English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The excellent performance of modern deep neural networks (DNNs) comes at anoften prohibitive training cost, limiting the rapid development of DNNinnovations and raising various environmental concerns. To reduce the dominantdata movement cost of training, process in-memory (PIM) has emerged as apromising solution as it alleviates the need to access DNN weights. However,state-of-the-art PIM DNN training accelerators employ either analog/mixedsignal computing which has limited precision or digital computing based on amemory technology that supports limited logic functions and thus requirescomplicated procedure to realize floating point computation. In this paper, wepropose a spin orbit torque magnetic random access memory (SOT-MRAM) baseddigital PIM accelerator that supports floating point precision. Specifically,this new accelerator features an innovative (1) SOT-MRAM cell, (2) fulladdition design, and (3) floating point computation. Experiment results showthat the proposed SOT-MRAM PIM based DNN training accelerator can achieve3.3$\times$, 1.8$\times$, and 2.5$\times$ improvement in terms of energy,latency, and area, respectively, compared with a state-of-the-art PIM based DNNtraining accelerator.

A New MRAM-based Process In-Memory Accelerator for Efficient Neural Network Training with Floating Point Precision