Towards a Benchmarking Suite for Kernel Tuners

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216224%3A14610%2F23%3A00131587" target="_blank" >RIV/00216224:14610/23:00131587 - isvavai.cz</a>

Výsledek na webu

<a href="http://dx.doi.org/10.1109/IPDPSW59300.2023.00124" target="_blank" >http://dx.doi.org/10.1109/IPDPSW59300.2023.00124</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1109/IPDPSW59300.2023.00124" target="_blank" >10.1109/IPDPSW59300.2023.00124</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Towards a Benchmarking Suite for Kernel Tuners

Popis výsledku v původním jazyce

As computing system become more complex combining CPUs and GPUs, it is becoming harder and harder for programmers to keep their codes optimized as the hardware gets updated. Autotuners try to alleviate this by hiding as many architecture-based optimization details as possible from the end-user, so that the code can be used efficiently across different generations of systems. Several autotuning frameworks have emerged, but a comparative analysis between these related works is scarce, owing to the significant manual effort required to port a tunable kernel from one tuner another. In this article we introduce a new benchmark suite for evaluating the performance of optimization algorithms used by modern autotuners targeting GPUs. The suite contains tunable GPU kernels that are representative of real-world applications, allowing for comparisons between optimization algorithms and the examination of code optimization, search space difficulty, and performance portability. Our framework facilitates easy integration of new autotuners and benchmarks by defining a shared problem interface. Our benchmark suite is evaluated based on five characteristics: convergence rate, local minima centrality, optimal speedup, Permutation Feature Importance (PFI), and performance portability. The results show that optimization parameters greatly impact performance and the need for global optimization. The importance of each parameter is consistent across GPU architectures, however, the specific values need to be optimized for each architecture. Our portability study highlights the crucial importance of autotuning each application for a specific target architecture. The results reveal that simply transferring the optimal configuration from one architecture to another can result in a performance ranging from 58.5% to 99.9% of the optimal performance, depending on the GPU architecture. This highlights the importance of autotuning in modern computing systems and the value of our benchmark suite in facilitating the study of optimization algorithms and their effectiveness in achieving optimal performance for specific target architectures.

Název v anglickém jazyce

Towards a Benchmarking Suite for Kernel Tuners

Popis výsledku anglicky

As computing system become more complex combining CPUs and GPUs, it is becoming harder and harder for programmers to keep their codes optimized as the hardware gets updated. Autotuners try to alleviate this by hiding as many architecture-based optimization details as possible from the end-user, so that the code can be used efficiently across different generations of systems. Several autotuning frameworks have emerged, but a comparative analysis between these related works is scarce, owing to the significant manual effort required to port a tunable kernel from one tuner another. In this article we introduce a new benchmark suite for evaluating the performance of optimization algorithms used by modern autotuners targeting GPUs. The suite contains tunable GPU kernels that are representative of real-world applications, allowing for comparisons between optimization algorithms and the examination of code optimization, search space difficulty, and performance portability. Our framework facilitates easy integration of new autotuners and benchmarks by defining a shared problem interface. Our benchmark suite is evaluated based on five characteristics: convergence rate, local minima centrality, optimal speedup, Permutation Feature Importance (PFI), and performance portability. The results show that optimization parameters greatly impact performance and the need for global optimization. The importance of each parameter is consistent across GPU architectures, however, the specific values need to be optimized for each architecture. Our portability study highlights the crucial importance of autotuning each application for a specific target architecture. The results reveal that simply transferring the optimal configuration from one architecture to another can result in a performance ranging from 58.5% to 99.9% of the optimal performance, depending on the GPU architecture. This highlights the importance of autotuning in modern computing systems and the value of our benchmark suite in facilitating the study of optimization algorithms and their effectiveness in achieving optimal performance for specific target architectures.

Druh

D - Stať ve sborníku

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/LM2023054" target="_blank" >LM2023054: e-Infrastruktura CZ</a><br>

Návaznosti

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

Rok uplatnění

2023

Kód důvěrnosti údajů

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

Název statě ve sborníku

2023 IEEE International Parallel and Distributed Processing Symposium Workshops (IPDPSW)

ISBN

9798350311990

ISSN

2164-7062

e-ISSN

—

Počet stran výsledku

10

Strana od-do

724-733

Název nakladatele

IEEE

Místo vydání

neuveden

Místo konání akce

St. Petersburg, FL, USA

Datum konání akce

1. 1. 2023

Typ akce podle státní příslušnosti

WRD - Celosvětová akce

Kód UT WoS článku

001055030700088