Real-Time GPU Monitoring on Polybar Using Bash + NVIDIA-SMI

🚀 Introduction

As part of optimizing my Linux development environment, I recently implemented a lightweight GPU monitoring widget using Polybar and a custom Bash script. This setup fetches real-time GPU utilization and temperature using nvidia-smi, integrates it seamlessly into Polybar, and refreshes at a short interval all without relying on bulky third-party tools.

The goal? Clean visibility into GPU performance, directly from my desktop bar, with zero overhead.

---

🔧 Stack & Tools

This kind of setup is ideal if you're running a custom lightweight window manager (like Openbox or bspwm), and want telemetry that aligns with the rest of your minimalist workflow.

---

💻 Bash Script

Here’s the core script I wrote:

bashCopyEdit#!/bin/bash

TEMP=$(nvidia-smi --query-gpu=temperature.gpu --format=csv,noheader,nounits)
UTIL=$(nvidia-smi --query-gpu=utilization.gpu --format=csv,noheader,nounits)

echo "GPU: ${UTIL}% | Temp: ${TEMP}°C"

You can drop this in ~/.config/polybar/scripts/gpu.sh and make it executable:

bashCopyEditchmod +x ~/.config/polybar/scripts/gpu.sh

This leverages nvidia-smi's CLI output directly to avoid additional dependencies and overhead.

---

🧩 Integrating with Polybar

Add this module to your ~/.config/polybar/config.ini:

iniCopyEdit[module/gpu]
type = custom/script
exec = ~/.config/polybar/scripts/gpu.sh
interval = 2
label = %output%

And then include it in your active bar setup:

iniCopyEditmodules-right = gpu ...

This ensures your GPU stats are updated every 2 seconds in real-time - no lag, no noise.

---

🖼️ Final Output

You’ll see clean stats like this on your bar:

yamlCopyEditGPU: 12% | Temp: 49°C

It’s functional, non-intrusive, and blends well into any custom theme.

Bonus: You can extend this with icons, conditional coloring, or integration into a dotfiles repo for portability.

---

🧠 Why I Did This

• Minimal setups deserve minimal tools.

• I prefer scriptable, version-controlled configurations over GUIs.

• This is part of a larger push: building a fully automated, dev-optimized Linux system from the ground up.

---

🔮 What’s Next

This was step one. I’m currently:

• Writing shell scripts to automate full dev machine provisioning

• Extending Polybar to show CPU, RAM, and battery metrics

• Exploring integrations with systemd timers and cron for power-aware switching

• Designing a Linux distro tailored for developers, with smart GPU/battery management and LLM-aware optimization baked in

---

💬 Final Thoughts

If you're building your own Linux stack - start here.
Don’t install 10 apps to monitor your machine. Just write 10 lines of code.

If this helped or inspired you, connect on LinkedIn.