nvidia-smi
nvidia-smi is NVIDIA’s command-line management and monitoring utility, built on NVML. It reports device state and exposes supported administrative operations.
How it works
Its familiar GPU utilization value measures the share of a sample interval during which one or more kernels executed. Memory utilization measures memory-interface activity over its sample interval; memory used is an allocation quantity. These fields answer different questions.
Work through an example
A small kernel running throughout the interval can produce high GPU utilization while leaving much arithmetic capacity unused. Use a timeline and kernel profiler to investigate useful throughput.
Where the term can mislead
The CUDA version field (called CUDA UMD Version in newer documentation) describes driver support. It does not prove that nvcc, a matching toolkit or the correct framework libraries are installed.
Connect the idea
Take it into practice
Continue with the existing concept, apply it to an interview problem, or follow a course lesson. Access requirements are shown before you open a page.
New to GPUs? Start with the free execution-model concept.
Sources and scope
Examples here are teaching calculations, not measured GPU performance. Consult the linked documentation for the exact device, instruction and tool version. Source review: 25 October 2026.