Linux GPU monitoring

Linux NVIDIA GPU monitoring : summary

XorMon monitors NVIDIA GPU attached to a Linux server.
It is done vi XorMon NVIDIA agent which has linked necessary Nvidia libraries, install instructions.

Monitored metrics

  • CPU usage[%]
  • Memory usage [%]
  • Memory activity [%]
  • Memory used/free/reserved [GiB]
  • Power [Watt]
  • Temperature [°C]
  • Encoder [%]
  • Decoder [%]
  • Clocks [Hertz]
  • PCIe Rx/Tx [MB/sec]

Where to find it

  • Global: "Linux GPU" under "Server" menu.
    It lists all attached Nvidia GPU and providing overall:
    • Total graphs for all GPU and metrics
    • Heatmap
    • TOP tables (averaged per day/week/month)
    • Configuration
    • List of all Nvidia GPU hosts
    Linux NVIDIA GPU monitoring : GPU


  • Local: under standard "Linux" OS agent tabs is "GPU" one with all metrics
  • Linux NVIDIA GPU monitoring : GPU local

Examples

Memory
Linux NVIDIA GPU monitoring : memory

Power
Linux NVIDIA GPU monitoring : Power