Heterogeneous Multi-core Array-based DNN Accelerator

TL;DR

This paper analyzes how architectural parameters affect energy and performance in array-based DNN accelerators, proposing a heterogeneous multi-core scheme and algorithms to optimize energy efficiency and processing speed.

Contribution

It introduces a simulation tool for architectural analysis and proposes a heterogeneous multi-core scheme with algorithms for efficient neural network execution.

Findings

Up to 36% energy savings with near-optimal core configurations.

Up to 67% reduction in energy-delay product.

Effective layer distribution algorithm for faster processing.

Abstract

In this article, we investigate the impact of architectural parameters of array-based DNN accelerators on accelerator's energy consumption and performance in a wide variety of network topologies. For this purpose, we have developed a tool that simulates the execution of neural networks on array-based accelerators and has the capability of testing different configurations for the estimation of energy consumption and processing latency. Based on our analysis of the behavior of benchmark networks under different architectural parameters, we offer a few recommendations for having an efficient yet high performance accelerator design. Next, we propose a heterogeneous multi-core chip scheme for deep neural network execution. The evaluations of a selective small search space indicate that the execution of neural networks on their near-optimal core configuration can save up to 36% and 67% of energy consumption and energy-delay product respectively. Also, we suggest an algorithm to distribute the processing of network's layers across multiple cores of the same type in order to speed up the computations through model parallelism. Evaluations on different networks and with the different number of cores verify the effectiveness of the proposed algorithm in speeding up the processing to near-optimal values.