Compute and Storage
Compute
The ML4AI lab has the following compute VMs:
| VM Name | CPU | RAM | GPUs |
| kraken | AMD EPYC 7662 64-Core Processor (2.0 GHz) | 720 GB | 2x NVIDIA A100 (40 GB) |
| leviathan | AMD EPYC 7763 64-Core Processor (2.45 GHz) | 720 GB | 6x NVIDIA RTX A6000 |
| carp | AMD EPYC-Rome Processor | 95 GB | 1x NVIDIA GeForce RTX 3090 |
| orca | AMD EPYC 9474f, 48-core, 3.60 GHz, 256MB cache | 1.5 TB (tentative) | 2x NVIDIA H100 NVL (94GB hbm3, PCIE 5.0 x16) |
TODO:
- Add information about venti
Storage
The lab has a 20 TB NFS share mounted at /media/mule on the kraken, leviathan, and orca VMs.
There is a 90 TB NFS share mounted at /media/snail-ssd on the orca VM.
TODO:
- Add information about timelord
TODO:
- Add information about other legacy SISTA systems that are still operational.
Backup
- VMs running on InfoSci servers (e.g., eel, starfish) are backed up every 6 hours
- As of 2025-03-06, VMs running on lab servers (e.g., kraken, leviathan) are not backed up, but the plan is to include them in the backup system in the future.