ARAS: An Adaptive Low-Cost ReRAM-Based Accelerator for DNNs

TL;DR

ARAS is an adaptive, low-cost ReRAM-based accelerator for DNN inference that employs smart scheduling and weight re-encoding to optimize performance and energy efficiency, addressing the high write latency and resource limitations of ReRAM crossbars.

Contribution

It introduces ARAS, a novel ReRAM accelerator with adaptive scheduling and weight re-encoding to improve DNN inference efficiency and reduce energy consumption.

Findings

Up to 2.2x speedup over baseline PUM accelerators.

45% energy savings compared to baseline.

Up to 1.5x speedup and 61% energy savings over TPU-like accelerators.

Abstract

Processing Using Memory (PUM) accelerators have the potential to perform Deep Neural Network (DNN) inference by using arrays of memory cells as computation engines. Among various memory technologies, ReRAM crossbars show promising performance in computing dot-product operations in the analog domain. Nevertheless, the expensive writing procedure of ReRAM cells has led researchers to design accelerators whose crossbars have enough capacity to store the full DNN. Given the tremendous and continuous increase in DNN model sizes, this approach is unfeasible for some networks, or inefficient due to the huge hardware requirements. Those accelerators lack the flexibility to adapt to any given DNN model, facing an challenge. To address this issue we introduce ARAS, a cost-effective ReRAM-based accelerator that employs a smart scheduler to adapt different DNNs to the resource-limited hardware. ARAS also overlaps the computation of a layer with the weight writing of several layers to mitigate the high writing latency of ReRAM. Furthermore, ARAS introduces three optimizations aimed at reducing the energy overheads of writing in ReRAM. Our key optimization capitalizes on the observation that DNN weights can be re-encoded to augment their similarity between layers, increasing the amount of bitwise values that are equal or similar when overwriting ReRAM cells and, hence, reducing the amount of energy required to update the cells. Overall, ARAS greatly reduces the ReRAM writing activity. We evaluate ARAS on a popular set of DNNs. ARAS provides up to 2.2x speedup and 45% energy savings over a baseline PUM accelerator without any optimization. Compared to a TPU-like accelerator, ARAS provides up to 1.5x speedup and 61% energy savings.