gem5 Co-Pilot: AI Assistant Agent for Architectural Design Space Exploration

2025-10-22

AI Summary

Design space exploration (DSE) in computer architecture faces challenges including combinatorial explosion of configuration parameters, prohibitively high simulation overhead, and low efficiency of manual analysis. To address these, this paper introduces gem5 Co-Pilot—an AI-powered assistant for architectural DSE—integrating three key innovations: (1) a domain-specific language (DSL) and a Design Space Database (DSDB) for structured knowledge modeling; (2) a retrieval-augmented generation (RAG) framework enabling reuse of historical experimental knowledge and context-aware reasoning; and (3) an automated agent coupled with a web-based GUI to close the loop on parameter recommendation, result interpretation, and interactive optimization. Evaluated under four cost constraints, gem5 Co-Pilot accelerates identification of optimal configurations by 2.3–4.1× over baseline approaches, significantly improving both DSE efficiency and decision intelligence.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Cognitive Modeling & Cognitive Systems: Agent Architectures
• Constraint Satisfaction and Optimization: Search

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Generative AI is increasing the productivity of software and hardware development across many application domains. In this work, we utilize the power of Large Language Models (LLMs) to develop a co-pilot agent for assisting gem5 users with automating design space exploration. Computer architecture design space exploration is complex and time-consuming, given that numerous parameter settings and simulation statistics must be analyzed before improving the current design. The emergence of LLMs has significantly accelerated the analysis of long-text data as well as smart decision making, two key functions in a successful design space exploration task. In this project, we first build gem5 Co-Pilot, an AI agent assistant for gem5, which comes with a webpage-GUI for smooth user interaction, agent automation, and result summarization. We also implemented a language for design space exploration, as well as a Design Space Database (DSDB). With DSDB, gem5 Co-Pilot effectively implements a Retrieval Augmented Generation system for gem5 design space exploration. We experiment on cost-constraint optimization with four cost ranges and compare our results with two baseline models. Results show that gem5 Co-Pilot can quickly identify optimal parameters for specific design constraints based on performance and cost, with limited user interaction.

Problem

• Automating complex gem5 architectural design space exploration
• Accelerating analysis of simulation parameters and performance statistics
• Optimizing system configurations under specific cost constraints

Innovation

• AI agent automates gem5 design space exploration
• Retrieval Augmented Generation system with DSDB
• Webpage GUI enables interactive parameter optimization