MemTorch: an open-source simulation framework for memristive deep learning systems

Journal Publication ResearchOnline@JCU
Lammie, Corey;Xiang, Wei;Linares-Barranco, Bernabé;Rahimi Azghadi, Mostafa
Abstract

Memristive devices have shown great promise to facilitate the acceleration and improve the power efficiency of Deep Learning (DL) systems. Crossbar architectures constructed using these Resistive Random-Access Memory(RRAM) devices can be used to efficiently implement various in-memory computing operations, such as Multiply Accumulate (MAC) and unrolled-convolutions, which are used extensively in Deep Neural Network(DNN) and Convolutional Neural Network (CNN). However, memristive devices face concerns of aging and non-idealities, which limit the accuracy, reliability, and robustness of Memristive Deep Learning System(MDLS), that should be considered prior to circuit-level realization. This Original Software Publication(OSP) presents MemTorch, an open-source1 framework for customized large-scale memristive Deep Learning(DL) simulations, with a refined focus on the co-simulation of device non-idealities. MemTorch also facilitates co-modelling of key crossbar peripheral circuitry. MemTorch adopts a modernized software engineering methodology and integrates directly with the well-known PyTorch Machine Learning(ML) library.

Journal

Neurocomputing

Publication Name

N/A

Volume

485

ISBN/ISSN

1872-8286

Edition

N/A

Issue

N/A

Pages Count

5

Location

N/A

Publisher

Elsevier

Publisher Url

N/A

Publisher Location

N/A

Publish Date

N/A

Url

N/A

Date

N/A

EISSN

N/A

DOI

10.1016/j.neucom.2022.02.043