TU Darmstadt / ULB / TUbiblio

Extracting Clean Performance Models from Tainted Programs

Copik, Marcin ; Calotoiu, Alexandru ; Grosser, Tobias ; Wicki, Nicolas ; Wolf, Felix ; Hoefler, Torsten (2021)
Extracting Clean Performance Models from Tainted Programs.
26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming. Seoul, South Korea (27. 02.-03.03.2021)
doi: 10.1145/3437801.3441613
Conference or Workshop Item, Bibliographie

Abstract

Performance models are well-known instruments to understand the scaling behavior of parallel applications. They express how performance changes as key execution parameters, such as the number of processes or the size of the input problem, vary. Besides reasoning about program behavior, such models can also be automatically derived from performance data. This is called empirical performance modeling. While this sounds simple at the first glance, this approach faces several serious interrelated challenges, including expensive performance measurements, inaccuracies inflicted by noisy benchmark data, and overall complex experiment design, starting with the selection of the right parameters. The more parameters one considers, the more experiments are needed and the stronger the impact of noise. In this paper, we show how taint analysis, a technique borrowed from the domain of computer security, can substantially improve the modeling process, lowering its cost, improving model quality, and help validate performance models and experimental setups.

Item Type: Conference or Workshop Item
Erschienen: 2021
Creators: Copik, Marcin ; Calotoiu, Alexandru ; Grosser, Tobias ; Wicki, Nicolas ; Wolf, Felix ; Hoefler, Torsten
Type of entry: Bibliographie
Title: Extracting Clean Performance Models from Tainted Programs
Language: English
Date: 17 February 2021
Publisher: ACM
Book Title: PPoPP '21: Proceedings of the ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming
Event Title: 26th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming
Event Location: Seoul, South Korea
Event Dates: 27. 02.-03.03.2021
DOI: 10.1145/3437801.3441613
Corresponding Links:
Abstract:

Performance models are well-known instruments to understand the scaling behavior of parallel applications. They express how performance changes as key execution parameters, such as the number of processes or the size of the input problem, vary. Besides reasoning about program behavior, such models can also be automatically derived from performance data. This is called empirical performance modeling. While this sounds simple at the first glance, this approach faces several serious interrelated challenges, including expensive performance measurements, inaccuracies inflicted by noisy benchmark data, and overall complex experiment design, starting with the selection of the right parameters. The more parameters one considers, the more experiments are needed and the stronger the impact of noise. In this paper, we show how taint analysis, a technique borrowed from the domain of computer security, can substantially improve the modeling process, lowering its cost, improving model quality, and help validate performance models and experimental setups.

Uncontrolled Keywords: DFG|449683531, DFG|323299120, DFG
Divisions: 20 Department of Computer Science
20 Department of Computer Science > Parallel Programming
Date Deposited: 04 Apr 2024 10:00
Last Modified: 04 Apr 2024 10:00
PPN:
Corresponding Links:
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