Getting access to Viking is a simple two stage process. First research groups/projects apply for a Viking Project Account. The Project Account is assigned a Project Code. The PI passes this project code onto the other members of the project and they use this to apply for a user account. The PI is notified each time a user account is assigned to the project.
In order to use the Viking cluster research groups/projects must apply for a Viking Project Account. The project accounts will be used to monitor the usage of Viking and attribute its use by researchers across the University.
We ask that the PI who applies for the account is a permanent member of staff, as this will allow us to have a permanent contact for project reporting purposes. You can also nominate a deputy who can deal with any project enquiries on your behalf.
The PI of the project should complete this project application form.
When completed a Viking Project Code will be supplied to you. The Project PI should then distribute the project code to all research team members before they apply for user accounts.
A PI may request more than one Viking Project Code for multiple projects.
The Project Application Form can be found here.
Each user must apply for an individual user account. Your PI, or supervisor, will supply you with a Viking Project Code which you will need to complete this application for an account.
The User Application Form can be found here.
The login servers are to be used for program/application development, testing and job submission only. You can run your programs on the login nodes but not for long periods of time (> 10 minutes) - this is what the cluster is for!
In order to have reliable job execution and the best performance, you have to run your jobs from the high performance filestore:
Within your home directory is a sub-directory called "scratch". Place all your work in here.
You will initially have 3TB quota if you need more please speak to us.
This directory is accessible from all nodes.
This is not optional, running your jobs from your home directory will result in your jobs being killed when started by the scheduler. (N.B your home directory may not look exactly like this but you should have a folder called scratch)
[abs4@login2(viking) ~]$ ls bin Downloads Music Public Templates work Desktop easyconfigs orig-slurm-configs scratch tmp Documents intel Pictures src Videos [abs4@login2(viking) ~]$ cd scratch/ [abs4@login2(viking) scratch]$ pwd /users/abs4/scratch [abs4@login2(viking) scratch]$ ls Amber DynamicConvergenceMatrix gnuplot intel Tensorflow VASP AmberBuild gaussian gpu slurm Test VASPBuid [abs4@login2(viking) scratch]$ df -h . Filesystem Size Used Avail Use% Mounted on 10.12.3.1@o2ib:10.12.3.2@o2ib:/lustre 2.6P 47T 2.4P 2% /mnt/lustre [abs4@login2(viking) scratch]$ ***************************** *** DO ALL YORK WORK HERE *** *****************************
Please do all you work in your scratch directory.
If you still get this error and you are sure you are submitting your job from the scratch directory, you may have edited your file on a Windows PC and transferred it to scratch
Editing text files on Windows has the bad effect of inserting additional characters into the file. After you have transferred the file to the login server, use the following command to remove these characters:
We do not recommend that you edit your files on a Windows PC and transfer them to the cluster.