GPUs in use — last hour busy GPUs (util >5% or mem >1 GB)
Evaluations
| run | cells | success | purpose | settings | updated |
|---|
| cell | task / case | skill | layout | result | steps | turns | time | why |
|---|
Slurm queue
evaluation runningqueuedGPU freeGPU running a cellGPU held by a resident server
| job | user | name | state | GPUs | CPUs | nodes | elapsed | limit | reason / start | purpose (job comment — every evaluation must say why) |
|---|
GPU utilGPU memoryCPU load / RAM / NVMe
User access — add someone's public key to the nodes you pick (non-root account, own dir on /opt/dlami/nvme)
Accounts on the cluster (click a row to select its nodes)