
Manage Slurm Workload Manager jobs across multiple clusters and inspect utilization from Raycast.
Host entry in ~/.ssh/config↵ on all clusters you wanna monitor. If a password or 2FA prompt is needed, a Terminal window opens automatically — log in there, close terminal, then return to Raycast| Command | Description |
|---|---|
| My Slurm Jobs | Your jobs with status, elapsed time, resource usage, logs, and cancel option |
| All Slurm Jobs | Same view for all jobs |
| HPC Info | Hardware information for all compute nodes |
| HPC Util | Live per-node CPU load, memory utilization and GPU allocation |
| Slurm Menu Bar | Running and pending job counts in the macOS menu bar |
| Select Clusters | Activate or deactivate SSH connections to your clusters |
| Preference | Default | Description |
|---|---|---|
| Control Persist | 12h | How long the SSH connection stays alive after the last command (OpenSSH ControlPersist) |
You can keep several clusters active at once. All data pages show one section per active cluster. Press ↵ on a cluster in Select Clusters to toggle it; this keeps the SSH connection alive so re-activation is instant. Use ⌘⇧X to fully close the connection and logout.
List your Slurm jobs for all active clusters.
Hardware information for all active clusters.
Live per-node CPU load, memory utilization and GPU availability.
List all Slurm jobs for all active clusters.
Show running and pending job counts in the macOS menu bar for all active clusters.
Select the active SSH hosts and authenticate.
↵ on any job to open a detail view with live per-job GPU, CPU, and memory utilization (sampled once per second via a persistent srun --overlap step) alongside run and trailing-window averagestail directly in Raycast, with copy actions for the file path and buffered outputRtx Pro 6000) and VRAM resolved from AllocTRES/ReqTRESControlPersist)