gpu-benchy

A lightweight Python CLI tool that benchmarks GPU rendering performance across multiple resolutions using OpenGL and GLFW. Automatically detects available GPUs, runs a fixed-function triangle render test, and reports the highest resolution that maintains ≥30 FPS.

Features

• Cross-platform GPU detection (Linux, Windows, macOS)
• Auto-installs dependencies (glfw, PyOpenGL) on first run
• Multi-resolution testing from 128x128 up to 12288x6480
• 3-second benchmark window per resolution with real-time FPS tracking
• Smart cutoff reporting (finds max resolution at ≥30 FPS)
• Headless-friendly (benchmark windows are hidden during execution)

Prerequisites

• Python 3.6+
• Working OpenGL drivers
• Internet connection (first time)
• A running display server (Linux/macOS) or desktop session (Windows)

Usage

Simply run the script directly. Missing Python packages will be installed automatically:

python gpu-benchy.py

Example Output

Found 2 GPU(s):
  [0] NVIDIA GeForce RTX 4070
  [1] Intel(R) UHD Graphics 630

=== Benchmarking GPU 0: NVIDIA GeForce RTX 4070 ===
128x128: 14520.33 FPS
240x240: 9841.12 FPS
...
7680x4320: 45.20 FPS
12288x6480: 22.15 FPS
[RESULT] NVIDIA GeForce RTX 4070 can render up to 7680x4320 at ≥30 FPS

License

GNU Lesser General Public License v2.1 - Feel free to use and modify