Developing Synthesis Flows Without Human Knowledge

TL;DR

This paper introduces an autonomous framework utilizing CNNs to generate design-specific synthesis flows for ICs and SoCs, reducing reliance on expert knowledge and addressing the complexity of modern design processes.

Contribution

It presents a novel CNN-based method for automatically creating synthesis flows without human input, improving scalability and customization for complex IC designs.

Findings

Successfully designed logic synthesis flows for three large-scale designs

Demonstrated the feasibility of autonomous flow generation

Reduced dependence on expert knowledge in flow development

Abstract

Design flows are the explicit combinations of design transformations, primarily involved in synthesis, placement and routing processes, to accomplish the design of Integrated Circuits (ICs) and System-on-Chip (SoC). Mostly, the flows are developed based on the knowledge of the experts. However, due to the large search space of design flows and the increasing design complexity, developing Intellectual Property (IP)-specific synthesis flows providing high Quality of Result (QoR) is extremely challenging. This work presents a fully autonomous framework that artificially produces design-specific synthesis flows without human guidance and baseline flows, using Convolutional Neural Network (CNN). The demonstrations are made by successfully designing logic synthesis flows of three large scaled designs.