Estimating resource budgets to ensure autotuning efficiency

J 2025

Estimating resource budgets to ensure autotuning efficiency

OĽHA, Jaroslav; Jana HOZZOVÁ; Matej ANTOL a Jiří FILIPOVIČ

Základní údaje

Originální název

Estimating resource budgets to ensure autotuning efficiency

Autoři

OĽHA, Jaroslav; Jana HOZZOVÁ; Matej ANTOL a Jiří FILIPOVIČ

Vydání

PARALLEL COMPUTING, NETHERLANDS, ELSEVIER, 2025, 0167-8191

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: 2.100 v roce 2024

Organizační jednotka

Ústav výpočetní techniky

DOI

https://doi.org/10.1016/j.parco.2025.103126

UT WoS

001437878000001

Klíčová slova anglicky

Autotuning; Dynamic autotuning; Stopping condition; Tuning budget estimation; Regression models

Štítky

J-Q2, rivok

Příznaky

Mezinárodní význam, Recenzováno

Změněno: 20. 3. 2025 09:21, doc. RNDr. Jiří Filipovič, Ph.D.

Anotace

V originále

Many state-of-the-art HPC applications rely on autotuning to maintain peak performance. Autotuning allows a program to be re-optimized for new hardware, settings, or input — even during execution. However, the approach has an inherent problem that has yet to be properly addressed: since the autotuning process itself requires computational resources, it is also subject to optimization. In other words, while autotuning aims to decrease a program’s run time by improving its efficiency, it also introduces additional overhead that can extend the overall run time. To achieve optimal performance, both the application and the autotuning process should be optimized together, treating them as a single optimization criterion. This framing allows us to determine a reasonable tuning budget to avoid both undertuning, where insufficient autotuning leads to suboptimal performance, and overtuning, where excessive autotuning imposes overhead that outweighs the benefits of program optimization. In this paper, we explore the tuning budget optimization problem in detail, highlighting its interesting properties and implications, which have largely been overlooked in the literature. Additionally, we present several viable solutions for tuning budget optimization and evaluate their efficiency across a range of commonly used HPC kernels.

Návaznosti

LM2023054, projekt VaV

Název: e-Infrastruktura CZ

Investor: Ministerstvo školství, mládeže a tělovýchovy ČR, e-Infrastruktura CZ

Citovat

OĽHA, Jaroslav; Jana HOZZOVÁ; Matej ANTOL a Jiří FILIPOVIČ. Estimating resource budgets to ensure autotuning efficiency. PARALLEL COMPUTING. NETHERLANDS: ELSEVIER, 2025, roč. 123, March, s. 1-15. ISSN 0167-8191. Dostupné z: https://doi.org/10.1016/j.parco.2025.103126.

@article{2478850,
   author = {Oľha, Jaroslav and Hozzová, Jana and Antol, Matej and Filipovič, Jiří},
   article_location = {NETHERLANDS},
   article_number = {March},
   doi = {https://doi.org/10.1016/j.parco.2025.103126},
   keywords = {Autotuning; Dynamic autotuning; Stopping condition; Tuning budget estimation; Regression models},
   language = {eng},
   issn = {0167-8191},
   journal = {PARALLEL COMPUTING},
   title = {Estimating resource budgets to ensure autotuning efficiency},
   url = {https://www.sciencedirect.com/journal/parallel-computing/vol/123/suppl/C},
   volume = {123},
   year = {2025}
}

TY  - JOUR
ID  - 2478850
AU  - Oľha, Jaroslav - Hozzová, Jana - Antol, Matej - Filipovič, Jiří
PY  - 2025
TI  - Estimating resource budgets to ensure autotuning efficiency
JF  - PARALLEL COMPUTING
VL  - 123
IS  - March
SP  - 1-15
EP  - 1-15
PB  - ELSEVIER
SN  - 01678191
KW  - Autotuning
KW  - Dynamic autotuning
KW  - Stopping condition
KW  - Tuning budget estimation
KW  - Regression models
UR  - https://www.sciencedirect.com/journal/parallel-computing/vol/123/suppl/C
N2  - Many state-of-the-art HPC applications rely on autotuning to maintain peak performance. Autotuning allows a program to be re-optimized for new hardware, settings, or input — even during execution. However, the approach has an inherent problem that has yet to be properly addressed: since the autotuning process itself requires computational resources, it is also subject to optimization. In other words, while autotuning aims to decrease a program’s run time by improving its efficiency, it also introduces additional overhead that can extend the overall run time. To achieve optimal performance, both the application and the autotuning process should be optimized together, treating them as a single optimization criterion. This framing allows us to determine a reasonable tuning budget to avoid both undertuning, where insufficient autotuning leads to suboptimal performance, and overtuning, where excessive autotuning imposes overhead that outweighs the benefits of program optimization. In this paper, we explore the tuning budget optimization problem in detail, highlighting its interesting properties and implications, which have largely been overlooked in the literature. Additionally, we present several viable solutions for tuning budget optimization and evaluate their efficiency across a range of commonly used HPC kernels.
ER  -

OĽHA, Jaroslav; Jana HOZZOVÁ; Matej ANTOL a Jiří FILIPOVIČ. Estimating resource budgets to ensure autotuning efficiency. \textit{PARALLEL COMPUTING}. NETHERLANDS: ELSEVIER, 2025, roč.~123, March, s.~1-15. ISSN~0167-8191. Dostupné z: https://doi.org/10.1016/j.parco.2025.103126.

Přehled o publikaci