For Immediate Release
November 28, 2014
Dublin, Ireland – The third Irish SUPERCOMPUTER list, ranking the fastest high performance computers in Ireland was released today, and for the first time including entries from Northern Ireland.
The project was launched in November 2013 to raise the profile of High Performance Computing (HPC) in Ireland and abroad. The list, at www.IrishSupercomputerList.org, is updated twice annually (June and November) and there is a continuous open call for participation from users and maintainers of Irish HPC installations.
Marking Northern Ireland’s entry to the Irish Supercomputer List are two machines from Queen’s University Belfast. University College Dublin also has two new entries and one upgraded entry. The combined number of CPU cores over all machines on the current list is over 75,000. The list also shows that Ireland is following a global trend of harnessing coprocessor technologies for HPC. 25% of the machines on the list make use of coprocessors, with a combined total of over 106,000 coprocessor cores. The combined performance of all machines is 648 Tflops/s.
The Irish Supercomputer List is maintained by an independent group of computer science researchers representing industry and academia, including Dublin City University and the College of Computer Training, also in Dublin. Providing an international perspective, are representatives from the University of Oxford and the University of Helsinki. The advisory committee is made of members in Ireland, the UK and the USA.
Speaking of the launch of the November 2014 Irish Supercomputing List, list maintainer Dr. Brett Becker said:
“The Irish Supercomputer List is beginning to show signs of maturity – upgraded machines and decommissioned machines, as well as new entrants. HPC moves fast – machines that are not upgraded are decommissioned quickly, not only because what is considered fast today can often be considered slow tomorrow, but because the power consumption of machines only a few years old can be unsustainably high. It is often cheaper to decommission and replace than to continue operating older machines.”