Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SynthAI: A Multi Agent Generative AI Framework for Automated Modular HLS Design Generation

Seyed Arash Sheikholeslam, Andre Ivanov

TL;DR

This paper introduces SynthAI, a new method for the automated creation of High-Level Synthesis (HLS) designs that integrates ReAct agents, Chain-of-Thought prompting, web search technologies, and the Retrieval-Augmented Generation framework within a structured decision graph.

Abstract

In this paper, we introduce SynthAI, a new method for the automated creation of High-Level Synthesis (HLS) designs. SynthAI integrates ReAct agents, Chain-of-Thought (CoT) prompting, web search technologies, and the Retrieval-Augmented Generation (RAG) framework within a structured decision graph. This innovative approach enables the systematic decomposition of complex hardware design tasks into multiple stages and smaller, manageable modules. As a result, SynthAI produces synthesizable designs that closely adhere to user-specified design objectives and functional requirements. We further validate the capabilities of SynthAI through several case studies, highlighting its proficiency in generating complex, multi-module logic designs from a single initial prompt. The SynthAI code is provided via the following repo: \url{https://github.com/sarashs/FPGA_AGI}

SynthAI: A Multi Agent Generative AI Framework for Automated Modular HLS Design Generation

TL;DR

This paper introduces SynthAI, a new method for the automated creation of High-Level Synthesis (HLS) designs that integrates ReAct agents, Chain-of-Thought prompting, web search technologies, and the Retrieval-Augmented Generation framework within a structured decision graph.

Abstract

In this paper, we introduce SynthAI, a new method for the automated creation of High-Level Synthesis (HLS) designs. SynthAI integrates ReAct agents, Chain-of-Thought (CoT) prompting, web search technologies, and the Retrieval-Augmented Generation (RAG) framework within a structured decision graph. This innovative approach enables the systematic decomposition of complex hardware design tasks into multiple stages and smaller, manageable modules. As a result, SynthAI produces synthesizable designs that closely adhere to user-specified design objectives and functional requirements. We further validate the capabilities of SynthAI through several case studies, highlighting its proficiency in generating complex, multi-module logic designs from a single initial prompt. The SynthAI code is provided via the following repo: \url{https://github.com/sarashs/FPGA_AGI}
Paper Structure (9 sections, 1 equation, 3 figures, 1 table, 2 algorithms)

This paper contains 9 sections, 1 equation, 3 figures, 1 table, 2 algorithms.

Table of Contents

  1. Introduction
  2. LLM-based Technologies used in SynthAI
  3. MRKL
  4. Interactive Agents
  5. Retrieval Augmented Generation (RAG)
  6. SynthAI Architecture
  7. Hardware design
  8. Case studies
  9. Conclusion

Figures (3)

  • Figure 1: Individual agent. Each of the blue nodes in figures \ref{['SynthAI']} and \ref{['SynthAI-hw']} is an agent similar to the one depicted here.
  • Figure 2: Flow diagram of knowledge gathering component consisting of various agents, evaluators, and a vector database. The blue nodes are agents that employ LLMs and have an architecture similar to Figure \ref{['single_node']}
  • Figure 3: Flow diagram of hardware design component consisting of various agents and evaluators.