Huatuo, the open‑source deep‑observability platform backed by Didi, now supports real‑time monitoring of MetaX GPUs, offering detailed hardware metrics via Docker or Kubernetes deployments and exposing them through a /metrics endpoint for cloud‑native AI and operations use cases.
This guide introduces nvitop, an interactive NVIDIA‑GPU process viewer and resource manager, explains its key features, shows how to install it via uvx/pipx, demonstrates basic device and process commands as well as the real‑time monitoring mode, and provides troubleshooting tips for common issues.
The article introduces GPUprobe, an eBPF‑based tool that provides lightweight, continuous, application‑level monitoring of CUDA memory allocation, leaks, and kernel launches, compares it with NSight Systems and DCGM, and demonstrates near‑zero overhead integration with Prometheus and Grafana through detailed code examples and real‑world output analysis.
This article systematically reviews GPU monitoring for large‑scale AI training, covering MFU/HFU definitions, key DCGM metrics, NVLink bandwidth, common failure codes such as Xid and SXid, experimental insights on T4 and H100 GPUs, and practical case studies for diagnosing and mitigating performance drops.
This guide explains how to install and use three command‑line tools—lm‑sensors, Glances, and i7z—on Ubuntu to monitor CPU and GPU temperatures, fan speeds, and other hardware metrics, providing step‑by‑step commands and example outputs for effective laptop cooling diagnostics.
This article explains how to use TensorBoard and the Timeline tool to monitor TensorFlow GPU utilization, identify operation bottlenecks, and visualize execution times, including code examples and steps for exporting and merging profiling data for repeated runs.
