CSIRO launches Virga supercomputer with Dell Technologies
Australia's national science agency, CSIRO, has launched a new high-performance computing facility named Virga. Built in collaboration with Dell Technologies, this system is set to replace the previous generation machine, Bracewell. The new system comprises 111 Dell Technologies PowerEdge XE9640 servers and offers nearly three times the scale and processing power of its predecessor. It includes innovations such as direct liquid cooling and artificial intelligence (AI) capabilities.
Virga has been designed with sustainability in mind, in line with the Treasury's newly announced Sustainable Finance Roadmap and mandatory climate reporting policy. The system uses advanced direct cooling technology, which significantly reduces the energy required to maintain optimal operating temperatures. Unlike traditional high-performance computing clusters that rely on air conditioning, Virga immerses the servers in liquid, thus lowering power consumption and reducing the machine's carbon footprint.
The AI-powered computing capabilities of Virga allow the system to analyse vast data sets more quickly and improve simulation models, thereby accelerating experimentation and aiding scientific research. The partnership between Dell Technologies and CSIRO aims to provide long-term support for Australian science, promising a tangible return on investment in terms of research and the commercial development of technologies that could contribute to Australia's economic growth.
Professor Elanor Huntington, Executive Director of Digital, National Facilities & Collections at CSIRO, commented on the significance of the facility. "CSIRO is proud to be a steward of some of Australia's most important pieces of research infrastructure," she said. "AI is used in practically all fields of research at CSIRO, such as developing world-leading flexible printed solar panels, predicting fires, measuring wheat crops and developing vaccines, just to name a few." She added that high-performance computing systems like Virga also play a crucial role in CSIRO's robotics and sensing work, essential to the recently launched National Robotics Strategy aimed at improving the competitiveness and productivity of Australian industry.
Virga is housed at CDC's Hume Data Centre in Canberra and is named after the meteorological phenomenon where rain evaporates before reaching the ground. This naming honours CSIRO's decades-long research into cloud and rain physics. Dr Jason Dowling from CSIRO's Australian e-Health Research Centre highlighted the importance of the facility for medical imaging and diagnostics. "The new HPC facilities will allow researchers in our Australian e-Health Research Centre to train and validate new computational models, which will help us develop translational software in medical image analysis for image classification, segmentation, reconstruction, registration, synthesis, and automated radiology reporting," he said. One notable project benefiting from Virga is the training of AI models to diagnose pathology from MRI lung scans in children with cystic fibrosis, in collaboration with the Queensland Children's Hospital.
Key features of the Virga cluster include NVIDIA H100 Tensor Core GPUs, which support deep learning, machine learning, and AI, along with 94GB of high-bandwidth memory per GPU. It also boasts a Transformer Engine to enhance AI performance and capabilities, significantly speeding up training times for large models. The system uses 4th Gen Intel Xeon Scalable processors and incorporates hybrid direct liquid cooling to reduce the reliance on energy-intensive air cooling.
Angela Fox, Senior Vice President and Managing Director for Dell Technologies Australia and New Zealand, elaborated on the benefits of the system. "Dell Technologies provides organisations with solutions and expertise that accelerate AI and bring value to their data," she said. "With Dell PowerEdge servers as the foundation, Virga will help create new Australian scientific breakthroughs using its AI capabilities, all the while being both more sustainable and more energy efficient than previous generation clusters."
