ChatNeuroSim: An LLM Agent Framework for Automated Compute-in-Memory Accelerator Deployment and Optimization

TL;DR

ChatNeuroSim is an LLM-based framework that automates the deployment and optimization of compute-in-memory accelerators for neural networks, significantly reducing design time and improving efficiency.

Contribution

It introduces an automated LLM-driven workflow for CIM design space exploration, including task automation and a novel optimizer with design space pruning.

Findings

Successfully automates CIM deployment and optimization tasks.

Achieves up to 0.79x reduction in optimization runtime with pruning.

Validates effectiveness on DNN workloads like Swin Transformer Tiny.

Abstract

Compute-in-Memory (CIM) architectures have been widely studied for deep neural network (DNN) acceleration by reducing data transfer overhead between the memory and computing units. In conventional CIM design flows, system-level CIM simulators (such as NeuroSim) are leveraged for design space exploration (DSE) across different hardware configurations and DNN workloads. However, CIM designers need to invest substantial effort in interpreting simulator manuals and understanding complex parameter dependencies. Moreover, extensive design-simulation iterations are often required to identify optimal CIM configurations under hardware constraints. These challenges severely prolong the DSE cycle and hinder rapid CIM deployment. To address these challenges, this work proposes ChatNeuroSim, a large language model (LLM)-based agent framework for automated CIM accelerator deployment and optimization. ChatNeuroSim automates the entire CIM workflow, including task scheduling, request parsing and adjustment, parameter dependency checking, script generation, and simulation execution. It also integrates the proposed CIM optimizer using design space pruning, enabling rapid identification of optimal configurations for different DNN workloads. ChatNeuroSim is evaluated on extensive request-level testbenches and demonstrates correct simulation and optimization behavior, validating its effectiveness in automatic request parsing and task execution. Furthermore, the proposed design space pruning technique accelerates CIM optimization process compared to no-pruning baseline. In the case study optimizing Swin Transformer Tiny under 22 nm technology, the proposed CIM optimizer achieves a 0.42$\times$-0.79$\times$ average runtime reduction compared to the same optimization algorithm without design space pruning.