Reproduce “ALIGN: A System for Automating Analog Layout” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Tonmoy Dhar, Kishor Kunal, Yaguang Li, Meghna Madhusudan, Jitesh Poojary, Arvind K. Sharma, Wenbin Xu, Steven M. Burns and 6 more

ALIGN is an open-source system designed to automate analog integrated circuit layout generation, aiming for 24-hour, no-human-in-the-loop turnaround times. The system, developed by the University of Minnesota, Texas A&M University, and Intel Labs, leverages machine learning and a hierarchical approach to generate layouts for diverse analog circuits with electrical performance competitive with hand-crafted designs.

Key insights

  • Nanometer-scale technologies with restricted design rules simplify the analog layout design space, making automation more feasible.
  • The increasing demand for analog content with modest performance requirements in SoCs creates a 'sweet spot' for automation.
  • Machine learning, particularly Graph Neural Networks, can effectively automate complex recognition tasks like hierarchy detection and symmetry identification, which were previously manual bottlenecks.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.