gpustat

A local command-line accelerator monitor shows usage and running processes on NVIDIA and AMD GPUs and Huawei Ascend 910B NPUs.

gpustat is a local terminal monitor for people running AI workloads or sharing accelerator hardware. It helps you check device activity and identify the jobs using memory without opening a separate dashboard.

Device status: The default display shows each GPU's name, temperature, utilization, and used versus total memory.
Running processes: See the processes using a GPU alongside their memory consumption, with owner, command and process ID details to help identify individual jobs.
Live monitoring: Watch mode refreshes the terminal display periodically so you can follow resource use during a workload.
Python integration: Query accelerator status from your own Python code through the library API, with automatic backend detection or an explicitly selected backend.

The project is open source under the MIT license. It supports NVIDIA GPUs through NVML and the official nvidia-ml-py bindings, AMD GPUs through ROCm SMI, and Huawei Ascend 910B NPUs through the driver-provided npu-smi command. NVIDIA monitoring requires a compatible NVIDIA driver; Ascend monitoring requires npu-smi to be available on the machine.

GPU Management and Monitoring