Polly - Polyhedral optimization in LLVM

Tobias Grosser; Hongbin Zheng; Raghesh Aloor; Andreas Simbürger; Armin Größlinger; Louis-Noël Pouchet

Polly - Polyhedral optimization in LLVM

Tobias Grosser, Hongbin Zheng, Raghesh Aloor, Andreas Simbürger, Armin Größlinger, Louis-Noël Pouchet

Research output: Contribution to conference › Paper › peer-review

Abstract

Various powerful polyhedral techniques exist to optimize computation intensive programs effectively. Applying these techniques on any non-trivial program is still surprisingly difficult and often not as effective as expected. Most polyhedral tools are limited to a specific programming language. Even for this language, relevant code needs to match specific syntax that rarely appears in existing code. It is therefore hard or even impossible to process existing programs automatically. In addition, most tools target C or OpenCL code, which prevents effective communication with compiler internal optimizers. As a result target architecture specific optimizations are either little effective or not approached at all.

Original language	English
Number of pages	6
Publication status	Published - 3 Apr 2011
Event	First International Workshop on Polyhedral Compilation Techniques - Chamonix, France Duration: 3 Apr 2011 → 3 Apr 2011 http://impact2011.inrialpes.fr/en/index.html

Workshop

Workshop	First International Workshop on Polyhedral Compilation Techniques
Abbreviated title	IPACT 2011
Country/Territory	France
City	Chamonix
Period	3/04/11 → 3/04/11
Internet address	http://impact2011.inrialpes.fr/en/index.html

Access to Document

http://perso.ens-lyon.fr/christophe.alias/impact2011/impact-07.pdfLicence: All Rights Reserved

Cite this

@conference{fdbb2602c7884f9080bede48bcddb67f,

title = "Polly - Polyhedral optimization in LLVM",

abstract = "Various powerful polyhedral techniques exist to optimize computation intensive programs effectively. Applying these techniques on any non-trivial program is still surprisingly difficult and often not as effective as expected. Most polyhedral tools are limited to a specific programming language. Even for this language, relevant code needs to match specific syntax that rarely appears in existing code. It is therefore hard or even impossible to process existing programs automatically. In addition, most tools target C or OpenCL code, which prevents effective communication with compiler internal optimizers. As a result target architecture specific optimizations are either little effective or not approached at all.",

author = "Tobias Grosser and Hongbin Zheng and Raghesh Aloor and Andreas Simb{\"u}rger and Armin Gr{\"o}{\ss}linger and Louis-No{\"e}l Pouchet",

year = "2011",

month = apr,

day = "3",

language = "English",

note = "First International Workshop on Polyhedral Compilation Techniques, IPACT 2011 ; Conference date: 03-04-2011 Through 03-04-2011",

url = "http://impact2011.inrialpes.fr/en/index.html",

}

TY - CONF

T1 - Polly - Polyhedral optimization in LLVM

AU - Grosser, Tobias

AU - Zheng, Hongbin

AU - Aloor, Raghesh

AU - Simbürger, Andreas

AU - Größlinger, Armin

AU - Pouchet, Louis-Noël

PY - 2011/4/3

Y1 - 2011/4/3

N2 - Various powerful polyhedral techniques exist to optimize computation intensive programs effectively. Applying these techniques on any non-trivial program is still surprisingly difficult and often not as effective as expected. Most polyhedral tools are limited to a specific programming language. Even for this language, relevant code needs to match specific syntax that rarely appears in existing code. It is therefore hard or even impossible to process existing programs automatically. In addition, most tools target C or OpenCL code, which prevents effective communication with compiler internal optimizers. As a result target architecture specific optimizations are either little effective or not approached at all.

AB - Various powerful polyhedral techniques exist to optimize computation intensive programs effectively. Applying these techniques on any non-trivial program is still surprisingly difficult and often not as effective as expected. Most polyhedral tools are limited to a specific programming language. Even for this language, relevant code needs to match specific syntax that rarely appears in existing code. It is therefore hard or even impossible to process existing programs automatically. In addition, most tools target C or OpenCL code, which prevents effective communication with compiler internal optimizers. As a result target architecture specific optimizations are either little effective or not approached at all.

M3 - Paper

T2 - First International Workshop on Polyhedral Compilation Techniques

Y2 - 3 April 2011 through 3 April 2011

ER -

Polly - Polyhedral optimization in LLVM

Abstract

Workshop

Access to Document

Fingerprint

Cite this