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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Heterogeneous MPSoCs comprise diverse processing units of varying computecapabilities. To date, the mapping strategies of neural networks (NNs) ontosuch systems are yet to exploit the full potential of processing parallelism,made possible through both the intrinsic NNs' structure and underlying hardwarecomposition. In this paper, we propose a novel framework to effectively map NNsonto heterogeneous MPSoCs in a manner that enables them to leverage theunderlying processing concurrency. Specifically, our approach identifies anoptimal partitioning scheme of the NN along its `width' dimension, whichfacilitates deployment of concurrent NN blocks onto different hardwarecomputing units. Additionally, our approach contributes a novel scheme todeploy partitioned NNs onto the MPSoC as dynamic multi-exit networks foradditional performance gains. Our experiments on a standard MPSoC platform haveyielded dynamic mapping configurations that are 2.1x more energy-efficient thanthe GPU-only mapping while incurring 1.7x less latency than DLA-only mapping.

Map-and-Conquer: Energy-Efficient Mapping of Dynamic Neural Nets onto Heterogeneous MPSoCs