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

Application migration and dynamic voltage and frequency scaling (DVFS) areindispensable means for fully exploiting the available potential in thermaloptimization of a heterogeneous clustered multi-core processor underuser-defined quality of service (QoS) targets. However, selecting the core toexecute each application and the voltage/frequency (V/f) levels of each clusteris a complex problem because 1) the diverse characteristics and QoS targets ofapplications require different optimizations, and 2) per-cluster DVFS requiresa global optimization considering all running applications. State-of-the-artresource management techniques for power or temperature minimization eitherrely on measurements that are often not available (such as power) or fail toconsider all the dimensions of the problem (e.g., by using simplifiedanalytical models). Imitation learning (IL) enables to use the optimality of anoracle policy, yet at low run-time overhead, by training a model from oracledemonstrations. We are the first to employ IL for temperature minimizationunder QoS targets. We tackle the complexity by training a neural network (NN)and accelerate the NN inference using a neural processing unit (NPU). Whilesuch NN accelerators are becoming increasingly widespread on end devices, theyare so far only used to accelerate user applications. In contrast, we use anexisting accelerator on a real platform to accelerate NN-based resourcemanagement. Our evaluation on a HiKey 970 board with an Arm big.LITTLE CPU andan NPU shows significant temperature reductions at a negligible run-timeoverhead, with unseen applications and different cooling than used fortraining.

NPU-Accelerated Imitation Learning for Thermal Optimization of QoS-Constrained Heterogeneous Multi-Cores