HURRY: Highly Utilized, Reconfigurable ReRAM-based In-situ Accelerator with Multifunctionality

TL;DR

HURRY is a reconfigurable ReRAM-based accelerator that improves utilization and efficiency for neural network inference by combining dynamic activation schemes and multifunctional computation blocks.

Contribution

It introduces a reconfigurable design with block activation and multifunctional units, enhancing utilization and performance over existing ReRAM accelerators.

Findings

Up to 3.35x speedup achieved

5.72x higher energy efficiency

7.91x greater area efficiency

Abstract

Resistive random-access memory (ReRAM) crossbar arrays are suitable for efficient inference computations in neural networks due to their analog general matrix-matrix multiplication (GEMM) capabilities. However, traditional ReRAM-based accelerators suffer from spatial and temporal underutilization. We present HURRY, a reconfigurable and multifunctional ReRAM-based in-situ accelerator. HURRY uses a block activation scheme for concurrent activation of dynamically sized ReRAM portions, enhancing spatial utilization. Additionally, it incorporates functional blocks for convolution, ReLU, max pooling, and softmax computations to improve temporal utilization. System-level scheduling and data mapping strategies further optimize performance. Consequently, HURRY achieves up to 3.35x speedup, 5.72x higher energy efficiency, and 7.91x greater area efficiency compared to current ReRAM-based accelerators.