Supercomputing is at the heart of iVEC’s function. From the perspective of the end-user, a key source of capabilities is the iVEC Supercomputing Team. The Supercomputing Team encourages the uptake of iVEC supercomputers by the scientific community in Western Australia and beyond.

iVEC has complementary supercomputers to reflect different workflows and user requirements. The iVEC Supercomputing Team assists users across these resources.

• Magnus. A Cray XC40 Supercomputer, for highly parallel distributed programs.
• Zythos. A SGI UV2000 shared memory computer, with 6TB addressable memory.
• Galaxy. A Cray XC40 Supercomputer, used exclusively for radio astronomy.
• Epic. A large Hewlett Packard cluster, for parallel distributed programs.
• Fornax. A SGI distributed memory cluster, featuring Nvidia accelerators.

Access to the above systems is obtained by submitting an online application for a project. Applications for new users can be made at any time of year and are awarded compute time under the iVEC Director’s Share of the systems. iVEC issues calls for applications for production projects periodically, and these are awarded compute time in a competitive merit process.

Training courses on using the iVEC supercomputers are available via the iVEC training page.

The Supercomputing Team undertakes various activities, including:

• Assisting new researchers onto supercomputers.
• Engaging researchers, communities and disciplines with supercomputing and data-intensive computing.
• Assisting researchers to streamline and automate their workflows that incorporate supercomputing, data storage and visualisation (with the iVEC Data Team and iVEC Visualisation Team).
• Developing and delivering training material, focused on capability-scale and accelerator computing.
• Embedding iVEC supercomputing specialists into key research groups for major projects to stimulate capability and to grow supercomputing activities.
• Promoting success stories through (for example) conferences, presentations, publications, and networking.

Supercomputing is a term that covers a wide range of activities, which are captured by the diverse skill set of the Supercomputing Team. Highlights include:

• Domain-specific tertiary-level qualifications – for example, computer science; computational science; geosciences; physics; chemistry; mathematics; and engineering.
• Parallel programming – distributed-memory programming (for example, MPI) and shared-variable programming (for example, OpenMP).
• High-performance computing – hybridisation; performance analysis; optimisation and scaling.
• High-throughput computing – batch processing; virtualisation; job management.
• Accelerator programming.
• Software engineering – for example, application integration; porting; parallel debugging and software testing.
• Scientific computing – numerical libraries; application-specific packages; academic writing.
• Data management – database design; data-grid and distributed computing; metadata / mark-up.
• Training and teaching.
• (Software) project management.