nv-monitor
Share on:
Enter Your Mastodon Instance
Copy the link below to share to Mastodon
https://terminaltrove.com/nv-monitor/
A low-level Nvidia GPU system monitor.
Script Preview
Review this script before executing in your terminal.
nv-monitor is a low-level NVIDIA system monitor built for monitoring multiple high performance GPU clusters in the terminal.
Alongside monitoring the CPU cores and memory usage of the running machine, each connected GPU is monitored with it's temperature, power, VRAM, and aggregate GPU usage. The top used processes are shown below the GPU stats. It handles unified memory, includes synthetic load generation and can also monitor RDMA or InfiniBand throughput.
Developers or machine learning engineers specializing in high performance computing, CUDA kernel performance tuning, inference engineering, or DGX users would find nv-monitor useful when they need to monitor GPU health and cluster operations for scaling GPU workloads or inference performance in the terminal.
