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

Scientific applications often contain large, computationally-intensive, andirregular parallel loops or tasks that exhibit stochastic characteristics.Applications may suffer from load imbalance during their execution onhigh-performance computing (HPC) systems due to such characteristics. Dynamicloop self-scheduling (DLS) techniques are instrumental in improving theperformance of scientific applications on HPC systems via load balancing.Selecting a DLS technique that results in the best performance for differentproblems and system sizes requires a large number of exploratory experiments. Atheoretical model that can be used to predict the scheduling technique thatyields the best performance for a given problem and system has not yet beenidentified. Therefore, simulation is the most appropriate approach forconducting such exploratory experiments with reasonable costs. This workdevises an approach to realistically simulate computationally-intensivescientific applications that employ DLS and execute on HPC systems. Severalapproaches to represent the application tasks (or loop iterations) are comparedto establish their influence on the simulative application performance. A novelsimulation strategy is introduced, which transforms a native application codeinto a simulative code. The native and simulative performance of twocomputationally-intensive scientific applications are compared to evaluate therealism of the proposed simulation approach. The comparison of the performancecharacteristics extracted from the native and simulative performance shows thatthe proposed simulation approach fully captured most of the performancecharacteristics of interest. This work shows and establishes the importance ofsimulations that realistically predict the performance of DLS techniques fordifferent applications and system configurations.

An Approach for Realistically Simulating the Performance of Scientific Applications on High Performance Computing Systems