POP Project Ends but Impact Continues
In October 2015, the European Commission funded the POP project (https://pop-coe.eu) with the aim of establishing a parallel performance analysis Centre of Excellence (CoE). The project has now been brought to a close, having completed more than 125 assessments of application performance studies and proof-of-concept demonstrations on how to achieve higher application performance.
POP has refined and promoted a high-level analysis and modelling methodology. A high-level model quantifies factors such as load balance, serialization, and computational efficiencies, providing fundamental insight into how application performance scales and recommendations were made on the most appropriate directions to refactor the code and thus improve performance.
Service orientation has been a cornerstone of CoE activities, which customers from different scientific domains having benefited from. Around 25 % of the studies were for small- and medium-sized enterprises in an attempt to promote performance analysis in non-traditional sectors.
A culture change is required to improve awareness of the importance of understanding application performance and promote programming best practices. POP has made a significant contribution in this respect. Many of the insights and suggestions provided have been well-received by customers. In many cases, performance gains of between 10 % and up to ten times - or even twenty times – have been achieved by the proof-of-concept service or through code optimizations by the developers themselves.
The POP CoE will continue to pursue its activities during the next few months with limited resources. The POP partners expect to continue these activities in the future, supporting the improvement of productivity and efficiency with respect to how research and industry sustainably use available HPC resources. Whilst the project has come to a close, its impact will continue in various dimensions.
(Contact: Dr. Bernd Mohr, b.mohr@fz-juelich.de)
JSC News No. 257, 24 April 2018