What it monitors: Utilization, temperature, memory, power draw, fan speed, clock speeds, PCIe info, P-State, throttle status, encoder/decoder sessions, active processes plus host CPU and RAM.
Everything streams over WebSocket and renders in real-time charts.
Multi-node support: Same image scales from a single machine to a cluster.
Run it on each GPU server, then point a hub instance at them one dashboard for all your GPUs, no extra infrastructure needed:
docker run -d -p 1312:1312 -e GPU_HOT_MODE=hub -e NODE_URLS=http://server1:1312,http://server2:1312 ghcr.io/psalias2006/gpu-hot:latest
Stack: FastAPI + NVML (nvidia-ml-py) on the backend, vanilla JS + Chart.js on the frontend.
MIT licensed. Happy to hear feedback — especially from anyone running larger GPU clusters.
Quick start: docker run -d --gpus all -p 1312:1312 ghcr.io/psalias2006/gpu-hot:latest
Live demo: https://psalias2006.github.io/gpu-hot/demo.html
GitHub: https://github.com/psalias2006/gpu-hot
What it monitors: Utilization, temperature, memory, power draw, fan speed, clock speeds, PCIe info, P-State, throttle status, encoder/decoder sessions, active processes plus host CPU and RAM. Everything streams over WebSocket and renders in real-time charts.
Multi-node support: Same image scales from a single machine to a cluster. Run it on each GPU server, then point a hub instance at them one dashboard for all your GPUs, no extra infrastructure needed:
docker run -d -p 1312:1312 -e GPU_HOT_MODE=hub -e NODE_URLS=http://server1:1312,http://server2:1312 ghcr.io/psalias2006/gpu-hot:latest
Stack: FastAPI + NVML (nvidia-ml-py) on the backend, vanilla JS + Chart.js on the frontend. MIT licensed. Happy to hear feedback — especially from anyone running larger GPU clusters.