Today the ARCHER national supercomputing service in the U.K. announced that it has deployed the Allinea MAP profiling tool to enable computational scientists and developers to tackle scalability and performance in key scientific applications that run on their 2.5 Petaflop Cray XC30 system.
The ARCHER system, with 118,080 Intel Ivy Bridge cores, is jointly owned by the EPSRC and NERC research councils.
As core counts have increased, so has the need for tools that can handle these levels of parallelism. Allinea MAP provides us with a scalable way to visualize performance bottlenecks and give hints for optimization. This can only become more important and we see Allinea MAP as a key tool in our armoury,” said Mark Parsons, Executive Director of the Edinburgh Parallel Computing Centre that hosts and manages the ARCHER system.
Scientists rely on ARCHER to provide capability computing that cannot be achieved with their own university systems.
The ARCHER Computational Science and Engineering (CSE) team works with the scientists to ensure the applications are able to use the system efficiently and able to produce those world class research results.” says Andy Turner, CSE Team Leader. “Running applications at these larger scales often exposes issues that are not visible at smaller scales. We evaluated Allinea MAP and found it complimented other profiling tools available on ARCHER, giving clear and easily interpreted performance profiles even when running on tens of thousands of cores,” added Andy Turner.
The ARCHER system already uses the Allinea DDT debugger at extreme scale and adding Allinea MAP is multiplying the benefits for scientists and the application support team already.
The system and its applications team deliver a platform that is enabling the UK’s scientists to demonstrate world-class results and publications,” said David Lecomber, CEO of Allinea Software. “We are delighted to be providing the developers and scientists on the ARCHER system with our highly scalable unified tools for debugging and performance profiling. This helps them to work more effectively and their codes to use the machine more efficiently – which means that they will have an even greater impact on UK and worldwide science.”
