Struct GpuUtilizationReader 

pub struct GpuUtilizationReader { /* private fields */ }

One-shot accumulator that opens NVML once and reads per-GPU utilisation for every NVIDIA device on each load tick. Holding the NVML handle across reads avoids the init cost (microseconds) on every tick and is the documented pattern.

Implementations

impl GpuUtilizationReader

pub fn new() -> Self

Build a reader. NVML init failure is non-fatal — the reader folds to “all zeroes” on every NVIDIA device and the rest of the load-tick path stays alive. Logged once at startup so operators can tell “no NVIDIA card” from “NVIDIA card but driver missing”.

pub fn read(&self, device: &GpuDevice) -> GpuUtilization

Read the per-tick snapshot for one device. Cheap when NVML is available (handful of FFI calls); free when it’s not (returns the zero-initialised default).

Trait Implementations

impl Default for GpuUtilizationReader

fn default() -> Self

Returns the “default value” for a type.