[CSRVM] [CSRGPU] [Current GPU VMs] [The Rules] [Request a GPU VM] The Rules 1. VM Ownership - All requests MUST come from a CS faculty member as that person will be responsible for the VM. 2. VM Lifecycle - All VMs must have an end date within the Maximum extension limits. Currently you can renew your VM up to 6 months. 3. Resource hoarding - Hoarding or squatting on resources will not be tolerated. You should only ask for what your current project requires. Resources that go unused will be recycled. Please only ask for what you need. 4. Cores & RAM - CPU cores and RAM are fixed and divided evenly across the allocated GPU VM slots. More cores or RAM may be requested after the initial setup. Additional resources may be granted depending on need and availability. 5. One is enough - All VMs on this system are limited to a single GPU. If you need more consider purchasing your own equipment. 6. L40S GPUs - Existing VMs can request the use of one of our L40S GPUs if available. The L40S GPUs are available for a SHORT time period of Two(2) Months MAX. After that you will be cycled back to a smaller GPU. This is to avoid monopolizing these high demand resources for long periods of time. Also, I'm going to ask that labs limit themselves to only one(1) L40S request at a time. To request an L40S send an email to techstaff with the name of your VM. [This rule is subject to change with demand.] 7. Expectations - When you are given a VM you are expected to know how to administrate it. We give you a bare bones install and you are responsible for what happens to it after that. Including but not limited to: security, user management, backups, package deployment, tool generation, etc. 8. AUP Violations - If the VM becomes a security risk we will simply shut it off in accordance with the VT Acceptable Use Policy. We will, however, work with you to try to recover your data and set it up in a new VM.