ROCm Testing Suite

A comprehensive testing suite for validating ROCm (Radeon Open Compute) functionality with Python machine learning frameworks and libraries.

Overview

This project provides a collection of tests and utilities to verify ROCm installation and compatibility with popular Python ML frameworks including PyTorch, TensorFlow, Keras, and pgmpy. It's designed to help developers ensure their ROCm setup is working correctly for GPU-accelerated machine learning workloads.

Features

• ROCm Installation Verification: Check if ROCm is properly installed and accessible
• GPU Information Retrieval: Display detailed GPU information using rocm-smi
• Framework Testing: Test PyTorch, TensorFlow, and Keras with ROCm backend
• Bayesian Network Testing: Validate pgmpy functionality for probabilistic modeling
• Comprehensive Test Suite: Multiple test scenarios for different use cases

Prerequisites

System Requirements

• AMD GPU with ROCm support
• Ubuntu 20.04/22.04 or compatible Linux distribution
• Python 3.8+ with virtual environment support

ROCm Installation

Ensure ROCm is installed on your system. Follow the official ROCm Installation Guide.

Quick Start

1. Clone and Setup

git clone <repository-url>
cd ROCM

2. Create Virtual Environment

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install Dependencies

pip install -r requirements.txt

4. Run Tests

# Run all tests
python -m pytest src/tests/ -v

# Run specific tests
python src/tests/test_rocm.py
python src/tests/test_pytorch.py
python src/tests/test_tensorflow.py

Project Structure

ROCM/
├── README.md                 # Project documentation
├── requirements.txt          # Python dependencies
├── pyproject.toml           # Project configuration
├── .github/                 # GitHub workflows and templates
├── .copilot/                # GitHub Copilot configuration
├── .vscode/                 # VS Code settings
├── docs/                    # Additional documentation
│   ├── directions.txt       # Setup instructions
│   ├── git_directions.txt   # Git workflow
│   └── github_directions.txt # GitHub integration
└── src/                     # Source code
    ├── tests/               # Test modules
    │   ├── test_rocm.py     # ROCm installation tests
    │   ├── test_pytorch.py  # PyTorch ROCm tests
    │   ├── test_tensorflow.py # TensorFlow ROCm tests
    │   ├── test_keras.py    # Keras compatibility tests
    │   ├── test_pgmpy.py    # Bayesian network tests
    │   ├── test_tf_help.py  # TensorFlow utilities
    │   └── test_torch_min.py # Minimal PyTorch test
    ├── tools/               # Utility tools
    └── utils/               # Helper utilities

Test Descriptions

ROCm Core Tests (test_rocm.py)

• Verifies ROCm installation
• Checks rocm-smi availability
• Displays GPU information and status

PyTorch Tests (test_pytorch.py)

• Tests PyTorch ROCm backend availability
• Performs tensor operations on GPU
• Validates device placement and computation

TensorFlow Tests (test_tensorflow.py)

• Complete MNIST neural network training
• Tests TensorFlow with ROCm backend
• Includes detailed training process explanation

Keras Tests (test_keras.py)

• Verifies Keras availability within TensorFlow
• Checks version compatibility

Probabilistic Modeling Tests (test_pgmpy.py)

• Bayesian network construction and inference
• Demonstrates probabilistic reasoning capabilities

Development

Adding New Tests

1. Create test files in src/tests/
2. Follow the existing naming convention: test_<framework>.py
3. Include proper error handling and informative output
4. Update this README with test descriptions

Contributing

1. Fork the repository
2. Create a feature branch
3. Add tests and documentation
4. Submit a pull request

Troubleshooting

Common Issues

ROCm Not Detected

• Verify ROCm installation: /opt/rocm/bin/rocm-smi
• Check GPU compatibility with ROCm
• Ensure proper driver installation

PyTorch ROCm Issues

• Install ROCm-compatible PyTorch version
• Verify PYTORCH_ROCM_ARCH environment variable
• Check GPU memory availability

Virtual Environment Issues

• Install python3-venv: sudo apt install python3-venv
• Use system Python if distribution packages conflict
• Consider using conda for complex dependency management

Getting Help

• Check the docs/ directory for detailed setup instructions
• Review test output for specific error messages
• Consult ROCm documentation for hardware-specific issues

License

This project is open source. See LICENSE file for details.

Acknowledgments

• AMD ROCm team for GPU compute platform
• PyTorch and TensorFlow communities for ROCm support
• Contributors to the Python scientific computing ecosystem