Application of alternating decision trees in selecting sparse linear solvers

Sanjukta Bhowmick, Victor Eijkhout, Yoav Freund, Erika Fuentes, David Keyes

Research output: Chapter in Book/Report/Conference proceedingChapter

7 Citations (Scopus)

Abstract

The solution of sparse linear systems, a fundamental and resource-intensive task in scientific computing, can be approached through multiple algorithms. Using an algorithm well adapted to characteristics of the task can significantly enhance the performance, such as reducing the time required for the operation, without compromising the quality of the result. However, the best solution method can vary even across linear systems generated in course of the same PDE-based simulation, thereby making solver selection a very challenging problem. In this paper, we use a machine learning technique, Alternating Decision Trees (ADT), to select efficient solvers based on the properties of sparse linear systems and runtime-dependent features, such as the stages of simulation. We demonstrate the effectiveness of this method through empirical results over linear systems drawn from computational fluid dynamics and magnetohydrodynamics applications. The results also demonstrate that using ADT can resolve the problem of over-fitting, which occurs when limited amount of data is available.

Original languageEnglish (US)
Title of host publicationSoftware Automatic Tuning
Subtitle of host publicationFrom Concepts to State-of-the-Art Results
PublisherSpringer New York
Pages153-173
Number of pages21
ISBN (Print)9781441969347
DOIs
StatePublished - Dec 1 2010

Fingerprint

Decision trees
Linear systems
Natural sciences computing
Magnetohydrodynamics
Learning systems
Computational fluid dynamics

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Bhowmick, S., Eijkhout, V., Freund, Y., Fuentes, E., & Keyes, D. (2010). Application of alternating decision trees in selecting sparse linear solvers. In Software Automatic Tuning: From Concepts to State-of-the-Art Results (pp. 153-173). Springer New York. https://doi.org/10.1007/978-1-4419-6935-4_10

Application of alternating decision trees in selecting sparse linear solvers. / Bhowmick, Sanjukta; Eijkhout, Victor; Freund, Yoav; Fuentes, Erika; Keyes, David.

Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York, 2010. p. 153-173.

Research output: Chapter in Book/Report/Conference proceedingChapter

Bhowmick, S, Eijkhout, V, Freund, Y, Fuentes, E & Keyes, D 2010, Application of alternating decision trees in selecting sparse linear solvers. in Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York, pp. 153-173. https://doi.org/10.1007/978-1-4419-6935-4_10
Bhowmick S, Eijkhout V, Freund Y, Fuentes E, Keyes D. Application of alternating decision trees in selecting sparse linear solvers. In Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York. 2010. p. 153-173 https://doi.org/10.1007/978-1-4419-6935-4_10
Bhowmick, Sanjukta ; Eijkhout, Victor ; Freund, Yoav ; Fuentes, Erika ; Keyes, David. / Application of alternating decision trees in selecting sparse linear solvers. Software Automatic Tuning: From Concepts to State-of-the-Art Results. Springer New York, 2010. pp. 153-173
@inbook{cfd881d83d0b4cd0a756ce2a26b79e6d,
title = "Application of alternating decision trees in selecting sparse linear solvers",
abstract = "The solution of sparse linear systems, a fundamental and resource-intensive task in scientific computing, can be approached through multiple algorithms. Using an algorithm well adapted to characteristics of the task can significantly enhance the performance, such as reducing the time required for the operation, without compromising the quality of the result. However, the best solution method can vary even across linear systems generated in course of the same PDE-based simulation, thereby making solver selection a very challenging problem. In this paper, we use a machine learning technique, Alternating Decision Trees (ADT), to select efficient solvers based on the properties of sparse linear systems and runtime-dependent features, such as the stages of simulation. We demonstrate the effectiveness of this method through empirical results over linear systems drawn from computational fluid dynamics and magnetohydrodynamics applications. The results also demonstrate that using ADT can resolve the problem of over-fitting, which occurs when limited amount of data is available.",
author = "Sanjukta Bhowmick and Victor Eijkhout and Yoav Freund and Erika Fuentes and David Keyes",
year = "2010",
month = "12",
day = "1",
doi = "10.1007/978-1-4419-6935-4_10",
language = "English (US)",
isbn = "9781441969347",
pages = "153--173",
booktitle = "Software Automatic Tuning",
publisher = "Springer New York",

}

TY - CHAP

T1 - Application of alternating decision trees in selecting sparse linear solvers

AU - Bhowmick, Sanjukta

AU - Eijkhout, Victor

AU - Freund, Yoav

AU - Fuentes, Erika

AU - Keyes, David

PY - 2010/12/1

Y1 - 2010/12/1

N2 - The solution of sparse linear systems, a fundamental and resource-intensive task in scientific computing, can be approached through multiple algorithms. Using an algorithm well adapted to characteristics of the task can significantly enhance the performance, such as reducing the time required for the operation, without compromising the quality of the result. However, the best solution method can vary even across linear systems generated in course of the same PDE-based simulation, thereby making solver selection a very challenging problem. In this paper, we use a machine learning technique, Alternating Decision Trees (ADT), to select efficient solvers based on the properties of sparse linear systems and runtime-dependent features, such as the stages of simulation. We demonstrate the effectiveness of this method through empirical results over linear systems drawn from computational fluid dynamics and magnetohydrodynamics applications. The results also demonstrate that using ADT can resolve the problem of over-fitting, which occurs when limited amount of data is available.

AB - The solution of sparse linear systems, a fundamental and resource-intensive task in scientific computing, can be approached through multiple algorithms. Using an algorithm well adapted to characteristics of the task can significantly enhance the performance, such as reducing the time required for the operation, without compromising the quality of the result. However, the best solution method can vary even across linear systems generated in course of the same PDE-based simulation, thereby making solver selection a very challenging problem. In this paper, we use a machine learning technique, Alternating Decision Trees (ADT), to select efficient solvers based on the properties of sparse linear systems and runtime-dependent features, such as the stages of simulation. We demonstrate the effectiveness of this method through empirical results over linear systems drawn from computational fluid dynamics and magnetohydrodynamics applications. The results also demonstrate that using ADT can resolve the problem of over-fitting, which occurs when limited amount of data is available.

UR - http://www.scopus.com/inward/record.url?scp=84870911452&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84870911452&partnerID=8YFLogxK

U2 - 10.1007/978-1-4419-6935-4_10

DO - 10.1007/978-1-4419-6935-4_10

M3 - Chapter

AN - SCOPUS:84870911452

SN - 9781441969347

SP - 153

EP - 173

BT - Software Automatic Tuning

PB - Springer New York

ER -