Getting Started

Before proceeding here, you should follow the instructions on logging on to your account.

Account Information

When you are connected to a Baskerville login node, run my_baskerville. This will give you information on the projects and QoS that you have access to.

Each project has access to project storage and this is of the form /bask/projects/p/project, where p is the first letter of the project code and project is the project code.

The project code is also the code you supply to Slurm for --account=.

Further information about your account is available in Baskerville Admin

Sample Slurm Job Script

A sample Slurm job script is:

#!/bin/bash
#SBATCH --qos _qos_
#SBATCH --account _account_
#SBATCH --time 1:0:0
#SBATCH --ntasks 4
#SBATCH --gres gpu:1

module purge
module load baskerville

# your commands

From the account information above, replace:

• _account_ with the project code
• _qos_ with the QoS to use

Multiple Projects and QoSes

If you are a member of multiple projects you should specify the correct project code. If you have access to multiple QoS then you need to specify a project code that corresponds to the correct QoS.

The jobs section contains further information on running jobs on Baskerville.

Acknowledging the use of Baskerville in your Research

A condition of access to Baskerville is that the service is acknowledged in the research output generated through the use of Baskerville. We have provided a suggested form for this acknowledgment.