CCF (Core Computational Facility) @ UQ run by ITS / SMP
[Home] [why HPC?] [UQ Resources] [getafix] [dogmatix] [asterix] [ghost] [contacts/help]
howto: [data] [Slurm] [OpenMP/MKL] [Nvidia GPU] [Intel Phi] [MPI] [matlab] [mathematica] [FAQ]
ghost partition
Ghost compute nodes are now part of the dogmatix cluster in a separate partition called ghost in the slurm queuing system.
Ghost jobs should be submitted to the SLURM queue manager by adding -p ghost to your submission command or adding this to your submission script:
#SBATCH --partition ghost
See the Nvidia GPU howto page for help submitting GPU jobs but note that the older GPUs on ghost require cuda-6 or earlier.
ghost information
Holger Baumgardt's Future Fellowship GPU Cluster.
- Purchase starting in 2010 $117K ex-GST
- One whole Rack... Xenon Nitro T5 with 34 Tesla C1060 GPUs (30 still operational at 30 May 2017).
-
- 10 nodes 4U... 120 cores Intel X5650 @ 2.66GHz (2010)
- 4 nodes have 4 x Nvidia Tesla C1060
- 2 nodes have 3 x Nvidia Tesla C1060
- 4 nodes have 2 x Nvidia Tesla C1060
- Located in Prentice Data Centre.
System status (last updated 17th May 2017)
This system is out of warranty.
This page last updated 13th March 2018. [Contacts/help]