RMIT supercomputer open for business

Australia’s first university cloud supercomputing facility, which allows researchers to test ideas and solutions up to 80 times faster than existing onsite servers, is now open to industry partners.

RMIT University’s AWS Cloud Supercomputing facility, or RACE, opened in July this year for RMIT researchers, ­­­who are now using it to power advances into battery technologies, photonics and geospatial science. External research partners are now able to use it too.

RACE provides fast, secure and private connections — powered by Amazon Web Services (AWS) and AARNet — ideal for workloads that require higher speed and fewer delays than the internet. RACE Director Dr Robert Shen said the increased bandwidth gives researchers, students and industry partners the ability make discoveries faster and for RMIT to fast-track the time between initial concepts and products going to market.

­­“RACE will enable researchers to test out ideas and solutions up to 80 times faster compared to the existing on-premises servers,” Shen said.

“Research typically involves many failures before success: this facility lets researchers fail quickly so they can fine-tune their solutions and improve them.”

AWS Chief Technologist for Australia and New Zealand Simon Elisha said high-performance computing is key to solving the most complex problems across many industries.

“AWS’s portfolio of cloud services allows researchers at RMIT to focus on ground-breaking research, across a broad range of sectors, and innovate faster. Using AWS, RMIT can securely deliver advanced computer performance, memory capacity, and scalability,” he said.

AARNet CEO Chris Hancock said the high-speed internet and communication services provided for RACE were designed to service both current and future demand.

“The network AARNet has deployed for RACE is high capacity and engineered to scale to 400 Gbps to provide RMIT researchers with plenty of headroom for transferring massive amounts of data to AWS on demand, now and into the future,” Hancock said.

Early adopters forging ahead

The new service is already making a difference for the RMIT research groups who’ve used it since July this year, prior to its public launch.

Professor Michelle Spencer has used it to analyse data and communicate a new ultra-fast way to screen hundreds of potential molecules that could make suitable electrolytes for lithium-metal batteries, which could potentially increase storage capacity by tenfold.

“We are using RACE to analyse our data and produce high-resolution animations that help us to interpret our data and communicate our research findings,” Spencer said.

Matt Duckham’s team is using the computing power to design new ways to automatically pinpoint a person’s exact location using just a verbal description of the features around them. This approach could be especially important in emergencies if satellite positioning fails.

Duckham said his team now has the ability to process massive information streams including drone imagery, satellite data, data from sensor networks and crowd-sourced data that could overwhelm conventional computing facilities.

“Enabling us to analyse these huge volumes of data from new sources can help better inform evidence-based policy decisions to improve public transport, traffic, infrastructure and many other aspects of quality of life,” he said.

Associate Professor Thach Nguyen and team at the Integrated Photonics and Applications Centre rely heavily on high-performance computing to design fingernail-sized photonic chips that can plug into optical fibre networks to make our internet faster, or plug into medical diagnostic tools to analyse how cancer cells spread in real time. They are now using RACE to conduct research that was almost impossible with standard computing power.

“Direct access to RACE means that when designing and simulating brain-like chips or creating a chip which could break the record for the world’s fastest internet, the team can run multiple processes at once with computing capability that expands and scales as needed,” Nguyen said.

“RACE has provided our team with on-demand computing power anywhere, anytime to simulate our photonic chips at 10 times faster than was previously possible.”

This work opens the door to new opportunities including the design of chips that could make our internet faster, helping drones to more accurately inspect railway infrastructure, and building handheld devices to detect ovarian cancer more accurately.

RACE is supported by the Victorian Government under the Victorian Higher Education State Investment Fund and represents a step change in how universities and industries access high-performance computing capabilities for advanced data processing. RACE is now officially open for industry partners with an interest in driving digital innovation in research and education.

More information is available on the RMIT website.

Image credit: iStock.com/gorodenkoff