Discamus continentiam augere, luxuriam coercere
Home -> Publications
Home
  Publications
    
all years
    2017
    2016
    2015
    2014
    2013
    2012
    2011
    2010
    2009
    2008
    2007
    2006
    2005
    2004
    theses
    techreports
    presentations
    edited volumes
    conferences
  Awards
  Research
  Teaching
  BLOG
  Miscellaneous
  Full CV [pdf]






  Events








  Past Events





Publications of Torsten Hoefler
Copyright Notice:

The documents distributed by this server have been provided by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a noncommercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, notwithstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.

T. Gysi, J. Baer, T. Hoefler:

 dCUDA: Hardware Supported Overlap of Computation and Communication

(In Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (SC16), presented in Salt Lake City, Utah, pages 52:1--52:12, IEEE Press, ISBN: 978-1-4673-8815-3, Nov. 2016)

Abstract

Over the last decade, CUDA and the underlying GPU hardware architecture have continuously gained popularity in various high-performance computing application domains such as climate modeling, computational chemistry, or machine learning. Despite this popularity, we lack a single coherent programming model for GPU clusters. We therefore introduce the dCUDA programming model, which implements device-side remote memory access with target notification. To hide instruction pipeline latencies, CUDA programs over-decompose the problem and over-subscribe the device by running many more threads than there are hardware execution units. Whenever a thread stalls, the hardware scheduler immediately proceeds with the execution of another thread ready for execution. This latency hiding technique is key to make best use of the available hardware resources. With dCUDA, we apply latency hiding at cluster scale to automatically overlap computation and communication. Our benchmarks demonstrate perfect overlap for memory bandwidth-bound tasks and good overlap for compute-bound tasks.

Documents

download article:
download slides:


Recorded talk (best effort)

 

BibTeX

@inproceedings{,
  author={T. Gysi and J. Baer and T. Hoefler},
  title={{dCUDA: Hardware Supported Overlap of Computation and Communication}},
  year={2016},
  month={Nov.},
  pages={52:1--52:12},
  booktitle={Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis (SC16)},
  location={Salt Lake City, Utah},
  publisher={IEEE Press},
  isbn={978-1-4673-8815-3},
  source={http://www.unixer.de/~htor/publications/},
}

serving: 54.224.102.26:55982© Torsten Hoefler