelbrot set
Session 3 - Numerical Scientific Computing
- floating-point numbers
- efficient data storage and access
- testing and validation
- basic numerical analysis
- linear algebra and optimisation packages
- experiment with serial code simulating comet orbits round the sun
- optimisation problem based on comet code
- search for initial position and velocity that result in given final state
Session 4 - Graphical Processing Units (GPUs)
- GPU architecture
- Programming GPUs using CUDA
- Performance optimisation techniques
- CPU <-> GPU data transfer
- Image processing exercise
This course is free to all academics.
Intended learning outcomes
On completion of this course students should be able to:
- explain the motivation for the use of parallel supercomputers in computational science
- describe the main models of parallel programming and propose parallelisation methods for standard problems
- understand the way real numbers are stored on a computer and the way that this affects the accuracy of results
- explain why random numbers are used in many simulations
The course will be delivered over four half-days in person and via live streaming online.
There are no strict pre-requisites for this course, though familiarity with a programming language and concepts will be beneficial.
The course will be held in Lecture Theatre 2, Appleton Tower, 11 Crichton St, Edinburgh, EH8 9LE.
Please use the registration page to register for this ARCHER course.
If you have any questions please contact the ARCHER Helpdesk.