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

Performance and energy are the two most important objectives for optimisationon modern parallel platforms. Latest research demonstrated the importance ofworkload distribution as a decision variable in the bi-objective optimisationfor performance and energy on homogeneous multicore clusters. We show in thiswork that bi-objective optimisation for performance and energy on heterogeneousprocessors results in a large number of Pareto-optimal optimal solutions(workload distributions) even in the simple case of linear performance andenergy profiles. We then study performance and energy profiles of real-lifedata-parallel applications and find that their shapes are non-linear, complexand non-smooth. We, therefore, propose an efficient and exact globaloptimisation algorithm, which takes as an input most general discreteperformance and dynamic energy profiles of the heterogeneous processors andsolves the bi-objective optimisation problem. The algorithm is also used as abuilding block to solve the bi-objective optimisation problem for performanceand total energy. We also propose a novel methodology to build discrete dynamicenergy profiles of individual computing devices, which are input to thealgorithm. The methodology is based purely on system-level measurements andaddresses the fundamental challenge of accurate component-level energymodelling of a hybrid data-parallel application running on a heterogeneousplatform integrating CPUs and accelerators. We experimentally validate theproposed method using two data-parallel applications, matrix multiplication and2D fast Fourier transform (2D-FFT).

Bi-objective Optimisation of Data-parallel Applications on Heterogeneous Platforms for Performance and Energy via Workload Distribution