High Performance Computing with Data Science MSc, PgDip, PgCert, PgProfDev (Online Learning) The University of Edinburgh

This programme aims to provide students with in-demand (for both a wide range of industries and academic disciplines) skills and knowledge of the techniques and technologies underpinning parallelism and High-Performance Computing (HPC).

HPC is the use of powerful processors, networks and parallel supercomputers to tackle problems that are very computationally or data-intensive. Data Science involves the manipulation, processing and analysis of data to extract knowledge, and High-Performance Computing (HPC) provides the power that underpins it. You will learn leading-edge HPC technologies and skills to exploit the full potential of the world’s largest supercomputers and multicore processors. This is a well-established programme that has been successful in training generations of specialists in parallel programming. High-Performance Computing is a key area supporting most areas of scientific research and industry.

The flexible structure ensures students acquire the core principles required before proceeding to their choice of more advanced topics and allows students to take on the programme at their own pace. You can study:

• MSc
• Postgraduate Diploma
• Postgraduate Certificate
• Postgraduate Professional Development Level

Further information can be found in the programme structure section, below.

EPCC is the UK’s leading supercomputing centre with staff who are experienced HPC practitioners and is a major provider of HPC training in Europe with an international reputation for excellence in HPC education and research and a well-established on-campus MSc programme that has been successful in training generations of specialists in parallel programming. Students benefit from access to HPC systems with recent examples including ARCHER2 (the UK national Tier 1 supercomputing service ) and Cirrus, a heterogeneous Tier-2 National HPC Facility.

Online learning

Our online learning technology is fully interactive, and award-winning and enables you to communicate with our highly qualified teaching staff from the comfort of your own home or workplace.

Our online students not only have access to the University of Edinburgh’s excellent resources but also become part of a supportive online community, bringing together students and tutors from around the world.

This programme will not require you to run code locally as you will have access to HPC systems provided as part of the programme, however, the ability to code on your device is required - therefore a laptop or desktop computer running Windows, iOS, or Linux is recommended.

The University of Edinburgh is currently ranked 15th in the World in the QS World University Rankings 2023.

The learning outcomes of the programme are to:

• Equip students with an understanding of HPC architectures and technologies.
• Equip students with expertise in advanced tools and techniques for HPC and Data Science software development.
• Enable students to apply this knowledge in order to exploit modern parallel and multicore computing systems and Data Science techniques in key scientific and commercial application areas.
• Enable students to develop skills in problem-solving, project management, independent and critical thinking, teamwork, professionalism and communication.
• Enable students to develop as HPC and Data Science practitioners, able to apply current and emergent technologies in both industry and research.
• Teach the leading-edge programming techniques required to exploit the power of the world’s largest parallel supercomputers.

Graduates from EPCC’s on-campus MSc programmes are in high demand from a wide range of companies ranging from multinationals to SMEs both within the UK, Europe, and internationally as well as strong demand from within academia both as researchers within HPC, computational science fields, data science, and professionally for HPC services and centres underpinning research.

Initial graduate destinations for on-campus students over recent years include ARM, Intel, Amazon, MathWorks, NCR, Avaloq, Global Surface Intelligence, Boston Ltd, ECMWF, Leonardo, STFC, ICHEC, and, EPCC itself with 10 current EPCC staff being graduates of the on-campus programme. Many students also go on to further study opportunities, with 8 current University of Edinburgh PhD students being graduates of the programme.

Programme structure

This programme is available on a part-time intermittent basis: i.e. it is inherently flexible in nature.

During the taught component students are permitted to take up to 30 credits per University semester (Semester 1 runs from early/mid-September to mid-December, Semester 2 runs from early/mid-January to mid/late-May), but in an individual Semester may take zero credits of courses. As fees are due at the course level this means that students have flexibility both financially and in terms of their time commitment.

The credit sizes and course availability listed below may be subject to change but are not expected to. Should this occur applicants/students would be given as much notice as possible.

MSc Structure

The MSc programme may be completed in as little as three academic years or as many as six, with the taught component (comprising 120 credits of taught courses) taking between two and five years and the dissertation component (comprising a 60-credit dissertation course) taking one year.

MSc compulsory courses

• Practical Introduction to High-Performance Computing (20 credits, Semesters 1 & 2)
• Practical Introduction to Data Science (20 credits, Semesters 1 & 2)
• Message Passing Programming (10 credits, Semester 2)*
• Threaded Programming (10 credits, Semester 2)*
• Programming Skills (10 credits, Semester 1)
• Software Development (10 credits, Semester 1)
• Project Preparation (10 credits, Semester 2) **
• Dissertation (60 credits: September - August) ***

*Requires Practical Introduction to High-Performance Computing as pre/co-requisite.

**Cannot be taken prior to Semester 2 of Year 2 and should only be taken as one of the final taught credits.

***Cannot be taken prior to Year 3 and students must meet progression requirements on taught components.

MSc optional courses

• Parallel Design Patterns (10 credits, Semester 1) ~
• Performance Programming (10 credits, Semester 1) ~
• Design and Analysis of Parallel Algorithms (10 credits, Semester 2)
• Advanced Parallel Techniques (10 credits, Semester 2) ~
• Advanced Message Passing Programming (10 credits, Semester 1) * ~
• Numerical Algorithms for High-Performance Computing (10 credits, Semester 2) *
• Plus some optional courses available from the School of Informatics or elsewhere in the College of Science and Engineering (subject to availability)

*Expected to launch in Academic Year 2022/23

~Requires a prerequisite course or course(s) from the compulsory courses

PGDip Structure

The PGDip programme may be completed in as few as two academic years or as many as four. The PGDip comprises the MSc programme taught component (comprising 120 credits of taught courses) and has compulsory/optional course options the same as above, but with the only difference being that the Project Preparation Course is optional for PGDip students, but compulsory for MSc students and that PGDip students do not take a dissertation course.

PGCert Structure

The PGCert programme may be completed in as few as one academic years or as many as two. It comprises 60 credits, all compulsory:

Compulsory Courses:

• Practical Introduction to Data Science (20 credits, Semesters 1 & 2)
• Practical Introduction to High-Performance Computing (20 credits, Semesters 1 & 2)
• Message Passing Programming (10 credits, Semester 2)*
• Threaded Programming (10 credits, Semester 2)*

*Requires Practical Introduction to High-Performance Computing as pre/co-requisite (or equivalent experience).

PPD Structure

Postgraduate Professional Development (PPD) is an unstructured programme of study allowing students to take up to 50 credits of courses from the PGCert Degree Programme Table (DPT) (see list of available courses above) over up to two academic years. The PPD does not offer a final accredited exit award, but certificates for modules completed can be provided. Students interested in an accredited award may wish to instead apply for the PGCert, although a student enrolled on the PPD may apply to transfer to the PGCert, subject to performance on courses taken.