Detail publikace

GPU-accelerated Evolutionary Design of the Complete Exchange Communication on Wormhole Networks

JAROŠ, J. TYRALA, R.

Originální název

GPU-accelerated Evolutionary Design of the Complete Exchange Communication on Wormhole Networks

Typ

článek ve sborníku ve WoS nebo Scopus

Jazyk

angličtina

Originální abstrakt

The communication overhead is one of the main challenges in the exascale era, where millions of compute cores are expected to collaborate on solving complex jobs. However, many algorithms will not scale since they require complex global communication and synchronisation. In order to perform the communication as fast as possible, contentions, blocking and deadlock must be avoided. Recently, we have developed an evolutionary tool producing fast and safe communication schedules reaching the lower bound of the theoretical time complexity. Unfortunately, the execution time associated with the evolution process raises up to tens of hours, even when being run on a multi-core processor. In this paper, we propose a revised implementation accelerated by a single Graphic Processing Unit (GPU) delivering speed-up of 5 compared to a quad-core CPU. Subsequently, we introduce an extended version employing up to 8 GPUs in a shared memory environment offering a speed-up of almost 30. This significantly extends the range of interconnection topologies we can cover.

Klíčová slova

Complete exchange communication, Collective communications, communication scheduling, evolutionary design, GPU-based acceleration, multi-GPU systems.

Autoři

JAROŠ, J.; TYRALA, R.

Rok RIV

2014

Vydáno

12. 7. 2014

Nakladatel

Association for Computing Machinery

Místo

New York, NY

ISBN

978-1-4503-2662-9

Kniha

GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference

Strany od

1023

Strany do

1030

Strany počet

8

URL

http://dl.acm.org/citation.cfm?id=2576768.2598315

BibTex

@inproceedings{BUT111530,
  author="Jiří {Jaroš} and Radek {Tyrala}",
  title="GPU-accelerated Evolutionary Design of the Complete Exchange Communication on Wormhole Networks",
  booktitle="GECCO 2014 - Proceedings of the 2014 Genetic and Evolutionary Computation Conference",
  year="2014",
  pages="1023--1030",
  publisher="Association for Computing Machinery",
  address="New York, NY",
  doi="10.1145/2576768.2598315",
  isbn="978-1-4503-2662-9",
  url="http://dl.acm.org/citation.cfm?id=2576768.2598315"
}