By Prof. David Abramson, RCC Director
I recently spent a few weeks visiting colleagues at the Barcelona Supercomputing Center in Catalonia. Interestingly, I arrived just in time for the National Day of Catalonia (Diada), on 11 September, which commemorates the fall of Barcelona during the War of the Spanish Succession in 1714. These days the event also serves as an opportunity to protest and push for secession from Spain, and I was greeted by millions of Catalonians lining the Ave Diagonal, which runs through Barcelona. (I suspected they weren’t just out to welcome me!)
The Barcelona Supercomputing Center–Centro Nacional de Supercomputación (BSC-CNS) is the national supercomputing centre in Spain. They specialise in high-performance computing (HPC) and manage MareNostrum, one of the most powerful supercomputers in Europe, located in the Torre Girona chapel.
I held specific meetings with two groups at BSC-CNS, with potential benefits for UQ researchers. These were the Performance Tools group led by Judit Gimenez and the Workflows and Distributed Computing group led by Rosa Badia. I also met the Director of the Computer Sciences Department, Jesus Labarta, who is involved in a number of groups as well as running the overall centre.
At RCC, we are planning collaboration with both groups over the coming year. In particular, the tools developed in the Performance Tools group may allow us to analyse the behaviour of software on UQ HPCs Tinaroo, FlashLite and Wiener, to discover performance and scalability bottlenecks. We are in the process of installing these tools, and once we know more about how they work, we will engage with selected RCC users to test them further.
The BSC-CNS Workflows group has developed a range of tools with similar functionality to our Nimrod software, and we are planning a number of collaborations that combine the best features of these two systems. Initially, we will install the Barcelona workflow system, COMPSs, which is a Python-based workflow tool, at UQ, and will install Nimrod at BSC-CNS. While there is a great degree of overlap, each of these systems caters for particular sweet spots, and having all tools available will be of significant advantage for UQ researchers.
While at BSC-CNS, I gave a talk titled, “Data Centric Debugging: Scaling to Infinity and Beyond?”
Before visiting Spain, I was in Lyon, France, to give a talk titled, “Energy Efficiency Modelling of Parallel Applications” at the Clusters, Clouds and Data for Scientific Computing 2018 workshop. Speakers presented their research and interacted with workshop participants on the future software technologies that will provide for easier use of parallel computers. My talk was about work performed with RCC researchers Mark Endrei, Dr Chao Jin and Dr Minh Dinh, colleagues at Cray, Dr Luiz DeRose and Heidi Poxon, and Dr Bronis de Supinski from Lawrence Livermore National Laboratories in the US. The work has been funded by the Australian Research Council in an ARC Linkage grant.