HPC CLUSTERING POWER FOR TRILLIONS OF POSSIBILITIES High-performance computing (HPC) pushes the envelope of traditional computing performance to solve advanced computation problems across a wide range of vertical markets. HPC uses supercomputers and computer clusters to solve complex engineering problems and includes computational fluid dynamics, simulation, modelling, and seismic tomography (particularly in the petrochemical industry). XENON delivers power-efficient and performance-optimised scalable supercomputing solutions based on CPU and GPU's providing many-/multi-core systems within large clusters. XENON's customised HPC solutions enable organisations to process, transfer and manage huge data sets using the latest industry technologies. Scientific research, education, defence, manufacturing & film are some of the demanding industries where XENON's HPC solutions have been successfully employed.

Sciences - CSIRO

XENON deployed Australia's first GPU High Performance Compute Cluster, based on hybrid computation technology, which combines the Central Processing Units (CPUs) with more powerful Graphics Processing Units (GPUs), for the CSIRO.

The GPU supercomputer which is located in the Canberra Data Centres (CDC), contains 1024 Intel Xeon 2.8GHz processor cores which is complemented by 61,440 NVIDIA Tesla streaming processor cores and will help scientists to explore the use of GPU technology for parallel processing.

CSIRO's supercomputer will be among the world's first to combine traditional CPU computing with more powerful GPU highly-parallel computing, providing a world class computational and simulation science facility. In addition, to its immense computational power, the GPU cluster also heralds major environmental benefits. This unique heterogeneous cluster architecture is more energy efficient and significantly cheaper to run than a traditional CPU-only supercomputer.

In April 2012 CSIRO choose XENON to upgrade their supercomputing facility with Australia's first Intel® Xeon® E5-2600 GPU HPC Cluster. This GPU cluster upgrade gave researchers significant productivity gains to help speed up scientific discovery and be more internationally competitive.

CSIRO choose XENON on account of its ability to best meet highly specific technology requirements, while helping CSIRO to minimise its impact on the environment, remotely manage the system and extract the highest performance possible from its investment in scientific and engineering computational technology.

CSIRO's head of Computational and Simulation Sciences, Dr John A Taylor, leads the CSIRO's research team to develop new methodologies for analysing and exploiting large and complex data sets and data streams. The new GPU HPC cluster helps CSIRO scientists increase productivity by performing computations in a single morning that used to take weeks. It also helps them create considerable scientific opportunity by giving the ability to perform tasks that were once computationally infeasible.

"We're in the business of turning information into knowledge and quantifying uncertainty to help people make better decisions, whether that's in bio-chemical marine models, climate modelling or for medical imaging," said Dr Taylor. "The faster we can process data, the faster we can get solutions to market – and that's in Australia's best interests. We chose XENON due to a combination of its expertise in delivering the project according to our very specific requirements, and our history of knowing the quality of work the team delivers."

Higher Education - NeCTAR (University of Melbourne) - Powering Australia's First Research Cloud

XENON delivered the first national research cloud purpose-built for Australian researchers at The University of Melbourne.

The world-leading NeCTAR research cloud, which went live on February 2012, now enables Australian researchers of all disciplines to respond rapidly to new developments by providing instant access to scalable computing resources and applications. Computational results are easily shared with national and international collaboration partners. It fosters innovation in research software applications and services by reducing barriers to rapid deployment and sharing of research applications.

Researchers' needed a dedicated cloud to allow scalable and elastic processing, storage, security and billing models for their vast computations and research requirements, and to take research workflows to the next level.

The University of Melbourne won the contract to commission the first node of the NeCTAR research cloud in October 2011, and further nodes of the cloud will be commissioned by other research institutions throughout 2012. It went to tender for a supplier of compute servers (EIA-310-D rack-mountable chassis containing CPUs, RAM modules, HDDs, motherboard, power supplies and all other necessary components), and awarded the deal to, XENON on account of its expertise in providing high performance and environmentally sustainable computing systems, project team fit, together with the ability to deliver the project within the required budget.

Defence – DSTO - Air Operations Division (AOD)

XENON supplied multiple Image Generator Nodes to the Air Operations Simulation Centre (AOSC) of the Air Operations Division (AOD) of the DSTO. These nodes are part of a high bandwidth, low latency visualisation cluster interconnected by high speed Infiniband. The cluster generates, in real time, a set of coherent images for the purpose of flight simulation. The graphics sub system is driven by NVIDIA Quadro graphics.

Higher Education - (VPAC)

Victorian Partnership for Advanced Computing (VPAC) selected XENON to supply and commission their LINUX based 64-bit cluster system. The cluster currently consists of 95 AMD64 Opteron Compute nodes totalling 760 processor cores, 3TB of RAM and 125TB of useable disk space interconnected by high-speed Infiniband technology.

The implementation of a new cluster is part of a long term strategy by VPAC to sustain a growth in Victorian and Australian Advanced computing in line with international technology trends. XENON's proposed solution was able to meet VPAC's criteria of achieving the maximum number of cores and the minimum latency within a nominated power budget. XENON's proven and demonstrable capability to deliver cluster solutions within a stipulated time-frame was another critical factor in the selection process.