gpu-benchmark 0.4.3

GPU benchmarking tool for training and inference

GPU Benchmark by United Compute

A simple CLI tool to benchmark your GPU's performance across various AI models (Stable Diffusion, LLMs, etc.) and compare results in our global benchmark results.

Installation

pip install gpu-benchmark

Usage

Run the benchmark (takes 5 minutes after the pipeline is loaded):

gpu-benchmark

Available Benchmarks

You can specify which model to benchmark using the --model flag:

Stable Diffusion 1.5 (Default)

gpu-benchmark --model stable-diffusion-1-5

Qwen 3.0 6B (LLM Inference)

gpu-benchmark --model qwen3-0-6b

nanoGPT (LLM Training)

gpu-benchmark --model nanogpt-train

Optional Arguments

If you're running on a cloud provider, specify it with the --provider flag:

gpu-benchmark --provider runpod

For multi-GPU systems, you can select a specific GPU like this:

1. Using the --gpu flag:

gpu-benchmark --gpu 1  # Uses GPU index 1

The tool will:

1. Load the selected model (Stable Diffusion, Qwen, or nanoGPT)
2. Run the benchmark for 5 minutes
3. Track performance metrics (throughput/iterations) and GPU temperature
4. Upload results to the United Compute Benchmark Results

What it measures

• Benchmark Score: Number of iterations or images generated in 5 minutes (model-dependent)
• GPU Model: The specific model of your GPU (e.g., NVIDIA GeForce RTX 4090)
• Max Heat: Maximum GPU temperature reached (°C)
• Avg Heat: Average GPU temperature during the benchmark (°C)
• Country: Your location (detected automatically)
• GPU Power: Power consumption in watts (W)
• GPU Memory: Total GPU memory in gigabytes (GB)
• Platform: Operating system information
• Acceleration: CUDA version
• PyTorch Version: PyTorch library version

Requirements

• CUDA-compatible NVIDIA GPU
• Python 3.8+

Links

• Official Website
• GPU Benchmark Results