Advanced Computing

Do you need to collect and store large amounts of data? Is data analysis or processing taking far too long on your computer? Do you need to use specialised hardware or software for your research?

We work directly with UON staff and students to help solve problems like these through the use of advanced technology. We can provide research groups with:

  • Data storage for large research data sets
  • Access to and help using High Performance Computing and Virtual Machines to improve data analysis
  • A secure and reliable hosting environment for research servers and software
  • Help solving difficult IT-related research problems

To contact us about a particular problem you think could be resolved using advanced computing, email arcs@newcastle.edu.au.

High Performance Computing (HPC) is the use of large-scale computers and parallel processing techniques for solving computational problems. Aggregating computing power allows analysis of very large data sets and solving of complex, time-consuming problems in areas such as science, engineering, health and medicine, or business and marketing. The terms High Performance Computing and Supercomputing are used interchangeably. For more information on HPC, see the Introduction to High Performance Computing knowledge base article.

UON researchers are able to access a number of HPC systems. Some are directly supported by the University, others are accessible through competitive, merit-based schemes.

If you are new to HPC, we highly recommend that you contact us for assistance getting started.

UON Supported Facilities

Research Compute Grid (RCG)

A UON facility managed by ARCS with over 3648 usable cores are available, with 700 TB of usable shared storage space running on a Lustre filesystem. Technical specifications are:

  • 16 x 512GB nodes powered by 2.7 GHz Intel Xeon Gold (E5-6150) processor
  • 32 x 128GB nodes powered by 2.3 GHz Intel Xeon Haswell (E5-2698 v3) processor
  • 32 x 512GB nodes powered by 2.6 GHz Intel Xeon Broadwell (E5-2697A v4) processor
  • 16 x 256GB nodes powered by 2.8 GHz AMD Opteron (6386) processor
  • 1 x 192GB Nvidia GPU node (4 x single V100)
  • 6 x 128GB Nvidia GPU nodes (4 x single K20, 2 x dual K80)
How To Access

Request access through the UON High Performance Computing Cluster request webform in serviceUON.

More Information

See the Introduction to Using the Unix Grid knowledge base article.

Raijin

Raijin is Australia’s peak supercomputing facility, managed by the National Computational Infrastructure (NCI). Raijin includes more than 84,656 cores with 40 PBytes of usable disk space delivering over 1.37 PFlops peak performance. The NCI also maintains an extensive library of pre-installed software packages, many of which can be accessed at no cost. Users are able to request computing time on Raijin through Intersect, which has a partner share of this facility. Technical specifications are:

  • 4,416 x compute nodes (32GB to 1TB memory configurations) powered by 2.6 GHz Intel Xeon Sandy Bridge or Broadwell processors
  • 30 x NVIDIA K80 GPU nodes (each node contains 4 x NVIDIA K80 GPUs)
  • 2 x NVIDIA P100 GPU nodes (each node contains 4 x NVIDIA P100 GPUs)
  • 32 x 192GB nodes powered by the 1.3 GHz Intel Xeon Phi (Knights Landing) processor
How To Access

Request access through the UON High Performance Computing Cluster request webform in serviceUON.

More Information

See https://intersect.org.au/time and the Raijin User Guide.

Other Facilities

Pawsey Supercomputing Centre

The Pawsey Supercomputing Centre operates multiple HPC systems, including Magnus, which consists of 35,712 CPU cores for computational research, and Galaxy, which consists of 9,440 CPU cores and is dedicated to radio-astronomy. Researchers are able to apply for access to these systems through an annual National Computational Merit Based Allocation Scheme (NCMAS).

MASSIVE (Multi-modal Australian ScienceS Imaging and Visualisation Environment)

MASSIVE is a specialised HPC facility for imaging and visualisation. It encompasses three systems (M1, M2 and M3), each supporting different configurations of CPUs, memory and GPUs. MASSIVE is best suited to computational imaging projects. Researchers are able to apply through an annual National Computational Merit Based Allocation Scheme (NCMAS).

A Virtual Machine (VM) is a computer that appears to the user as a real physical computer, however the hardware is actually simulated or virtual. Usually, a single powerful physical server runs several VMs at a time, distributing the workload onto its physical CPUs and memory as needed. Each virtual machine is a fully simulated computer which can have its own operating system and software running.

So why is this helpful in research?

  • Software can be installed onto a VM just like any other computer. This means they can complete long-running tasks, e.g. complex processing and data analysis, without slowing down the computer for other daily tasks.
  • It is possible to provide many people access to one VM, which enables your collaborators and students access to both the data and analysis software in one location.
  • VMs can be more powerful than desktop or laptop computers, which may speed up or make certain types of analysis possible.
  • Remote access to a VM and your research data allows you to run your analysis and access the results from anywhere.
  • VMs can be powered on 24/7, making them perfect for running websites, web applications and databases.

UON supports two VM platforms for research use. If you think you need a VM in your research, we recommend that you contact us to discuss your needs.

Virtual Workstations

ARCS can provide VMs to researchers who want to operate their own software from a secure, centrally managed data centre. The ongoing maintenance, patching, and support of the virtual machine is undertaken by IT professionals within our data centre, reducing the risk of interrupting long processing jobs, hardware failure, data loss or virus infection.

How To Access

To request a virtual workstation use the virtual workstation request webform in serviceUON.

NeCTAR

The National eResearch Collaboration Tools and Resources project (NeCTAR) is a Commonwealth Government funded initiative available to Australian based researchers.  It is a self-service system allowing researchers to start small-medium Unix/Linux VMs for testing purposes at any time under a personal trial project. After the trial period, users can request access to larger VMs and more time through an online form on the NeCTAR website. As there is a cost to UON, project requests will be directed to IT Services to approve.

How To Access

Login to the NeCTAR dashboard using your UON credentials (number plate). After login you will be able to start VMs for trial purposes.

More Information

For self-paced training see training.nectar.org.au. For advice in selecting the most appropriate VM, or assistance with an allocation contact NeCTAR Research Cloud Support.

IT Services provide a secure, stable and recoverable environment where research servers can be physically or virtually hosted, reducing the risk of interrupting long processing jobs, hardware failure, data loss or virus infection. The ongoing maintenance, patching, and support of the virtual machine is undertaken by IT professionals within our data centre.

The main benefits to researchers are:

  • Physical hardware is no longer required to be held in the researcher’s office or lab
  • Regular server patching by IT Services
  • Backups are regular and automated
  • Data is secure and recoverable

How To Access

Request physical or virtual hosting services through the hosting webform in serviceUON.