Which tool shows the GPUs, driver version and running processes on one host?
Why A is the answer
nvidia-smi
nvidia-smi is the standard command-line view.
In detail
- nvidia-smi is the standard command-line view of a host.
- It is the standard command-line view.
Use case
An illustrative scenario, not a real incident.
An engineer runs nvidia-smi on one host, and sees GPUs, driver version and processes.
See it in different ways
flowchart TD A["DCGM"] --> B["Utilization"] A --> C["Memory"] A --> D["Temperature"] A --> E["Power"] A --> F["ECC and Xid errors"]
mindmap
root((Operations))
Monitoring
nvidia-smi
DCGM
Scheduling
Slurm
Kubernetes
Provisioning
Base Command Manager
Sharing
MIG
vGPUflowchart LR S0["Log in to host"] --> S1["Run nvidia-smi"] --> S2["Read the table"]
Key point: nvidia-smi shows one host.
Why the other options are wrong
- B. ping
- ping tests reachability.
- C. traceroute
- traceroute shows network paths.
- D. top-k
- That is a sampling setting.