A simple GPU monitoring tool with real-time data display and historical usage charts.
- Real-time GPU data display in separate tables for GPU info and processes
- Ability to select and monitor a specific GPU
- Historical GPU usage chart with adjustable time window
- Configurable refresh rate (supports both spinbox arrows and manual input with Enter key)
- Basic process information (which processes are using the selected GPU)
- Language switching between English and Chinese
- Python 3.6+
- NVIDIA GPU with drivers installed
- PySide6
- GPUtil
- matplotlib
- numpy
- psutil
- Clone this repository:
git clone https://github.com/yourusername/gpu-top.git
cd gpu-top
- Install the required dependencies:
pip install -r requirements.txt
Run the application:
python main.py
- Refresh Rate: Adjust how frequently the GPU data is updated (in milliseconds). You can use the spinbox arrows or manually enter a value and press Enter.
- Time Window: Set the time range for the historical GPU usage chart
- Select GPU: Choose which GPU to monitor when multiple GPUs are available
- Language: Switch between English and Chinese interface
- The process information feature is simplified and may not accurately show all GPU processes
- For more accurate process information, consider using NVIDIA's System Management Interface (nvidia-smi)
- The application allows you to select which GPU to monitor when multiple GPUs are present
MIT