Kaya (Hello!) comes from the Noongar language spoken by the indigenous Noongar people who live in the south-west of Western Australia. The UWA HPC system is named Kaya, and it provides a way to introduce researchers at UWA to HPC. The Kaya system has two main purposes, the first is to be as a 'launch pad' to enable researchers to develop their computational skills and workflows with a view to successfully migrating to other systems such as Pawsey, NCI (National Computational Infrastructure), or to the Nimbus/Nectar research clouds. The second is to provide a platform to run work that is too big for a desktop system, but not suitable for a Pawsey or NCMAS allocation.

The Kaya system has a mix of Central Processing Unit (CPU) and Graphics Processing Unit (GPU) compute nodes, including some large memory machines, high-performance solid-state storage, and high-speed networking interconnects. The system uses Simple Linux Utility for Resource Management (SLURM) to ensure fair access for all users. Standard software includes Message Passing Interface (MPI) libraries, C/C++/Fortran compilers and debugging tools. We can provide Jupyter Notebooks, MATLAB, and RStudio through Open OnDemand. The HPC team can assist with more specialised requirements, including commercial software.

There is no cost associated with the use Kaya. It is free to use for all UWA staff and students.

When applying for access to Kaya you will need to provide:

• A Principal Investigator (PI), normally a senior researcher or member of the academic staff.
• A project description and some initial information about how long the project is expected to run and how many users are likely to need to work on the project.
• A GitHub repository and link.
• ORCIDs for project members.
• An Institutional Research Data Store (IRDS) share for long term storage of project data.

The Project will be created and then accounts for PI and project members are added to the project. The PI has the ultimate responsibility for the data at completion of the project when all the associated project data must be removed either to the UWA Library archive storage or UWA IRDS in-line with the projects data governance processes.

UWA staff and students can access the Kaya user documentation here.

The University of Western Australia is one of the founding partners of the Pawsey Supercomputing Research Centre. The Pawsey Supercomputing Research Centre is a world-class high-performance computing facility accelerating scientific discoveries for Australia’s researchers. Located in Perth, Pawsey brings together the critical infrastructure, expert staff, sector knowledge and focus needed to solve pressing problems in every field.

The University of Western Australia is a Pawsey Partner and together with Australia’s national science agency – Commonwealth Scientific and Industrial Research Organization (CSIRO) and the other three Western Australia public universities. The UWA DVCR office funds the Pawsey Partnership that allows UWA staff and students to apply for Pawsey resources at no cost through the Pawsey Partner allocation schemes on an annual basis. The National Computational Merit Allocation Scheme is also a way that UWA researchers and students can access the Australian Flagship systems. This gives our researchers the opportunity to scale up their projects and the grand challenges in their research domains. Access to the Nimbus research cloud, managed storage and visualisation services and expertise is also available, and it can be part of your supercomputing allocation or accessed independently.

Home of Setonix, the most powerful supercomputer in the southern hemisphere, Pawsey supports science of national and international importance. Setonix gives users access to both CPU and GPU resources. Pawsey has carefully considered sustainability in building this system and it is the fourth greenest supercomputer in the GREEN500 lists.

The Centre is also known for the training and education opportunities provided to current and potential new Pawsey researchers. Extensive training is scheduled and delivered throughout the year. In addition, several training videos are produced and shared online for accessibility purposes. Researchers can also benefit from Pawsey’s summer internship program, where undergraduate students develop skills in computational science while immersing themselves in advanced computing research projects for ten weeks between November and February every year.

More information can be found on the Pawsey website. 

Access to the Pawsey supercomputer is through the NCMAS or the Pawsey Partner scheme. The call for applications is in September each year and allocations are made in December. If you are a UWA researcher and would like to access the Pawsey supercomputer, please contact the HPC team before September for assistance with the application process.

Project applications will need to:

• estimate the number of jobs or simulations.
• estimate the average number of cores/GPUs per jobs.
• the total number of cores hours for the entire project. 

All proposed projects that are planning to use the Pawsey Supercomputing Research Centre or National Compute Infrastructure facilities, will need to include a technical assessment as part of their application. The projects need to ensure that it can make effective use of the HPC capabilities and capacities as well and define the quality of research project itself.

Nimbus and Nectar

Research cloud services provide Australia’s research community with fast, interactive, self-service access to large-scale computing infrastructure, software, and data, and are a powerful platform for collaboration. The research clouds are ideal solutions for a researcher who wants a no-cost, on-demand compute resource at their disposal. These resources allow for doing testing and development work to support HPC research projects with great flexibility, including:

• Root access to your Virtual Machine (VM).
• Complete control over your software environment.
• The ability to control access with collaborators.
• Burst out to multiple VM instances to run many simulations at a time.
• Run interactively with no queues and no wall times.

The UWA HPC team has extensive expertise in supporting research projects on cloud services with common tools such as RStudio and Jupyter Notebooks.

The UWA HPC team can help find the best solution to your research group’s needs which may include bespoke services. We can provide UWA researchers with an avenue to buy into a centralized dedicated service with specialised computer resources for needs that cannot be met by using Kaya or national facilities. This may occur if you have large memory (>1TB) requirements or sensitive data.

These dedicated resources are integrated into the Kaya system which is centrally managed by the UWA HPC team using the Kaya management system. The University Information Technology (IT) HPC systems are located off-campus in a commercial data centre and University IT funds the associated operational costs (power, cooling, networking, and rack-space). In this case, researchers do not need to manage user accounts, cyber security, or networking for their resources, so they can focus on their research and not maintaining a local server.

Researchers may be required to upgrade their current standard primary Computer Provisioning Scheme (CPS) device with certain components or purchase a secondary non-standard laptop/workstation for their computational or visualisation workloads. The HPC team can help with this by finding the best resource for your research needs.

Software Carpentry provides basic courses for computing on their website.

Training videos and courses are also provided by Pawsey Supercomputing Centre and can be found at their website. 

We provide several training courses for UWA researchers to help them use HPC services for their projects. If you would like a dedicated training course to be run for your group or research area, please contact the UWA HPC Team to discuss this.

The UWA HPC team can be contacted by emailing [email protected] or using the UWA IT Servicenow Portal.