A Feasibility Study in Iterative Compilation

Toru Kisuki; Peter M. W. Knijnenburg; Michael F. P. O'Boyle; François Bodin; Harry A. G. Wijshoff

doi:10.1007/BFb0094916

A Feasibility Study in Iterative Compilation

Toru Kisuki
, Peter M. W. Knijnenburg
, Michael F. P. O'Boyle
, François Bodin
, Harry A. G. Wijshoff

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

In this paper we investigate the feasibility of iterative compilation in program optimisation. This technique enables compilers to deliver efficient code by searching for the best sequence of optimisations. In embedded systems, long compilation time can be afforded since the application is an integral part of the shipped product. However, in practice search spaces may be extremely large. Our experimental results show that in the case of large transformation spaces, near optimal transformations can be found by visiting only a small fraction of the entire search space by using a simple search algorithm.

Original language	English
Title of host publication	High Performance Computing
Subtitle of host publication	Second International Symposium, ISHPC'99 Kyoto, Japan, May 26–28, 1999 Proceedings
Publisher	Springer
Pages	121-132
Number of pages	12
ISBN (Electronic)	978-3-540-48821-7
ISBN (Print)	978-3-540-65969-3
DOIs	https://doi.org/10.1007/BFb0094916
Publication status	Published - 19 Oct 2006

Access to Document

10.1007/BFb0094916

http://link.springer.com/chapter/10.1007%2FBFb0094916

Cite this

@inproceedings{bff46f2dc7014d3c8ceec0120e004efa,

title = "A Feasibility Study in Iterative Compilation",

abstract = "In this paper we investigate the feasibility of iterative compilation in program optimisation. This technique enables compilers to deliver efficient code by searching for the best sequence of optimisations. In embedded systems, long compilation time can be afforded since the application is an integral part of the shipped product. However, in practice search spaces may be extremely large. Our experimental results show that in the case of large transformation spaces, near optimal transformations can be found by visiting only a small fraction of the entire search space by using a simple search algorithm.",

author = "Toru Kisuki and Knijnenburg, \{Peter M. W.\} and O'Boyle, \{Michael F. P.\} and Fran{\c c}ois Bodin and Wijshoff, \{Harry A. G.\}",

year = "2006",

month = oct,

day = "19",

doi = "10.1007/BFb0094916",

language = "English",

isbn = "978-3-540-65969-3",

pages = "121--132",

booktitle = "High Performance Computing",

publisher = "Springer",

address = "United Kingdom",

}

TY - GEN

T1 - A Feasibility Study in Iterative Compilation

AU - Kisuki, Toru

AU - Knijnenburg, Peter M. W.

AU - O'Boyle, Michael F. P.

AU - Bodin, François

AU - Wijshoff, Harry A. G.

PY - 2006/10/19

Y1 - 2006/10/19

N2 - In this paper we investigate the feasibility of iterative compilation in program optimisation. This technique enables compilers to deliver efficient code by searching for the best sequence of optimisations. In embedded systems, long compilation time can be afforded since the application is an integral part of the shipped product. However, in practice search spaces may be extremely large. Our experimental results show that in the case of large transformation spaces, near optimal transformations can be found by visiting only a small fraction of the entire search space by using a simple search algorithm.

AB - In this paper we investigate the feasibility of iterative compilation in program optimisation. This technique enables compilers to deliver efficient code by searching for the best sequence of optimisations. In embedded systems, long compilation time can be afforded since the application is an integral part of the shipped product. However, in practice search spaces may be extremely large. Our experimental results show that in the case of large transformation spaces, near optimal transformations can be found by visiting only a small fraction of the entire search space by using a simple search algorithm.

U2 - 10.1007/BFb0094916

DO - 10.1007/BFb0094916

M3 - Conference contribution

SN - 978-3-540-65969-3

SP - 121

EP - 132

BT - High Performance Computing

PB - Springer

ER -

A Feasibility Study in Iterative Compilation

Abstract

Access to Document

Fingerprint

Cite this