NVIDIA Management Library
NVML is NVIDIA’s management and monitoring API for supported devices. It exposes information such as device identity, memory use, utilization and power under the platform’s permissions and capabilities.
How it works
Tools use NVML to collect sampled device state. Those samples have defined intervals and support limits; they are not a reconstruction of every kernel or request. On Linux, NVIDIA documents the library name as libnvidia-ml.so, rather than the informal libnvml.so name.
Work through an example
A device collector can record memory used and a utilization sample alongside request outcomes. Correlating them helps investigation, but the device sample cannot identify which request caused a stall.
Where the term can mislead
Management API support differs by product and partitioning mode. An unavailable field should remain unavailable, not be reported as zero.
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.