⚡ Power Gating Decoder ML Optimizer

Advanced Machine Learning System for Low-Power Decoder Design Optimization

📋 Setup Instructions

1. GitHub Repository Setup

1. Create a new GitHub repository (or use existing one)
2. Upload these files:
   • app.py
   • requirements.txt
   • packages.txt
   • decoder_power_delay_area_dataset.csv

2. Update Dataset Path in app.py

In app.py, find line ~65 and update with your GitHub username and repo name:

url = "https://raw.githubusercontent.com/YOUR_USERNAME/YOUR_REPO/main/decoder_power_delay_area_dataset.csv"

Example:

url = "https://raw.githubusercontent.com/john_doe/power-decoder-ml/main/decoder_power_delay_area_dataset.csv"

3. Deploy on Streamlit Cloud

1. Go to streamlit.io/cloud
2. Sign in with GitHub
3. Click "New app"
4. Select your repository
5. Set main file: app.py
6. Click "Deploy"!

4. Local Testing (Optional)

# Install dependencies
pip install -r requirements.txt

# Run locally
streamlit run app.py

📊 Dataset Format

Your decoder_power_delay_area_dataset.csv must have these columns:

Column	Description	Unit
decoder_size	Decoder bits (2-6)	bits
tech_node	Technology node	nm
supply_voltage	Supply voltage	V
threshold_voltage	Threshold voltage	V
transistor_width	Transistor width	µm
load_capacitance	Load capacitance	fF
pg_efficiency	Power gating efficiency	0-1
switching_activity	Switching activity	0-1
leakage_factor	Leakage factor	0-1
temperature	Operating temperature	°C
power	Power consumption (target)	mW
delay	Propagation delay (target)	ns
area	Silicon area (target)	µm²

🚀 Features

• 🤖 4 ML Algorithms: Random Forest, Gradient Boosting, Neural Networks, SVR
• 📊 Interactive Dashboard: Real-time predictions and visualizations
• 🎯 Multi-Objective Optimization: Power-Delay-Area trade-off analysis
• 📈 3D Pareto Front: Beautiful interactive visualization
• 📄 Auto-Generated Reports: Conference paper ready documentation
• 💾 Export Functionality: Download results and reports

🎯 Usage

1. Data Overview Tab: Explore your dataset statistics and distributions
2. ML Training Tab: Train models and compare performance
3. Predictions Tab: Make real-time predictions for custom configurations
4. Optimization Tab: Find Pareto-optimal designs
5. Report Tab: Generate comprehensive project report

🛠️ Technologies

• Framework: Streamlit
• ML Library: scikit-learn
• Visualization: Plotly
• Data Processing: Pandas, NumPy

📝 Project Details

Project: Power Gating in Decoders - Low Power Strategy
Course: B.Tech Final Year Project (045-320)
Topic: ML-Driven Multi-Objective Optimization

🎓 For Conference Paper

The app automatically generates:

• Comprehensive methodology section
• Results and comparisons
• Feature importance analysis
• Pareto-optimal configurations
• Complete project report

💡 Tips

• Use the sidebar to configure ML algorithms and optimization targets
• Export datasets and reports for your documentation
• The 3D Pareto front is perfect for presentations!
• Use the generated report for your conference paper

📞 Support

For issues or questions:

1. Check that dataset path in app.py is correct
2. Verify CSV file has all required columns
3. Ensure GitHub repository is public

---

Good luck with your project! 🚀