Developing mini-apps for performance prediction and hardware experimentation

Developing mini-apps for performance prediction and hardware experimentation:

The research area that I am working on is building mini-applications to develop a deeper understanding of the limitations of Rolls-Royce compute codes and exploring new avenues for improving performance with both hardware and software techniques. This is being done in partnership with Rolls-Royce in their Prosperity Partnership Advanced Simulation and Modelling of Virtual Systems (ASiMoV) project

The strategic vision of this project is to enable the research and development of the next generation of engineering simulation and modelling techniques. The aim is to achieve the world's first high-fidelity simulation of a complete gas-turbine engine during operation, simultaneously including the effects of thermo-
mechanics, electromagnetics, and CFD. This level of simulation will require breakthroughs at all
levels, including physical models, numerical solvers, algorithms, software infrastructure, and
Exascale HPC hardware. The partnership uniquely combines fundamental engineering and
computational science research with two high tech SMEs and Rolls-Royce plc to address a challenge
that is well beyond the capabilities of today's numerical solvers.

ASiMoV aligns with the EPSRC Prosperity outcomes, particularly the Productive Nation. Enabling
large scale simulations with unprecedented modelling accuracy will drive innovation in product
design and technology and business innovation through digital transformation.
The lean computing theme will deliver productivity improvements and affordable solutions. A
significant focus of the partnership is to develop a talent pipeline for scientists to provide future
technical leadership in an area critical to UK competitiveness. The extreme computing theme
focuses on large-scale data and mathematical algorithms for Exascale which contribute to the
Connected Nation ambition. This theme will also develop solutions to cyber threats particularly for large distributed simulations in the cloud and so contributes to the Resilient Nation ambition.

Simulation is now well established as the third pillar of scientific investigation and, therefore, of
fundamental national importance. Moreover, each jump in computer power yields corresponding
jumps in simulation capability and, therefore, scientific progress. The computational advances made
in ASiMoV will have be broadly applicable and will therefore contribute to the health of other
simulation-based research disciplines. It is expected ASiMoV to provide a significant stimulus to new industrial players in simulation across the UK.