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

Autotuning techniques are a promising approach to minimize the otherwisetedious manual effort of optimizing scientific applications for a specifictarget platform. Ideally, an autotuning approach is capable of reliablyidentifying the most efficient implementation variant(s) for a new targetsystem or new characteristics of the input by applying suitable programtransformations and analytic models. In this work, we introduce Offsite, anoffline autotuning approach which automates this selection process atinstallation time by rating implementation variants based on an analyticperformance model without requiring time-consuming runtime experiments. Fromabstract multilevel YAML description languages, Offsite automatically derivesoptimized, platform-specific and problem-specific code of possibleimplementation variants and applies the performance model to theseimplementation variants.  We apply Offsite to parallel numerical methods for ordinary differentialequations (ODEs). In particular, we investigate tuning a specific class ofexplicit ODE solvers (PIRK methods) for various initial value problems (IVPs)on shared-memory systems. Our experiments demonstrate that Offsite is able toreliably identify a set of the most efficient implementation variants for giventest configurations (ODE solver, IVP, platform) and is capable of effectivelyhandling important autotuning scenarios.

Offsite Autotuning Approach -- Performance Model Driven Autotuning Applied to Parallel Explicit ODE Methods