RED: A ReRAM-Based Efficient Accelerator for Deconvolutional Computation

TitleRED: A ReRAM-Based Efficient Accelerator for Deconvolutional Computation
Publication TypeJournal Article
Year of Publication2020
AuthorsZ Li, B Li, Z Fan, and H Li
JournalIeee Transactions on Computer Aided Design of Integrated Circuits and Systems
Volume39
Start Page4736
Issue12
Pagination4736 - 4747
Date Published12/2020
Abstract

Deconvolution is a key component in contemporary neural networks, especially, generative adversarial networks (GANs) and fully convolutional networks (FCNs). Due to extra operations of deconvolution compared to convolution, considerable degradation of performance, as well as energy efficiency is incurred when implementing deconvolution on the existing resistive random access memory (ReRAM)-based processing-in-memory (PIM) accelerators. In this article, we propose an ReRAM-based accelerator design, RED, for providing high-performance and low-energy deconvolution. We analyze the deconvolution execution on the existing ReRAM-based PIMs and utilize its interior computation pattern for design optimization. RED includes two major contributions: 1) pixel-wise mapping scheme and 2) zero-skipping data flow. Pixel-wise mapping scheme removes the zero insertion and performs convolutions over several ReRAM arrays and thus enables parallel computations with nonzero inputs. Zero-skipping data flow, assisted with customized input buffers design, enhances the computation parallelism and input data reuse. In evaluation, we compare RED against the existing ReRAM-based PIMs and CMOS-based counterpart with a variety of GAN and FCN models, each of which contains multiple deconvolution layers. The experimental results show that RED achieves a 4.0times - 56.16times speedup and a 1.05times - 18.17times energy efficiency improvement over previous related accelerator designs.

DOI10.1109/TCAD.2020.2981055
Short TitleIeee Transactions on Computer Aided Design of Integrated Circuits and Systems