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

Nowadays, GPU accelerators are commonly used to speed up general-purposecomputing tasks on a variety of hardware. However, due to the diversity of GPUarchitectures and processed data, optimization of codes for a particular typeof hardware and specific data characteristics can be extremely challenging. Theautotuning of performance-relevant source-code parameters allows for automaticoptimization of applications and keeps their performance portable. Although theautotuning process typically results in code speed-up, searching the tuningspace can bring unacceptable overhead if (i) the tuning space is vast and fullof poorly-performing implementations, or (ii) the autotuning process has to berepeated frequently because of changes in processed data or migration todifferent hardware.  In this paper, we introduce a novel method for searching tuning spaces. Themethod takes advantage of collecting hardware performance counters (also knownas profiling counters) during empirical tuning. Those counters are used tonavigate the searching process towards faster implementations. The methodrequires the tuning space to be sampled on any GPU. It builds aproblem-specific model, which can be used during autotuning on various, evenpreviously unseen inputs or GPUs. Using a set of five benchmarks, weexperimentally demonstrate that our method can speed up autotuning when anapplication needs to be ported to different hardware or when it needs toprocess data with different characteristics. We also compared our method tostate of the art and show that our method is superior in terms of the number ofsearching steps and typically outperforms other searches in terms ofconvergence time.

Using hardware performance counters to speed up autotuning convergence on GPUs