Usage Policy¶
important
The cluster is a shared resources, always be mindful of other users, deliberate attempt to circumvent the system to gain an advantage will incur negative consequences. We reserve the rights to pause, cancel any running jobs as well as temporarily suspending your access to maintain overall usability to the cluster and the best interests of all users on the cluster.
Running job or CPU intensive process on sc.stanford.edu (headnode) and scdt.stanford.edu (data-transfer) themselves is strictly prohitbited, we will kill such processes without warning.
No direct access (eg. SSH) to compute node unless otherwise arranged. You may SSH to the compute node where you have an active running jobs within SLURM.
If you need to run your job outside the queuing system and/or need to reserve a batch of machines for other purpose, please submit a request via action@cs.stanford.edu, we will review the situation and properly assign you a batch of machines for exlusive use.
Please be mindful of others. We practice all honor codes and we expect every user to be a good citizen. We also encourage collaboration and sharing, but written permission is required before using other group resources.