SimulatorCoder: DNN Accelerator Simulator Code Generation and Optimization via Large Language Models

2026-02-19

AI Summary

This work proposes the first large language model (LLM)-based framework for automatically generating deep neural network (DNN) accelerator simulators, addressing the lengthy development cycles and limited adaptability of traditional hand-coded approaches in the face of rapidly evolving architectures. By leveraging domain-specific prompt engineering, in-context learning (ICL), and chain-of-thought (CoT) reasoning, the framework translates natural language descriptions into high-fidelity, cycle-accurate simulator code in an end-to-end manner. Functional correctness and performance are ensured through iterative feedback loops and validation against the SCALE-Sim benchmark. The generated simulators achieve cycle-level accuracy with less than 1% error and demonstrate significantly higher runtime efficiency compared to manual implementations, drastically reducing development time.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

This paper presents SimulatorCoder, an agent powered by large language models (LLMs), designed to generate and optimize deep neural network (DNN) accelerator simulators based on natural language descriptions. By integrating domain-specific prompt engineering including In-Context Learning (ICL), Chain-of-Thought (CoT) reasoning, and a multi-round feedback-verification flow, SimulatorCoder systematically transforms high-level functional requirements into efficient, executable, and architecture-aligned simulator code. Experiments based on the customized SCALE-Sim benchmark demonstrate that structured prompting and feedback mechanisms substantially improve both code generation accuracy and simulator performance. The resulting simulators not only maintain cycle-level fidelity with less than 1% error compared to manually implemented counterparts, but also consistently achieve lower simulation runtimes, highlighting the effectiveness of LLM-based methods in accelerating simulator development. Our code is available at https://github.com/xiayuhuan/SimulatorCoder.

Problem

• DNN accelerator simulator
• code generation
• large language models
• natural language description
• simulation optimization

Innovation

• Large Language Models
• DNN Accelerator Simulation
• Prompt Engineering
• Code Generation
• Feedback-Verification Loop