Plato is a Linux-based, heterogeneous, high-performance computing (HPC) cluster designed to support University research projects. It is also used for training highly-qualified personnel and students on advanced research computing techniques.
Plato has a total of 120 compute nodes with an aggregate 2 000 CPU cores and 7.4 TB RAM, yielding a theoretical 64 TFLOPS. There are 94 general-purpose nodes, 2 GPU nodes, and 2 large-memory nodes. In addition, 22 nodes were contributed by specific groups, and are reserved for them. Plato runs RedHat Enterprise Linux (REHL) 7.3 and uses the SLURM scheduler.
Plato is managed by ICT’s Advanced Research Computing (ARC) team. Although it is not a Compute Canada machine, Plato works in a similar manner. For instance, through CVMFS, Plato provides University researchers with access to Compute Canada’s scientific software stack.
Plato cannot provide all the computing power required for all research projects at the University. Researchers with important computing needs are invited to consider Plato as a stepping stone to test their projects locally before requesting access to larger WestGrid and Compute Canada systems.