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

Convolutional neural networks (CNN) have become a ubiquitous algorithm withgrowing applications in mobile and edge settings. We describe acompute-in-memory (CIM) technique called FPIRM using Racetrack Memory (RM) toaccelerate CNNs for edge systems. Using transverse read, a technique that candetermine the number of '1's multiple adjacent domains, FPIRM can efficientlyimplement multi-operand bulk-bitwise and addition computations, and two-operandmultiplication. We discuss how FPIRM can implement both variable precisioninteger and floating point arithmetic. This allows both CNN inference andon-device training without expensive data movement to the cloud. Based on thesefunctions we demonstrate implementation of several CNNs with back propagationusing RM CIM and compare these to state-of-the-art implementations of CIMinference and training in Field-Programmable Gate Arrays. During training FPIRMimproves by 2$\times$ the efficiency, by reducing the energy consumption by atleast 27% and increasing the throughput by at least 18% against FPGA.

FPIRM: Floating-point Processing in Racetrack Memories