Semantics-preserving data layout transformations for improved vectorisation

Artjoms Sinkarovs; Sven Bodo Scholz

doi:10.1145/2502323.2502332

Semantics-preserving data layout transformations for improved vectorisation

Artjoms Sinkarovs, Sven Bodo Scholz

School of Mathematical & Computer Sciences

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

4 Citations (Scopus)

Abstract

Data-Layouts that are favourable from an algorithmic perspective often are less suitable for vectorisation, i.e., for an effective use of modern processor's vector instructions. This paper presents work on a compiler driven approach towards automatically transforming data layouts into a form that is suitable for vectorisation. In particular, we present a program transformation for a first-order functional array programming language that systematically modifies they layouts of all data structures. At the same time, the transformation also adjusts the code that operates on these structures so that the overall computation remains unchanged. We define a correctness criterion for layout modifying program transformations and we show that our transformation abides to this criterion.

Original language	English
Title of host publication	Proceedings of the ACM SIGPLAN International Conference on Functional Programming, ICFP
Pages	59-69
Number of pages	11
DOIs	https://doi.org/10.1145/2502323.2502332
Publication status	Published - 21 Oct 2013
Event	2nd ACM SIGPLAN Workshop on Functional High-Performance Computing, Co-located with the 18th ACM SIGPLAN International Conference on Functional Programming - Boston, MA, United Kingdom Duration: 23 Sept 2013 → 23 Sept 2013

Conference

Conference	2nd ACM SIGPLAN Workshop on Functional High-Performance Computing, Co-located with the 18th ACM SIGPLAN International Conference on Functional Programming
Abbreviated title	FHPC 2013 & ICFP 2013
Country/Territory	United Kingdom
City	Boston, MA
Period	23/09/13 → 23/09/13

Keywords

Correctness
Program transformation
Type systems
Vectorisation

ASJC Scopus subject areas

Software

Access to Document

10.1145/2502323.2502332

Cite this

@inproceedings{1564f4c02cdb4c02a96d7ec3ff8446fa,

title = "Semantics-preserving data layout transformations for improved vectorisation",

abstract = "Data-Layouts that are favourable from an algorithmic perspective often are less suitable for vectorisation, i.e., for an effective use of modern processor's vector instructions. This paper presents work on a compiler driven approach towards automatically transforming data layouts into a form that is suitable for vectorisation. In particular, we present a program transformation for a first-order functional array programming language that systematically modifies they layouts of all data structures. At the same time, the transformation also adjusts the code that operates on these structures so that the overall computation remains unchanged. We define a correctness criterion for layout modifying program transformations and we show that our transformation abides to this criterion.",

keywords = "Correctness, Program transformation, Type systems, Vectorisation",

author = "Artjoms Sinkarovs and Scholz, {Sven Bodo}",

year = "2013",

month = oct,

day = "21",

doi = "10.1145/2502323.2502332",

language = "English",

isbn = "9781450323819",

pages = "59--69",

booktitle = "Proceedings of the ACM SIGPLAN International Conference on Functional Programming, ICFP",

note = "2nd ACM SIGPLAN Workshop on Functional High-Performance Computing, Co-located with the 18th ACM SIGPLAN International Conference on Functional Programming, FHPC 2013 & ICFP 2013 ; Conference date: 23-09-2013 Through 23-09-2013",

}

Sinkarovs, A & Scholz, SB 2013, Semantics-preserving data layout transformations for improved vectorisation. in Proceedings of the ACM SIGPLAN International Conference on Functional Programming, ICFP. pp. 59-69, 2nd ACM SIGPLAN Workshop on Functional High-Performance Computing, Co-located with the 18th ACM SIGPLAN International Conference on Functional Programming, Boston, MA, United Kingdom, 23/09/13. https://doi.org/10.1145/2502323.2502332

TY - GEN

T1 - Semantics-preserving data layout transformations for improved vectorisation

AU - Sinkarovs, Artjoms

AU - Scholz, Sven Bodo

PY - 2013/10/21

Y1 - 2013/10/21

N2 - Data-Layouts that are favourable from an algorithmic perspective often are less suitable for vectorisation, i.e., for an effective use of modern processor's vector instructions. This paper presents work on a compiler driven approach towards automatically transforming data layouts into a form that is suitable for vectorisation. In particular, we present a program transformation for a first-order functional array programming language that systematically modifies they layouts of all data structures. At the same time, the transformation also adjusts the code that operates on these structures so that the overall computation remains unchanged. We define a correctness criterion for layout modifying program transformations and we show that our transformation abides to this criterion.

AB - Data-Layouts that are favourable from an algorithmic perspective often are less suitable for vectorisation, i.e., for an effective use of modern processor's vector instructions. This paper presents work on a compiler driven approach towards automatically transforming data layouts into a form that is suitable for vectorisation. In particular, we present a program transformation for a first-order functional array programming language that systematically modifies they layouts of all data structures. At the same time, the transformation also adjusts the code that operates on these structures so that the overall computation remains unchanged. We define a correctness criterion for layout modifying program transformations and we show that our transformation abides to this criterion.

KW - Correctness

KW - Program transformation

KW - Type systems

KW - Vectorisation

U2 - 10.1145/2502323.2502332

DO - 10.1145/2502323.2502332

M3 - Conference contribution

AN - SCOPUS:84885640033

SN - 9781450323819

SP - 59

EP - 69

BT - Proceedings of the ACM SIGPLAN International Conference on Functional Programming, ICFP

T2 - 2nd ACM SIGPLAN Workshop on Functional High-Performance Computing, Co-located with the 18th ACM SIGPLAN International Conference on Functional Programming

Y2 - 23 September 2013 through 23 September 2013

ER -

Semantics-preserving data layout transformations for improved vectorisation

Abstract

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Fingerprint

Cite this