TU Darmstadt / ULB / TUbiblio

Noise-Resilient Empirical Performance Modeling with Deep Neural Networks

Ritter, Marcus ; Geiß, Alexander ; Wehrstein, Johannes ; Calotoiu, Alexandru ; Reimann, Thorsten ; Hoefler, Torsten ; Wolf, Felix (2021)
Noise-Resilient Empirical Performance Modeling with Deep Neural Networks.
35th International Parallel and Distributed Processing Symposium (IPDPS'21). virtual Conference (17.-21.05.2021)
doi: 10.1109/IPDPS49936.2021.00012
Conference or Workshop Item, Bibliographie

Abstract

Empirical performance modeling is a proven instrument to analyze the scaling behavior of HPC applications. Using a set of smaller-scale experiments, it can provide important insights into application behavior at larger scales. Extra-P is an empirical modeling tool that applies linear regression to automatically generate human-readable performance models. Similar to other regression-based modeling techniques, the accuracy of the models created by Extra-P decreases as the amount of noise in the underlying data increases. This is why the performance variability observed in many contemporary systems can become a serious challenge. In this paper, we introduce a novel adaptive modeling approach that makes Extra-P more noise resilient, exploiting the ability of deep neural networks to discover the effects of numerical parameters, such as the number of processes or the problem size, on performance when dealing with noisy measurements. Using synthetic analysis and data from three different case studies, we demonstrate that our solution improves the model accuracy at high noise levels by up to 25% while increasing their predictive power by about 15%.

Item Type: Conference or Workshop Item
Erschienen: 2021
Creators: Ritter, Marcus ; Geiß, Alexander ; Wehrstein, Johannes ; Calotoiu, Alexandru ; Reimann, Thorsten ; Hoefler, Torsten ; Wolf, Felix
Type of entry: Bibliographie
Title: Noise-Resilient Empirical Performance Modeling with Deep Neural Networks
Language: English
Date: 28 June 2021
Place of Publication: Piscataway, NJ
Publisher: IEEE
Book Title: Proceedings: 2021 35th International Parallel and Distributed Processing Symposium
Event Title: 35th International Parallel and Distributed Processing Symposium (IPDPS'21)
Event Location: virtual Conference
Event Dates: 17.-21.05.2021
DOI: 10.1109/IPDPS49936.2021.00012
Abstract:

Empirical performance modeling is a proven instrument to analyze the scaling behavior of HPC applications. Using a set of smaller-scale experiments, it can provide important insights into application behavior at larger scales. Extra-P is an empirical modeling tool that applies linear regression to automatically generate human-readable performance models. Similar to other regression-based modeling techniques, the accuracy of the models created by Extra-P decreases as the amount of noise in the underlying data increases. This is why the performance variability observed in many contemporary systems can become a serious challenge. In this paper, we introduce a novel adaptive modeling approach that makes Extra-P more noise resilient, exploiting the ability of deep neural networks to discover the effects of numerical parameters, such as the number of processes or the problem size, on performance when dealing with noisy measurements. Using synthetic analysis and data from three different case studies, we demonstrate that our solution improves the model accuracy at high noise levels by up to 25% while increasing their predictive power by about 15%.

Divisions: 20 Department of Computer Science
20 Department of Computer Science > Data and AI Systems
20 Department of Computer Science > Parallel Programming
Zentrale Einrichtungen
Zentrale Einrichtungen > University IT-Service and Computing Centre (HRZ)
Zentrale Einrichtungen > University IT-Service and Computing Centre (HRZ) > Hochleistungsrechner
Date Deposited: 02 Feb 2023 08:03
Last Modified: 13 Feb 2024 15:08
PPN: 507736818
Export:
Suche nach Titel in: TUfind oder in Google
Send an inquiry Send an inquiry

Options (only for editors)
Show editorial Details Show editorial Details