MIREDO: MIP-Driven Resource-Efficient Dataflow Optimization for Computing-in-Memory Accelerator

TL;DR

MIREDO is a novel MIP-based framework that optimizes dataflow in CIM accelerators, significantly improving DNN performance by effectively navigating complex design constraints.

Contribution

It introduces a hierarchical hardware abstraction and analytical latency model to systematically optimize dataflow configurations in CIM accelerators.

Findings

Achieves up to 3.2x performance improvement on DNNs

Effectively models CIM-specific data transfer behaviors

Outperforms existing optimization approaches

Abstract

Computing-in-Memory (CIM) architectures have emerged as a promising solution for accelerating Deep Neural Networks (DNNs) by mitigating data movement bottlenecks. However, realizing the potential of CIM requires specialized dataflow optimizations, which are challenged by an expansive design space and strict architectural constraints. Existing optimization approaches often fail to fully exploit CIM accelerators, leading to noticeable gaps between theoretical and actual system-level efficiency. To address these limitations, we propose the MIREDO framework, which formulates dataflow optimization as a Mixed-Integer Programming (MIP) problem. MIREDO introduces a hierarchical hardware abstraction coupled with an analytical latency model designed to accurately reflect the complex data transfer behaviors within CIM systems. By jointly modeling workload characteristics, dataflow strategies, and CIM-specific constraints, MIREDO systematically navigates the vast design space to determine the optimal dataflow configurations. Evaluation results demonstrate that MIREDO significantly enhances performance, achieving up to $3.2\times$ improvement across various DNN models and hardware setups.