Chip Performance Analyzer Chatbot

A specialized AI assistant for chip design and performance analysis, focusing on RISC-V and MIPS architectures.

Features

• 🤖 Specialized AI Assistant: Expert-level chip design analysis using Google's Gemini AI
• 📊 File Upload Support: Upload CSV, JSON, and text files for performance data analysis
• 📝 Text File Analysis: Support for instruction traces, logs, and other text-based data files
• 🔬 Performance Focus: Optimized for IPC analysis, cache performance, and architectural bottlenecks
• ⚡ Modern Setup: Uses uv for fast dependency management with pip fallback
• 🔧 Environment Management: Uses python-dotenv for secure API key management

Quick Start

Option 1: Using uv (Recommended)

# Install uv if you haven't already
curl -LsSf https://astral.sh/uv/install.sh | sh

# Run the installation script
./install.sh

# Start the chatbot
uv run python app.py

Option 2: Using pip

# Run the installation script (will use pip automatically)
./install.sh

# Activate virtual environment and start
source venv/bin/activate
python app.py

Option 3: Manual setup

# Install dependencies
pip install -r requirements.txt

# Configure environment (copies .env.example to .env)
cp .env.example .env
# Edit .env with your Google API key

# Test the setup
python test_setup.py

# Run the chatbot
python app.py

Testing

Run the test script to verify your setup:

uv run python test_setup.py
# or
python test_setup.py

This will check:

• Python version compatibility
• Required dependencies
• Environment configuration
• Sample file availability
• Basic functionality

Configuration

1. Copy .env.example to .env
2. Get your Gemini API key from Google AI Studio
3. Set your API key in the .env file:

   GEMINI_API_KEY=your_actual_api_key_here
   GEMINI_MODEL=gemini-2.5-flash
   

Usage

Interactive Commands

• /csv <filepath> - Load CSV file for analysis
• /json <filepath> - Load JSON file for analysis
• /txt <filepath> - Load text file for analysis (e.g., traces, logs)
• /files - List uploaded files
• /clear - Clear all uploaded files
• /help - Show available commands
• /quit - Exit the chatbot

Sample Files

The project includes sample files to test the functionality:

• sample_performance.csv - Processor performance metrics
• sample_benchmark.json - Detailed benchmark results
• inst_trace.txt - RISC-V instruction trace example

Example Sessions

CSV Analysis

🔬 You: /csv sample_performance.csv
✅ Successfully loaded CSV: sample_performance.csv (6 rows, 8 columns)

🔬 You: Analyze the IPC performance across different architectures
🤖 Assistant: Based on the uploaded performance data, I can see significant 
variations in IPC across architectures...

Text File Analysis

🔬 You: /txt inst_trace.txt
✅ Successfully loaded text file: inst_trace.txt (149 lines, 892 words)

🔬 You: What RISC-V instructions are most frequent in this trace?
🤖 Assistant: Analyzing the instruction trace, I can see several patterns...
The most frequent instructions are memory operations (sd) and arithmetic...

System Prompt

The chatbot uses a specialized system prompt (system_prompt.md) that configures it as:

• Senior chip design engineer with 15+ years experience
• Expert in RISC-V and MIPS architectures
• Specialist in performance optimization and EDA workflows
• Focused on quantitative analysis and actionable recommendations

Development

Dependencies

• Python 3.9+
• google-genai[aiohttp] - Google Generative AI SDK
• python-dotenv - Environment variable management
• Standard library modules (json, csv, asyncio, pathlib)

Project Structure

├── app.py                    # Main chatbot application
├── system_prompt.md          # AI assistant configuration
├── requirements.txt          # pip dependencies
├── pyproject.toml           # Modern Python project config
├── install.sh               # Installation script
├── setup.py                 # Python setup script (legacy)
├── test_setup.py            # Environment test script
├── env-example              # Environment template
├── sample_performance.csv   # Sample CSV data
└── sample_benchmark.json    # Sample JSON data