A Learning-Based Oracle for Automatic Optimization

Abstract

Standing at the upstream of IDEA TA-I design flow, logic synthesis is a crucial step and profoundly impacts the performance of downstream tools. Especially in very challenging and competitive applications, such as individual combat equipment, radar, sonar, etc., the superiority of a design depends on how it is architected and optimized at the logic level. To enable "no-human-in-the-loop" unified layout generator and advancements in PPA, a high-quality and fast logic synthesis platform is required. This project provides a public release of a learning-based logic synthesis platform exploiting logic optimization for combinational benchmarks, with an improved platform for runtime and memory usage.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2023
Accession Number
AD1211996

Entities

People

  • Pierre-Emmanuel Gaillardon

Organizations

  • University of Utah

Tags

Communities of Interest

  • Autonomy
  • Energy and Power Technologies
  • Human Systems

DTIC Thesaurus Topics

  • Air Force
  • Air Force Facilities
  • Air Force Research Laboratories
  • Algorithms
  • Artificial Intelligence Software
  • Automatic
  • Circuits
  • Classification
  • Computational Science
  • Computer Programs
  • Computers
  • Computing System Architectures
  • Convolutional Neural Networks
  • Data Sets
  • Field Programmable Gate Arrays
  • Governments
  • Information Science
  • Logic Gates
  • Machine Learning
  • Neural Networks
  • Standards

Fields of Study

  • Engineering

Readers

  • Computational Fluid Dynamics (CFD)
  • Database Systems and Applications
  • Integrated Circuit Design and Technology.