Unicorn-CIM: Uncovering the Vulnerability and Improving the Resilience of High-Precision Compute-in-Memory

TL;DR

Unicorn-CIM identifies vulnerabilities in high-precision compute-in-memory architectures and proposes an algorithm-hardware co-design with lightweight error correction to enhance reliability with minimal overhead.

Contribution

This paper introduces Unicorn-CIM, a novel framework that uncovers soft error vulnerabilities in FP CIM and proposes a co-designed solution to improve its resilience.

Findings

High sensitivity of high-precision DNNs to exponent errors

Efficient ECC scheme with 8.98% logic overhead

Enhanced model robustness and accuracy

Abstract

Compute-in-memory (CIM) architecture has been widely explored to address the von Neumann bottleneck in accelerating deep neural networks (DNNs). However, its reliability remains largely understudied, particularly in the emerging domain of floating-point (FP) CIM, which is crucial for speeding up high-precision inference and on device training. This paper introduces Unicorn-CIM, a framework to uncover the vulnerability and improve the resilience of high-precision CIM, built on static random-access memory (SRAM)-based FP CIM architecture. Through the development of fault injection and extensive characterizations across multiple DNNs, Unicorn-CIM reveals how soft errors manifest in FP operations and impact overall model performance. Specifically, we find that high-precision DNNs are extremely sensitive to errors in the exponent part of FP numbers. Building on this insight, Unicorn-CIM develops an efficient algorithm-hardware co-design method that optimizes model exponent distribution through fine-tuning and incorporates a lightweight Error Correcting Code (ECC) scheme to safeguard high-precision DNNs on FP CIM. Comprehensive experiments show that our approach introduces just an 8.98% minimal logic overhead on the exponent processing path while providing robust error protection and maintaining model accuracy. This work paves the way for developing more reliable and efficient CIM hardware.