PENDRAM: Enabling High-Performance and Energy-Efficient Processing of Deep Neural Networks through a Generalized DRAM Data Mapping Policy

TL;DR

PENDRAM introduces a generalized DRAM data mapping policy that optimizes CNN accelerator performance and energy efficiency by exploring various DRAM architectures and data mappings, significantly reducing energy-delay-product.

Contribution

It proposes a novel methodology for exploring DRAM data mappings and architectures to optimize CNN acceleration performance and energy efficiency.

Findings

DRAM data mapping policy reduces energy-delay-product by up to 96%.

Method enables high-performance, energy-efficient CNN acceleration across diverse DRAM architectures.

Provides a systematic approach for design space exploration in CNN hardware accelerators.

Abstract

Convolutional Neural Networks (CNNs), a prominent type of Deep Neural Networks (DNNs), have emerged as a state-of-the-art solution for solving machine learning tasks. To improve the performance and energy efficiency of CNN inference, the employment of specialized hardware accelerators is prevalent. However, CNN accelerators still face performance- and energy-efficiency challenges due to high off-chip memory (DRAM) access latency and energy, which are especially crucial for latency- and energy-constrained embedded applications. Moreover, different DRAM architectures have different profiles of access latency and energy, thus making it challenging to optimize them for high performance and energy-efficient CNN accelerators. To address this, we present PENDRAM, a novel design space exploration methodology that enables high-performance and energy-efficient CNN acceleration through a generalized DRAM data mapping policy. Specifically, it explores the impact of different DRAM data mapping policies and DRAM architectures across different CNN partitioning and scheduling schemes on the DRAM access latency and energy, then identifies the pareto-optimal design choices. The experimental results show that our DRAM data mapping policy improves the energy-delay-product of DRAM accesses in the CNN accelerator over other mapping policies by up to 96%. In this manner, our PENDRAM methodology offers high-performance and energy-efficient CNN acceleration under any given DRAM architectures for diverse embedded AI applications.