ALIGN: Analog Layout, Intelligently Generated from Netlists

Abstract

The ALIGN project seeks to develop open-source software for analog/mixed-signal circuit layout to build a hierarchical, machine learning (ML) driven physical layout generator. Four classes of analog/mixed signal circuits are considered: low-frequency circuits, wireline systems, wireless systems, and power delivery circuits. The philosophy of ALIGN is to identify a set of building blocks from an input netlist and to hierarchically build layouts for these blocks from the cell level to the system level. ALIGN generates parameterized layouts for the lowest-level cells, composes these cells into larger sub-circuits through floor planning/placement strategies, and performs performance-aware routing. At each step, ALIGN comprehends design rule constraints and electrical constraints and translates these to directives for each layout step. The approaches use a mix of graph-based techniques and accelerated physical design methods, coupled with machine learning techniques that enable fast design optimization.

Document Details

Document Type
DoD Grant Award
Publication Date
Jan 13, 2022
Source ID
N660011824048

Entities

People

  • Sachin Sapatnekar

Organizations

  • Defense Advanced Research Projects Agency
  • Naval Information Warfare Center Pacific
  • Regents of the University of Minnesota

Tags

Fields of Study

  • Computer science

Readers

  • Defense Acquisition Program Management
  • Integrated Circuit Design and Technology.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Machine Learning Algorithms