27th April 2021
Giant compute grants to generate great Australian science
NCI Australia has today announced the recipients of the 2021 Australasian Leadership Computing Grants (ALCG) thanks to Australian Government support through the National Collaborative Research Infrastructure Strategy (NCRIS).
Australian researchers from a range of fields will use substantial grants of computing time to work on some of the most complex problems facing science today. Success in the highly-competitive ALCG process facilitates five teams accelerating their research and producing unprecedented results at incredible resolution.
NCI Director Professor Sean Smith says the ALCG once again identified meritorious research projects with demonstrated capacity to use Gadi effectively at scale. He says that the successful projects reflect the world-class, high-impact research that NCI enables.
“The Australian research community has once again demonstrated the high demand for high-performance computing across a wide range of fields. The high-quality applications received in this round build on the growing depth and breadth of compute-driven investigations and are a testament to the tremendous scientific talent in Australia,” said Professor Smith.
“Gadi, Australia's fastest supercomputer, will continue to boost Australian research with the 150 million units of computing time awarded as part of the Australasian Leadership Computing Grants.”
Five projects have been awarded a total of almost 150 million units of compute time, which is equivalent to 20,000 years of constant calculations on one single computer.
The five recipients of the 2021 Australasian Leadership Computing Grants are:
Professor Ben Corry, The Australian National University
Professor Alan E Mark, The University of Queensland
Dr Bernhard Müller, Monash University
Associate Professor Ekaterina Pas, Monash University
Professor Julio Soria, Monash University
Professor Ben Corry, Group Leader of the Corry Group at The Australian National University Research School of Biology, will attempt to get ahead of the next potential pandemic.
"Nobody wants to see another pandemic like COVID-19. But how do we know if any of the numerous coronaviruses currently circulating in animals such as bats and rodents have the potential to cause another outbreak?"
"Over the last year we learnt a lot about what makes SARS-CoV-2, the virus causing COVID-19, so transmissible and deadly. We want to apply this knowledge to assess the danger of a much wider variety of coronaviruses. But, analysing how the virus invades human cells at a molecular level requires enormous computer simulations that run for many weeks.
"This allocation makes it feasible to apply this analysis to a huge number of different coronaviruses, rather than just one. Such a large grant turns a specialised method into a predictive tool. We hope that this can be used to help monitor dangerous coronaviruses so that they can be isolated before they have the chance to spread to humans," Professor Corry said.
Professor Alan Mark, from the University of Queensland's School of Chemistry and Molecular Biosciences, said "The Australasian Leadership Computing Grant will help develop efficient, lightweight and low cost organic semiconductor devices for use in lighting and energy generation.
It will allow us to examine how these devices form at an atomic level. We will be able to simulate devices of direct commercial interest, not simplified model systems."
Dr Bernhard Müller, ARC Future Fellow at Monash University's School of Physics and Astronomy said "Astronomical observations tell us that most massive stars end their life in a supernova explosion, but to figure out how these explosions work we rely on simulations to ‘look inside the star.’
"Constructing accurate supernova models is tough, and the scale of the allocation finally allows us to perform the high-resolution simulations that we think are needed to achieve the necessary degree of realism. This grant will help us address many open question surrounding supernova explosions, neutron stars, and the origin of many chemical elements that we believe are made in massive stars," said Dr Müller.
Associate Professor Ekaterina Pas from Monash University's School of Chemistry says her team's work will be accelerated by the ALCG allocation.
"We are hoping to predict lattice energies of ionic organic salts from first principle for the first time. This unprecedented level of accuracy has not been available to us until our development of a highly accurate computational chemistry method. The Barca group were very successful in translating our group's theory into a very efficient GPU code. Without this code, the proposed calculations would not be feasible.
"The allocation on Gadi will allow us to run the GPU-based code on very large systems (up to 4500 atoms) and uncover the relationship between the electronic structure and bulk properties of ionic materials for the first time.
"These will be the largest quantum chemical calculations performed in Australia so far. The outcomes of the grant will have a significant impact on two branches of chemistry - theoretical and physical chemistry - by showcasing our ability to predict bulk properties of complex materials.
"Equally important, we will determine the importance of hydrogen bonding in ionic materials to exhibit desirable properties for their use as phase change materials for heating and cooling buildings. The latter will bring us closer to reaching the 2050 target of zero emissions for Australia."
Professor Julio Soria's work in the Department of Mechanical and Aerospace Engineering at Monash University will contribute to turbulent bushfire flow research.
"A high-fidelity direct numerical simulation of a turbulent thermal boundary layer with distributed high energy heat sources is an analogy for high-fidelity simulations of bushfires. This enables the detailed quantitative study of the interaction between buoyant and inertial forces, the spatio-temporal structure and transport of energy, mass and momentum within turbulent bushfire flows as a function of the fuel source parameters.
"This allocation will permit a unique DNS of a turbulent thermal boundary layer providing unprecedented quantitative information which will inform important bushfire parameters and their dependence like the spreading rate.
"This is the single largest grant provided for a DNS of a turbulent thermal boundary layer flow in Australia and will yield the first ever DNS of a turbulent thermal boundary layer flow with modelled fuel sources," Professor Soria said.
Research will take place on NCI’s Gadi supercomputer, commissioned in early 2020, and funded under the Australian Government's National Collaborative Research Infrastructure Strategy (NCRIS). NCI Australia brings the Australian Government and the Australian research sector together through a collaboration involving CSIRO, The Australian National University, Bureau of Meteorology, Geoscience Australia, universities, industry and the Australian Research Council.
Read more about the recipients and their research.
Research Project Details
Predicting the next coronavirus outbreak
Professor Ben Corry
Mr Josiah Bones
Dr Amanda Buyan
Mr Matthew Witney
The Australian National University
COVID-19 has resulted in millions of deaths and created enormous social and economic upheaval. However, this is not the only coronavirus circulating in the world as there are a large number of coronaviruses in animal populations; and so far we cannot judge which of these will be dangerous to humans. This project will fill this knowledge gap by using computer modelling of the coronavirus spike proteins to determine which circulating viruses have the potential to generate future outbreaks, helping to avoid future pandemics and allowing pre-development of vaccines.
Understanding organic semiconductor morphology at an atomic level: Simulating the formation of realistic devices.
Professor Alan E Mark
Professor Paul L. Burn
Dr Martin Stroet
Dr Paul Shaw
Ms Audrey Sanzogni
Mr Abhay Sharma
The University of Queensland
The aim is to understand how different manufacturing techniques affect the properties of certain semiconductors used in OLEDs, photovoltaic devices and sensors. The researchers will use computational techniques to replicate the physical processes that occur during manufacturing and thereby predict the effect of different production protocols on the properties of the final device. Because the active layers in these devices are so extremely thin the way the molecules are deposited has a large effect on how the molecules pack. This research has the potential to transform the way that thin-film devices are developed. Thin-film devices such as these can be used for efficient lighting systems, photovoltaic solar energy production and high-tech device innovation.
High-resolution Core-Collapse Supernova Simulations
Dr Bernhard Müller - Monash University
Dr Jade Powell - Swinburne University
Professor Alexander Heger - Monash University
Core-collapse supernovae, the spectacular explosions of massive stars, are one of the grand challenges in computational astrophysics, a true multi-physics problem that involves multi-dimensional turbulent fluid flow, neutrino radiation transport, extremely strong magnetic fields, general relativity, and nuclear physics. This project aims to conduct high-resolution 3D simulations to investigate the effects of turbulence, magnetic fields, and general relativity on supernova formation. With models that set new standards in terms of numerical accuracy and physical completeness, the researchers seek to find a solution to the supernova “energy problem” and to reliably predict the properties of the leftover neutron star, the gravitational waves emitted from the supernova core, and the chemical elements made by nuclear fusion during the explosion.
Design of Phase Change Materials of the Future
Associate Professor Ekaterina Pas - Monash University
Dr Giuseppe Maria Junior Barca – The Australian National University
This project aims to design improved Phase Change Materials (PCMs) for use in heat storage and conversion in buildings and industry. These PCMs store and release heat as they melt and freeze. Carefully designed PCMs with very precise melting and freezing points have significant commercial value and environmental benefits as they reduce energy use for heating and cooling buildings, thereby reducing emissions from the housing sector. The research team will use world-leading software running on NCI’s entire set of Graphic Processing Units to efficiently investigate the properties of various promising PCMs, certain organic ionic salts.
High-fidelity direct numerical simulation of high Reynolds number turbulent thermal boundary layer flow with distributed high energy heat sources: An analogy for high-fidelity simulations of bushfires
Professor Julio Soria
Dr Shahram Karami
Dr Callum Atkinson
This project aims to simulate the turbulent boundary layer of bushfires at high resolution, taking into account the interactions between different kinds of fuel sources and the atmosphere. Direct Numerical Simulations will be used to investigate flows of energy, mass and momentum within turbulent bushfire flows. Bushfires are complex to measure and predict. This research should help to provide detailed, physically accurate information about the processes occurring in and around a bushfire.
Media enquiries: Aidan Muirhead aidan.muirhead [@] anu.edu.au