Using hardware performance counters to speed up autotuning convergence on GPUs

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216224%3A14610%2F22%3A00125022" target="_blank" >RIV/00216224:14610/22:00125022 - isvavai.cz</a>

Výsledek na webu

<a href="https://www.sciencedirect.com/science/article/pii/S0743731521001945?via%3Dihub" target="_blank" >https://www.sciencedirect.com/science/article/pii/S0743731521001945?via%3Dihub</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1016/j.jpdc.2021.10.003" target="_blank" >10.1016/j.jpdc.2021.10.003</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Using hardware performance counters to speed up autotuning convergence on GPUs

Popis výsledku v původním jazyce

Nowadays, GPU accelerators are commonly used to speed up general-purpose computing tasks on a variety of hardware. However, due to the diversity of GPU architectures and processed data, optimization of codes for a particular type of hardware and specific data characteristics can be extremely challenging. The autotuning of performance-relevant source-code parameters allows for automatic optimization of applications and keeps their performance portable. Although the autotuning process typically results in code speed-up, searching the tuning space can bring unacceptable overhead if (i) the tuning space is vast and full of poorly-performing implementations, or (ii) the autotuning process has to be repeated frequently because of changes in processed data or migration to different hardware. In this paper, we introduce a novel method for searching generic tuning spaces. The tuning spaces can contain tuning parameters changing any user-defined property of the source code. The method takes advantage of collecting hardware performance counters (also known as profiling counters) during empirical tuning. Those counters are used to navigate the searching process towards faster implementations. The method requires the tuning space to be sampled on any GPU. It builds a problem-specific model, which can be used during autotuning on various, even previously unseen inputs or GPUs. Using a set of five benchmarks, we experimentally demonstrate that our method can speed up autotuning when an application needs to be ported to different hardware or when it needs to process data with different characteristics. We also compared our method to state of the art and show that our method is superior in terms of the number of searching steps and typically outperforms other searches in terms of convergence time.

Název v anglickém jazyce

Using hardware performance counters to speed up autotuning convergence on GPUs

Popis výsledku anglicky

Nowadays, GPU accelerators are commonly used to speed up general-purpose computing tasks on a variety of hardware. However, due to the diversity of GPU architectures and processed data, optimization of codes for a particular type of hardware and specific data characteristics can be extremely challenging. The autotuning of performance-relevant source-code parameters allows for automatic optimization of applications and keeps their performance portable. Although the autotuning process typically results in code speed-up, searching the tuning space can bring unacceptable overhead if (i) the tuning space is vast and full of poorly-performing implementations, or (ii) the autotuning process has to be repeated frequently because of changes in processed data or migration to different hardware. In this paper, we introduce a novel method for searching generic tuning spaces. The tuning spaces can contain tuning parameters changing any user-defined property of the source code. The method takes advantage of collecting hardware performance counters (also known as profiling counters) during empirical tuning. Those counters are used to navigate the searching process towards faster implementations. The method requires the tuning space to be sampled on any GPU. It builds a problem-specific model, which can be used during autotuning on various, even previously unseen inputs or GPUs. Using a set of five benchmarks, we experimentally demonstrate that our method can speed up autotuning when an application needs to be ported to different hardware or when it needs to process data with different characteristics. We also compared our method to state of the art and show that our method is superior in terms of the number of searching steps and typically outperforms other searches in terms of convergence time.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Projekt

<a href="/cs/project/EF16_013%2F0001802" target="_blank" >EF16_013/0001802: CERIT Scientific Cloud</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Journal of Parallel and Distributed Computing

ISSN

0743-7315

e-ISSN

—

Svazek periodika

160

Číslo periodika v rámci svazku

February

Stát vydavatele periodika

NL - Nizozemsko

Počet stran výsledku

20

Strana od-do

16-35

Kód UT WoS článku

000711621300002

EID výsledku v databázi Scopus

2-s2.0-85117699383