Using hardware performance counters to speed up autotuning convergence on GPUs

FILIPOVIČ, Jiří, Jana HOZZOVÁ, Amin NEZARAT, Jaroslav OĽHA a Filip PETROVIČ. Using hardware performance counters to speed up autotuning convergence on GPUs. Journal of Parallel and Distributed Computing. Elsevier, 2022, roč. 160, February, s. 16-35. ISSN 0743-7315. Dostupné z: https://dx.doi.org/10.1016/j.jpdc.2021.10.003.

Základní údaje

Originální název

Using hardware performance counters to speed up autotuning convergence on GPUs

Autoři

FILIPOVIČ, Jiří (203 Česká republika, garant, domácí), Jana HOZZOVÁ (703 Slovensko, domácí), Amin NEZARAT (364 Írán, domácí), Jaroslav OĽHA (703 Slovensko, domácí) a Filip PETROVIČ (703 Slovensko, domácí)

Vydání

Journal of Parallel and Distributed Computing, Elsevier, 2022, 0743-7315

---

Další údaje

Jazyk

angličtina

Typ výsledku

Článek v odborném periodiku

Obor

10201 Computer sciences, information science, bioinformatics

Stát vydavatele

Nizozemské království

Utajení

není předmětem státního či obchodního tajemství

Odkazy

URL

Impakt faktor

Impact factor: 3.800

Kód RIV

RIV/00216224:14610/22:00125022

Organizační jednotka

Ústav výpočetní techniky

DOI

http://dx.doi.org/10.1016/j.jpdc.2021.10.003

UT WoS

000711621300002

Klíčová slova anglicky

Auto-tuning; Search method; Performance counters; CUDA

Štítky

J-Q1, rivok

Příznaky

Mezinárodní význam, Recenzováno

---

Anotace

V originále

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.

---

Návaznosti

EF16_013/0001802, projekt VaV	Název: CERIT Scientific Cloud
MUNI/A/1145/2021, interní kód MU	Název: Rozsáhlé výpočetní systémy: modely, aplikace a verifikace XI. (Akronym: SV-FI MAV XI.) Investor: Masarykova univerzita, Rozsáhlé výpočetní systémy: modely, aplikace a verifikace XI.