Runtime Support and Compilation Methods for User-Specified Irregular Data Distributions

Ravi Ponnusamy; Joel Saltz; Yuan Shin Hwang; Alok Choudhary

doi:10.1109/71.406958

Runtime Support and Compilation Methods for User-Specified Irregular Data Distributions

Ravi Ponnusamy, Joel Saltz, Yuan Shin Hwang, Alok Choudhary

Electrical and Computer Engineering

Research output: Contribution to journal › Article › peer-review

37 Scopus citations

Abstract

This paper describes two new ideas by which a High Performance Fortran compiler can deal with irregular computations effectively. The first mechanism invokes a user specified mapping procedure via a set of proposed compiler directives. The directives allow use of program arrays to describe graph connectivity, spatial location of array elements, and computational load. The second mechanism is a conservative method for compiling irregular loops in which dependence arises only due to reduction operations. This mechanism in many cases enables a compiler to recognize that it is possible to reuse previously computed information from inspectors (e.g., communication schedules, loop iteration partitions, and information that associates off-processor data copies with on-processor buffer locations). This paper also presents performance results for these mechanisms from a Fortran 90D compiler implementation.

Original language	English (US)
Pages (from-to)	815-831
Number of pages	17
Journal	IEEE Transactions on Parallel and Distributed Systems
Volume	6
Issue number	8
DOIs	https://doi.org/10.1109/71.406958
State	Published - Aug 1995

Keywords

Runtime support
data distributions
distributed memory machines
high performance Fortran
irregular problems
language directives
parallelizing compilers

ASJC Scopus subject areas

Signal Processing
Hardware and Architecture
Computational Theory and Mathematics

Access to Document

10.1109/71.406958

Cite this

@article{38c9d9cc61524db9888d58777fea5905,

title = "Runtime Support and Compilation Methods for User-Specified Irregular Data Distributions",

abstract = "This paper describes two new ideas by which a High Performance Fortran compiler can deal with irregular computations effectively. The first mechanism invokes a user specified mapping procedure via a set of proposed compiler directives. The directives allow use of program arrays to describe graph connectivity, spatial location of array elements, and computational load. The second mechanism is a conservative method for compiling irregular loops in which dependence arises only due to reduction operations. This mechanism in many cases enables a compiler to recognize that it is possible to reuse previously computed information from inspectors (e.g., communication schedules, loop iteration partitions, and information that associates off-processor data copies with on-processor buffer locations). This paper also presents performance results for these mechanisms from a Fortran 90D compiler implementation.",

keywords = "Runtime support, data distributions, distributed memory machines, high performance Fortran, irregular problems, language directives, parallelizing compilers",

author = "Ravi Ponnusamy and Joel Saltz and Hwang, {Yuan Shin} and Alok Choudhary",

note = "Funding Information: This work was sponsored in part by ARPA (NAG-1-1485), NSF (ASC 9213821) and ONR (N00014-93-1-0158, N00014-94-1-0907). Also supported by NASA Contract No. NAS1-19480 while author Saltz was in residence at ICASE, NASA Langley Research Center, Hampton, Virginia. Author Choud-hary was also supported by NSF Young Investigator award (CCR-9357840) and matching funds Intel SSD division and IBM Corporation. The content of the information does not necessarily reflect the position or the policy of the Government and no official endorsement should be inferred.",

year = "1995",

month = aug,

doi = "10.1109/71.406958",

language = "English (US)",

volume = "6",

pages = "815--831",

journal = "IEEE Transactions on Parallel and Distributed Systems",

issn = "1045-9219",

publisher = "IEEE Computer Society",

number = "8",

}

TY - JOUR

T1 - Runtime Support and Compilation Methods for User-Specified Irregular Data Distributions

AU - Ponnusamy, Ravi

AU - Saltz, Joel

AU - Hwang, Yuan Shin

AU - Choudhary, Alok

N1 - Funding Information: This work was sponsored in part by ARPA (NAG-1-1485), NSF (ASC 9213821) and ONR (N00014-93-1-0158, N00014-94-1-0907). Also supported by NASA Contract No. NAS1-19480 while author Saltz was in residence at ICASE, NASA Langley Research Center, Hampton, Virginia. Author Choud-hary was also supported by NSF Young Investigator award (CCR-9357840) and matching funds Intel SSD division and IBM Corporation. The content of the information does not necessarily reflect the position or the policy of the Government and no official endorsement should be inferred.

PY - 1995/8

Y1 - 1995/8

N2 - This paper describes two new ideas by which a High Performance Fortran compiler can deal with irregular computations effectively. The first mechanism invokes a user specified mapping procedure via a set of proposed compiler directives. The directives allow use of program arrays to describe graph connectivity, spatial location of array elements, and computational load. The second mechanism is a conservative method for compiling irregular loops in which dependence arises only due to reduction operations. This mechanism in many cases enables a compiler to recognize that it is possible to reuse previously computed information from inspectors (e.g., communication schedules, loop iteration partitions, and information that associates off-processor data copies with on-processor buffer locations). This paper also presents performance results for these mechanisms from a Fortran 90D compiler implementation.

AB - This paper describes two new ideas by which a High Performance Fortran compiler can deal with irregular computations effectively. The first mechanism invokes a user specified mapping procedure via a set of proposed compiler directives. The directives allow use of program arrays to describe graph connectivity, spatial location of array elements, and computational load. The second mechanism is a conservative method for compiling irregular loops in which dependence arises only due to reduction operations. This mechanism in many cases enables a compiler to recognize that it is possible to reuse previously computed information from inspectors (e.g., communication schedules, loop iteration partitions, and information that associates off-processor data copies with on-processor buffer locations). This paper also presents performance results for these mechanisms from a Fortran 90D compiler implementation.

KW - Runtime support

KW - data distributions

KW - distributed memory machines

KW - high performance Fortran

KW - irregular problems

KW - language directives

KW - parallelizing compilers

UR - http://www.scopus.com/inward/record.url?scp=0029356841&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0029356841&partnerID=8YFLogxK

U2 - 10.1109/71.406958

DO - 10.1109/71.406958

M3 - Article

AN - SCOPUS:0029356841

SN - 1045-9219

VL - 6

SP - 815

EP - 831

JO - IEEE Transactions on Parallel and Distributed Systems

JF - IEEE Transactions on Parallel and Distributed Systems

IS - 8

ER -

Runtime Support and Compilation Methods for User-Specified Irregular Data Distributions

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this