Design and implementation of the PULSAR programming system for large scale computing

Jakub Kurzak; Piotr Luszczek; Ichitaro Yamazaki; Yves Robert; Jack Dongarra

doi:10.14529/jsfi170101

Design and implementation of the PULSAR programming system for large scale computing

Jakub Kurzak, Piotr Luszczek, Ichitaro Yamazaki, Yves Robert, Jack Dongarra

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

1 Scopus citations

Abstract

The objective of the PULSAR project was to design a programming model suitable for largescale machines with complex memory hierarchies, and to deliver a prototype implementation of a runtime system supporting that model. PULSAR tackled the challenge by proposing a programming model based on systolic processing and virtualization. The PULSAR programming model is quite simple, with point-to-point channels as the main communication abstraction. The runtime implementation is very lightweight and fully distributed, and provides multithreading, messagepassing and multi-GPU offload capabilities. Performance evaluation shows good scalability up to one thousand nodes with one thousand GPU accelerators.

Original language	English
Pages (from-to)	4-26
Number of pages	23
Journal	Supercomputing Frontiers and Innovations
Volume	4
Issue number	1
DOIs	https://doi.org/10.14529/jsfi170101
State	Published - 2017
Externally published	Yes

Funding

This work has been supported by the National Science Foundation, under grant SHF-1117062, Parallel Unified Linear algebra with Systolic ARrays (PULSAR). The authors would also like to thank the National Institute for Computational Sciences, the Georgia Institute of Technology and the Oak Ridge National Laboratory for generous computer allocations on their supercomputers. Yves Robert has been supported by Institut Universitaire de France. This paper is distributed under the terms of the Creative Commons Attribution-Non Commercial 3.0 License which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is properly cited.

Funders	Funder number
National Institute for Computational Sciences
National Science Foundation	SHF-1117062
Oak Ridge National Laboratory
Georgia Institute of Technology
Institut universitaire de France
National Science Foundation

Keywords

Dataflow scheduling
Distributed computing
Hardware accelerators
Massively parallel computing
Multicore processors
Runtime scheduling
Systolic arrays
Virtualization

Access to Document

10.14529/jsfi170101

Cite this

@article{6f3bd894a7f64a0a8c18f664a4061e53,

title = "Design and implementation of the PULSAR programming system for large scale computing",

abstract = "The objective of the PULSAR project was to design a programming model suitable for largescale machines with complex memory hierarchies, and to deliver a prototype implementation of a runtime system supporting that model. PULSAR tackled the challenge by proposing a programming model based on systolic processing and virtualization. The PULSAR programming model is quite simple, with point-to-point channels as the main communication abstraction. The runtime implementation is very lightweight and fully distributed, and provides multithreading, messagepassing and multi-GPU offload capabilities. Performance evaluation shows good scalability up to one thousand nodes with one thousand GPU accelerators.",

keywords = "Dataflow scheduling, Distributed computing, Hardware accelerators, Massively parallel computing, Multicore processors, Runtime scheduling, Systolic arrays, Virtualization",

author = "Jakub Kurzak and Piotr Luszczek and Ichitaro Yamazaki and Yves Robert and Jack Dongarra",

note = "Publisher Copyright: {\textcopyright} The Authors 2017.",

year = "2017",

doi = "10.14529/jsfi170101",

language = "English",

volume = "4",

pages = "4--26",

journal = "Supercomputing Frontiers and Innovations",

issn = "2409-6008",

number = "1",

}

TY - JOUR

T1 - Design and implementation of the PULSAR programming system for large scale computing

AU - Kurzak, Jakub

AU - Luszczek, Piotr

AU - Yamazaki, Ichitaro

AU - Robert, Yves

AU - Dongarra, Jack

N1 - Publisher Copyright: © The Authors 2017.

PY - 2017

Y1 - 2017

N2 - The objective of the PULSAR project was to design a programming model suitable for largescale machines with complex memory hierarchies, and to deliver a prototype implementation of a runtime system supporting that model. PULSAR tackled the challenge by proposing a programming model based on systolic processing and virtualization. The PULSAR programming model is quite simple, with point-to-point channels as the main communication abstraction. The runtime implementation is very lightweight and fully distributed, and provides multithreading, messagepassing and multi-GPU offload capabilities. Performance evaluation shows good scalability up to one thousand nodes with one thousand GPU accelerators.

AB - The objective of the PULSAR project was to design a programming model suitable for largescale machines with complex memory hierarchies, and to deliver a prototype implementation of a runtime system supporting that model. PULSAR tackled the challenge by proposing a programming model based on systolic processing and virtualization. The PULSAR programming model is quite simple, with point-to-point channels as the main communication abstraction. The runtime implementation is very lightweight and fully distributed, and provides multithreading, messagepassing and multi-GPU offload capabilities. Performance evaluation shows good scalability up to one thousand nodes with one thousand GPU accelerators.

KW - Dataflow scheduling

KW - Distributed computing

KW - Hardware accelerators

KW - Massively parallel computing

KW - Multicore processors

KW - Runtime scheduling

KW - Systolic arrays

KW - Virtualization

UR - http://www.scopus.com/inward/record.url?scp=85038430269&partnerID=8YFLogxK

U2 - 10.14529/jsfi170101

DO - 10.14529/jsfi170101

M3 - Article

AN - SCOPUS:85038430269

SN - 2409-6008

VL - 4

SP - 4

EP - 26

JO - Supercomputing Frontiers and Innovations

JF - Supercomputing Frontiers and Innovations

IS - 1

ER -

Design and implementation of the PULSAR programming system for large scale computing

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this