Welcome to ARE (short for Australian Research Environment). ARE is a web-based graphical interface for performing your computational research. It combines the familiarity of your regular desktop/laptop with the power of NCI’s world-class research HPC capabilities. ARE gives you access to NCI’s Gadi supercomputer and data collections, all from a simple, graphical interface. ARE consists of a number of applications that support your research such as Virtual Desktop, JupyterLab, Terminal, etc. If you are coming from the OOD then we strongly encourage you to visit the Changes section to see what has changed.
ARE can be accessed at are.nci.org.au
NCI is pleased to announce that the highly requested Internet access capability in the ARE service (Australian Research Environment) is now operational. Jobs running in ARE will now be able to communicate directly with the Internet, facilitating more interactive usage.
ARE supports the full range of Gadi capabilities and high-performance access to the underlying /g/data filesystems. It provides users with a variety of specialised queues to access the GPU-equipped Visualisation nodes, Large memory nodes and a variety of CPU generations based on processing requirements. Read the ARE User Guide to find out more.
NCI projects currently using OOD should migrate their workflows to the ARE environment as soon as possible. Please follow the instructions for copying files from OOD to ARE to complete this migration.
Please note that ARE uses NCI Service Units as its basis for charging the jobs that are run. An allocation of SUs, acquired through merit schemes, partner share or other access mechanism, is required to use ARE.
Please get in touch with NCI at email@example.com if you have any questions.
Any app can be run on any of the queues documented at https://opus.nci.org.au/x/2ABiBQ by typing the queue name into the box on the form; the drop down contains the queues most likely to be appropriate for that app.
Attempting to run a GPU-specific app (eg the GPU-enabled VDI app) on a non-GPU queue will fail.
There is an additional queue called 'analysis' which is specifically for use with the GPU-enabled VDI app.
Most queues do not have outbound internet access; if your workload needs interactive internet access, you will need to select either the 'copyq' or 'analysis' queues. (If possible, it would be preferable to do any data transfer stages in separate copyq job).
The sections below cover more details for each app: