CQUniversity Australia Opts for SGI to Expand HPC

Silicon Graphics International Corp. announced that CQUniversity Australia has selected SGI to supply new HPC server and storage infrastructure to accelerate its research projects.

With the deployment of the SGI Rackable HPC system, CQUniversity can analyse, store and process big data in record time, increasing ROI and results for healthcare and environmental projects.

Compared to existing facilities, the new HPC system will provide 40 times more storage, 14 times the available memory and significantly increase network performance. By offering such HPC capabilities, CQUniversity will be able to compete for and attract more students focused on research activities globally. The acquisition of the SGI system was jointly funded by the Queensland Cyber Infrastructure Foundation (QCIF) and the CQUniversity Research Office and Information Technology Directorate.

Demand for HPC at the university is expected to increase, driven primarily by projects from the CQU Institute for Health and Social Science Research (IHSSR), the Centre for Intelligent and Networked Systems and the Centre for Railway Engineering. In addition, increased usage is also expected as a result of a recent AUD$5.53M investment in health research collaboration between the University of Queensland, Queensland University of Technology and Curtin University.

"CQUniversity is experiencing significant growth in research activity. This new HPC facility not only allows researchers to continue their big data projects, but provides a substantial increase in capacity and capabilities," said Jason Bell, senior research technologies officer, information technology directorate, CQUniversity Australia. "With new areas of research being conducted, this facility will allow researchers access to unprecedented local research computing infrastructure."

One of the current projects expected to benefit from this increased HPC capability is the HEFRAIL Project which looks at the effects of grass cover on erosion on the embankments within the coal rail corridors in Central Queensland. As a consequence of erosion there is an increase in maintenance costs, risks of outages and derailments, interruptions of normal train operations and environmental degradation. By analysing and modelling rainfall-runoff-erosion data gathered from the experimental site, it has been established that 60% grass cover on railway embankment steep slopes reduces erosion by over 90%, compared with the bare scenario. This research has enabled development of conservation strategies, including drip irrigation modelling and design, to combat the effects of erosion.

The SGI Rackable HPC system includes 28 compute nodes each with Xeon E5-2670 processors and 12 GB memory. In addition, the system includes two GPU nodes and a memory node with 64 cores and 512GB of memory. The overall system comprises over 500 processor cores, 4TB of memory and some 300TB of disk storage. System software includes SGI Management Centre, SGI Performance Suite, PBS Pro Scheduler and SUSE Linux Enterprise Server operating system and SGI Data Migration Facility.