Compute Cluster

You would expect any Faculty of Engineering at any University around the world to be full of computers and gadgets, because that’s what engineering is about. Isn’t it? Well this page lists the technical information about the Engineering Compute Cluster at UNSW Faculty of Engineering – in case you need that level of detail.

The UNSW Faculty of Engineering boasts a medium-sized computational cluster for use by researchers in Engineering. Based on blade servers (HP BL685c G7 blade) the cluster currently consists of:

  • 2944 compute core (across 56 compute servers)
  • 5.8 TB Memory
  • 100 TB (usable) shared disk storage

Of the 56 compute servers, 40 consist of:

  • 48 compute cores
  • 96 GB memory
  • 144 GB high speed local disk

The other 16 compute servers consist of:

  • 64 compute cores
  • 128 GB memory
  • 144 GB high speed local disk

The cluster is fully interconnected with dual 10 Gb/sec data paths. The cluster nodes are Linux-based, running the standard Rocks clustering platform.

There are UNSW site licences available for a number of software packages which may be used on the cluster – including Matlab and Ansys. The cluster is also a viable platform for researchers writing their own software.

Hosted at the UNSW Data Centre at the Randwick Campus, a location with fully redundant networking, power and air-conditioning, the cluster is constantly monitored by UNSW systems and staff.

To request access to use the cluster, fill out the online application form. You can also access a documentation wiki for using the cluster.

Other available resources

Researchers running single-threaded jobs with very large memory requirements should consider using the large shared-memory mclaren system at AC3.

Researchers with very large parallel compute requirements should consider using the vayu cluster at NCI.

Both are available to UNSW researchers.  Access to either system can be made through Intersect.

Researchers wishing to run simultaneous discrete computations should also look at the condor pool running in CSE, by which the spare computational capacity of several hundred laboratory computers is made available.  Documentation on the condor pool is available here.  For access to the condor pool, contact System Support at CSE.

Share this