[hpc-announce] CFP: Special Issue on Scientific Programming - In Search of Extreme Scale Scientific Programming Paradigms

Justin Y. Shi shi at temple.edu
Wed Nov 28 09:50:26 CST 2018

The Special Issue of International Journal on Scientific Programming is
accepting papers on the topic of "In Search of Extreme Scale Scientific
Programming Paradigms," submission deadline is March 29, 2019 with
publication date in August 2019.

Call for Papers: In Search of Extreme Scale Scientific Programming Paradigms

It is known that application level program-program (or process-process)
communication and coordination protocols decide the ultimate deliverable
performance and reliability of parallel applications. It is also known that
parallel applications using traditional hop-by-hop protocols, such as MPI
(message passing interface) and OpenMP (program-memory-program), suffer
shrinking MTBF (mean time between failures), as the processing
infrastructure expands for meeting larger computing needs. Regardless of
optimization methods, checkpoint/restart routines generate growing energy
and resource inefficiencies as the applications scale up.

End-to-end protocols, such as blockchain and BitTorrent, can defy the
scaling challenges in distributed and parallel systems. These protocols
follow end-to-end distributed system design principles as opposed to the
hop-by-hop protocols. While the current blockchain and BitTorrent protocols
are too slow for high performance applications, the end-to-end nature of
their design should enable the development of end-to-end resource
optimization techniques. These include potentials for automatic fault
tolerant nonstop parallel computing, optimized energy and resource
efficiency, as well as the potential of best effort performances for
volatile environments. It is plausible that future decentralized
high-performance computing applications can exploit scaling benefits and be
limited only by the economic law of diminishing returns. The solution for
fault tolerant parallel computing can also be used for automated workload
distribution, thus cheaper but slower computing clouds may become
performant and energy efficient HPC work horses.

This special issue aims to publish original and previously unpublished
research results in the broad area of extreme scale scientific programming

Potential topics include but are not limited to the following:

   - Novel scientific programming paradigms that are free from the
   scalability dilemma
   - Legacy application reengineering concepts that can mitigate the
   scalability dilemma when the application up scales
   - Custom program-program communication and coordination protocols that
   can mitigate specific application scaling difficulties
   - Practice and experiences of large-scale scientific simulations with
   demonstrable performance and reliability levels
   - Compilers, libraries, and runtime systems that can mitigate
   application scaling difficulties in performance and reliability
   - Virtualization techniques, such as fault tolerant software containers,
   that can facilitate the scaling of large scale scientific applications
   - Parallel workload optimization methods for heterogeneous environments
   - Parallel workload optimization methods that can also mitigate
   component failures for nonstop computing
   - Parallel application scalability modeling methods for homogeneous and
   heterogeneous environments

Authors can submit their manuscripts through the Manuscript Tracking System
Submission Deadline Friday, 29 March 2019
Publication Date August 2019

Papers are published upon acceptance, regardless of the Special Issue
publication date.
Lead Guest Editor

   - Justin Shi <shi at temple.edu>, Temple University, Pennsylvania, USA

Guest Editors

   - Hsing-Bung Chen <hbchen at lanl.gov>, Los Alamos National laboratory, New
   Mexico, USA
   - Hanqi Guo <hguo at anl.gov>, Argonne National Laboratory, Illinois, USA
   - Qiang Guan <qguan at kent.edu>, Kent State University, Ohio, USA


Weblink: https://www.hindawi.com/journals/sp/si/640542/cfp/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.mcs.anl.gov/mailman/private/hpc-announce/attachments/20181128/e496645a/attachment-0001.html>

More information about the hpc-announce mailing list