We are delighted that our world-leading Computer Science Department has been included as a host to the UK’s first Intel oneAPI Academic Centre of Excellence.
The centre will conduct research on task-based and GPU programming using oneAPI, and it will also organise workshops and tutorials that will not just be open to our students and researchers, but colleagues from all over the world.
The research will allow Computer Science to deliver a software called OneExaHyPE which can be used to simulate wave phenomena ranging from gravitational waves to tsunamis to earthquakes.
oneAPI is an open and unified programming model that allows to write faster applications for a wide range of computers with multiple processors and Graphics Processing Units (GPUs).
It also enables software developers to solve a specific problem using oneAPI without needing to rewrite the software from scratch whenever they switch to a different machine.
OneAPI relies on an extension of C++, a general-purpose programming language that software developers can use to write codes and run it everywhere including on GPUs, massively parallel multiprocessors or even on programmable hardware such as FPGAs without needing to learn different programming languages.
The new Centre of Excellence at Durham will also work on extending the capabilities of a simulation engine, called ExaHyPE. This is similar to a graphics engine in computer games and allows users to write wave equation solvers quickly and easily with a lot of the technical complexity hidden.
Further development of ExaHyPE could enable it to operate across a wide range of computers powered by many cores and/or Graphics Processing Units (GPU).
The project is a joint collaboration between scientists and research software engineers from Intel, Durham's Advanced Research Computing, and Durham's Department of Computer Science with support from Durham's Institute for Computational Cosmology (ICC).
ExaHyPE is used by Durham's Institute for Computational Cosmology (ICC), Technical University Munich (TUM), the Ludwig Maximilans University (LMU) and the EU’s ChEESE consortium for different tasks such as the mergers of binary black holes, earthquake risk assessment and tsunami propagation.
The code is also used to assess and benchmark upcoming supercomputing technologies.