The UK can’t continue its shambolic stop-go approach to supercomputing

With the announcement that plans for a £800 million exascale supercomputer in Edinburgh have been scrapped, the authors point out that the new government has not prioritised bringing the UK up to speed with our counterparts in the US, Europe and China. 
 
The authors argue that it is very unfortunate to exclude traditional HPC users in the UK access to the new Isambard-AI centre in Bristol, a policy which was also being imposed on the more recent machines whose procurements have been abandoned. The thousands of high-performance GPUs could be of great value to all of UK computational research, not just within the realms of AI. 
 
Professor Coveney and Dr Highfield conclude that the UK requires much stronger commitment to this technology, which goes beyond the current focus on AI and merges with HPC and conventional physics-based approaches to supercomputing. Substantial long-term funding should be provided for the education, training, recruitment and retention of a workforce with the expertise to exploit this technology for the strategic benefit of UK. 
 
Read the full article here.