To enable businesses to take advantage of our computing and artificial intelligence (AI) technologies, we need to be able to do more than simply give access to hardware and code.
As an example of this, we're collaborating with Lawrence Livermore National Laboratory and IBM Research to port and optimise the Global Extensible Open Power Manager (GEOPM) framework for use on heterogeneous supercomputers – supercomputers which use more than type of one processor. This open-source framework, developed by Intel, is designed to improve power efficiency of future exascale supercomputers through software that can automatically make intelligent decisions based on the type of applications it is running. The ultimate goal is to improve performance and energy efficiency simultaneously, for more sustainable supercomputers that don't compromise on speed.
AI Technical Computing also encompasses workflow optimisation from our energy efficient computing research, to enable complex applications to be deployed in a low-cost way on our IBM Power platforms.
To find out about using the Hartree Centre's AI resources for your applications – contact us