MAxPy

MAxPy is a framework aimed for simulation and exploration of Approximate Computing techniques in VLSI designs. It is Ptyhon-based, free and open-source (available in Github).

MAxPy is part of the MAxPy Project - a growing collaborative effort to make Approximate Computing available to the community via easily applicable tools. Check out our organization page at Github.

Note

This project is under active development.

Contents

• Overview
• Installation
  • System packages
  • MAxPy package
• Basic tutorial
  • The problem
  • RTL design
  • Converting from Verilog to Python
  • Writing a testbench
  • Exploring Approximate Arithmetic Blocks
  • Probabilistic pruning
  • Conclusion
• Case Study: Sobel Edge Detection
  • Sobel Basics
  • Quality Metrics
  • RTL Exact Description
  • Using Approximate Adders
  • Using Probabilistic Pruning
  • Conclusion
• Development
  • Getting the repository
  • Installing the package locally
  • Building documentation
  • MAxPy’s internals
• Contact