Partitioned Global Address Space Languages

被引:30
|
作者
De Wael, Mattias [1 ]
Marr, Stefan [2 ]
De Fraine, Bruno [3 ]
Van Cutsem, Tom [1 ]
De Meuter, Wolfgang [1 ]
机构
[1] Vrije Univ Brussel, SOFT DINF, B-1050 Brussels, Belgium
[2] Johannes Kepler Univ Linz, A-4040 Linz, Austria
[3] Synopsys, B-3001 Heverlee, Belgium
关键词
Design; Languages; Parallel programming; HPC; PGAS; message passing; one-sided communication; data distribution; data access; survey;
D O I
10.1145/2716320
中图分类号
TP301 [理论、方法];
学科分类号
081202 ;
摘要
The Partitioned Global Address Space (PGAS) model is a parallel programming model that aims to improve programmer productivity while at the same time aiming for high performance. The main premise of PGAS is that a globally shared address space improves productivity, but that a distinction between local and remote data accesses is required to allow performance optimizations and to support scalability on large-scale parallel architectures. To this end, PGAS preserves the global address space while embracing awareness of nonuniform communication costs. Today, about a dozen languages exist that adhere to the PGAS model. This survey proposes a definition and a taxonomy along four axes: how parallelism is introduced, how the address space is partitioned, how data is distributed among the partitions, and finally, how data is accessed across partitions. Our taxonomy reveals that today's PGAS languages focus on distributing regular data and distinguish only between local and remote data access cost, whereas the distribution of irregular data and the adoption of richer data access cost models remain open challenges.
引用
收藏
页数:27
相关论文
共 50 条
  • [1] Productivity and performance using partitioned global address space languages
    University of California, Berkeley
    不详
    [J]. PASCO: Proc Int workshop Parallel Symb Comput, 2007, (24-32):
  • [2] Efficient Runtime Support for a Partitioned Global Logical Address Space
    Larkins, D. Brian
    Snyder, John
    Dinan, James
    [J]. PROCEEDINGS OF THE 47TH INTERNATIONAL CONFERENCE ON PARALLEL PROCESSING, 2018,
  • [3] Design and evaluation of efficient global data movement in partitioned global address space
    Murai, Hitoshi
    Sato, Mitsuhisa
    [J]. PARALLEL COMPUTING, 2020, 96
  • [4] Tuning collective communication for Partitioned Global Address Space programming models
    Nishtala, Rajesh
    Zheng, Yili
    Hargrove, Paul H.
    Yelick, Katherine A.
    [J]. PARALLEL COMPUTING, 2011, 37 (09) : 576 - 591
  • [5] A Runtime System for Finite Element Methods in a Partitioned Global Address Space
    Groth, Stefan
    Grunewald, Daniel
    Teich, Jurgen
    Hannig, Frank
    [J]. 17TH ACM INTERNATIONAL CONFERENCE ON COMPUTING FRONTIERS 2020 (CF 2020), 2020, : 39 - 48
  • [6] Scaling Data Race Detection for Partitioned Global Address Space Programs
    Park, Chang-Seo
    Sen, Koushik
    Iancu, Costin
    [J]. ACM SIGPLAN NOTICES, 2013, 48 (08) : 305 - 306
  • [7] Algorithms of Collective Operations for Distributed Arrays in Partitioned Global Address Space
    Paznikov, Alexey A.
    Kurnosov, Mikhail G.
    Kupriyanov, Mikhail S.
    [J]. 2017 IEEE II INTERNATIONAL CONFERENCE ON CONTROL IN TECHNICAL SYSTEMS (CTS), 2017, : 5 - 8
  • [8] Improving Data Locality for Irregular Partitioned Global Address Space Parallel Programs
    Larkins, D. Brian
    [J]. PROCEEDINGS OF THE 50TH ANNUAL ASSOCIATION FOR COMPUTING MACHINERY SOUTHEAST CONFERENCE, 2012,
  • [9] XpressSpace: a programming framework for coupling partitioned global address space simulation codes
    Zhang, Fan
    Docan, Ciprian
    Bui, Hoang
    Parashar, Manish
    Klasky, Scott
    [J]. CONCURRENCY AND COMPUTATION-PRACTICE & EXPERIENCE, 2014, 26 (03): : 644 - 661
  • [10] Evaluation of the partitioned global address space (PGAS) model for an inviscid Euler solver
    Prugger, Martina
    Einkemmer, Luk
    Ostermann, Alexander
    [J]. PARALLEL COMPUTING, 2016, 60 : 22 - 40