Unfortunately, the CHPC is subject to load-shedding (under the City of Cape Town) when Eskom has reduced generation.

While the CHPC has a 1000 kVA generator this is not sufficient to power the entire cluster during blackouts and only 10 of the compute racks are available.

Please plan accordingly

There are steps you can take to make more effective use of the cluster when it is operating at reduced capacity. Many of these are also useful outside of loadshedding events as well.

1. Please check whether each of your jobs is very necessary.

The following will allow your job to start running sooner and remain queued for a shorter time:

2. Choose as short a walltime as appropriate such that your job will have passed and written checkpoint files, so that you can then submit subsequent jobs to continue your calculation. You can also adjust (increase) the frequency of checkpoint file outputs when possible. You can also choose the waltimes according to your working hours and schedule.

3. Select fewer nodes for your job. Be aware of how well your specific jobs scale and what the sweetspot is. For example your job has speed S for 1 node, speed 1.8S for 2 nodes, 3S for 4 nodes, 4S for 6 nodes, 3.8S for 8 nodes. When there are a surplus of nodes on the cluster, one might choose to use 6 nodes (8 nodes would be a very poor choice as it's inefficient and wasteful). During loadshedding when the cluster is oversubscribed, choosing 2 nodes (or even 1 node) would be a good idea.

4. Make sure to be using checkpoints as well as possible.

5. Testing of a new type of job/calculation - either use an interactive job/node or a very very short job (maybe 15 mins). This way you will know your job will work, and then you can submit your actual loner job thereafter. You can thus avoid waiting hours or even days to find your job failed, due to perhaps a small error in your input file.