Segmented Exponent Alignment and Dynamic Wordline Activation for Floating-Point Analog CIM Macros

TL;DR

This paper introduces SEA and DWA techniques to improve floating-point CIM macro efficiency by reducing power and delay, leveraging exponent clustering and dynamic activation strategies.

Contribution

The paper proposes novel SEA and DWA methods that significantly reduce latency and power consumption in floating-point CIM operations by exploiting exponent clustering.

Findings

Saves 63.8% power consumption compared to traditional methods.

Achieves 40.87% delay reduction on VGG16-CIFAR10.

Reduces input mantissa shifting and eliminates maximum exponent detection.

Abstract

With the rise of compute-in-memory (CIM) accelerators, floating-point multiply-and-accumulate (FP-MAC) operations have gained extensive attention for their higher accuracy over integer MACs in neural networks. However, the hardware overhead caused by exponent comparison and mantissa alignment, along with the delay introduced by bit-serial input methods, remains a hinder to implement FP-MAC efficiently. In view of this, we propose Segmented Exponent Alignment (SEA) and Dynamic Wordline Activation (DWA) strategies. SEA exploits the observation that input exponents are often clustered around zero or within a narrow range. By segmenting the exponent space and aligning mantissas accordingly, SEA eliminates the need for maximum exponent detection and reduces input mantissa shifting, and thus reduces the processing latency. DWA further reduces latency and maintains accuracy by activating wordlines based on the exponent segments defined by SEA. Simulation results demonstrate that, when compared with conventional comparison tree based maximum exponent alignment method, our approach saves 63.8\% power consumption, and achieves a 40.87\% delay reduction on the VGG16-CIFAR10 benchmark.