Practical High Performance Computing (HPC)
Registration link to follow.
This module is suitable for postgraduate students, staff and professionals whose studies include HPC or research topics that would potentially rely on HPC techniques. This module is a broad introduction to this important topic aimed at all users of HPC. It is assumed that participants are already beginner to intermediate users of Linux.
This module follows on from the theory explained in the online course “Introduction to High Performance Computing" (https://moodle.chpc.ac.za/) and aims to build practical, hands-on skills in constructing parallel job scripts for your work flow, including domain-specific tips and tricks.
The module will cover the use of the CHPC cluster super-computer, from how to write, run and debug job scripts, how to build and install parallel software, and how to turn your scientific workflow into an HPC job script. This will include the use of GNU parallel for high throughput computing, and the options for parallel execution of R, python, and Octave. Specifically, participants will gain a thorough understand and practical skills in:
- Cluster Environment, Compilers, Shell Scripting and Job Submission.
- Planning and scripting an HPC workflow
- GNU parallel
- Shared and distributed memory parallel processing
There will be short lectures, interwoven with practical exercises. Participants will be encouraged to follow along with the exercises on their own laptops. Interactive lectures and computer tutorials will introduce students to a range of key aspects of HPC and further illustrate how these tools are currently being applied to address research questions.
The course is suitable for postgraduate students, staff and professionals who have actively programmed on a scientific project and have a current project they are seeking to improve. It is assumed that course participants are familiar with general quantitative scientific practice and at least one programming language, and routinely use those skills in their daily work. No particular scientific field or programming language is required. Participants must make sure they have completed the online course “Introduction to High Performance Computing" (https://moodle.chpc.ac.za/) before the start of this module.
Dr Kevin Colville, Dr Krishna Govender and Dr Andrew Gill are all based at the Centre for High Performance Computing (CHPC) in Cape Town. The CPHPC is one of the three pillars of the National Integrated Cyber Infrastructure System (NICIS). NICIS is a national initiative of the Department of Science and Technology, and implemented by the Council for Scientific and Industrial Research (CSIR). The respective fields of expertise of Dr Colville, Dr Govender and Dr Gill are computational and applied mathematics, computational chemistry and machine learning, and computational mechanics and computational fluid dynamics.