School of GeoSciences

Linux Servers

The school has a central compute server running Linux you can access to process data.

 

Connecting

If you are using Linux you can just open a terminal and ssh to the server detailed below. If you submit a job via ssh, you should use screen so the job remains running if you disconnect.

If you are using Windows (or working remotely using Linux/Mac) you should use NX

 

Usage

Our Linux server is accessible to anyone in GeoSciences. It is possible for you to consume all of the server's resources, so please be considerate when determing how many cores and memory you plan to use.

For big jobs, you should use the University's Eddie cluster (see http://www.ed.ac.uk/information-services/research-support/research-computing/ecdf/high-performance-computing)

 

 

Using Local Scratch

Local storage is far quicker than network storage. If you are finding your job is running slower than expected, use the local scratch for your input and output.

To use local scratch, create a directory in /scratch/local/ matching your username. Once your job has finished, copy the output to your personal storage (e.g. M: Drive/home directory or DataStore) and delete your data from local scratch.

Local scratch is a shared resource so please tidy up after yourself and be considerate with amount you store.

 

Burn

Address

burn.geos.ed.ac.uk

OS Scientific Linux 7
Hardware Dell PowerEdge R820
CPU 64 cores
RAM 256GB 
Local Scratch 1.3TB