You are in the main area:Organization
Headerimage for: EuroMPI/ASIA 2014 Workshop: Challenges in Data-Centric Computing (BigDataComputing'2014)

EuroMPI/ASIA 2014 Workshop: Challenges in Data-Centric Computing (BigDataComputing'2014)

This workshop is a 1-day event organized as a part of the EuroMPI/ASIA conference. It aims at bringing together researchers from the domains related to data-centric computing to discuss challenges and possible  solutions for data-centric application development in high-performance computing environments. The workshop is organized by the High Performance Center Stuttgart (HLRS) in cooperation with the Steinbuch Centre for Computing, Karlsruhe Institute of Technology. It is further supported by the EC FP7 projects JUNIPER and DreamCloud

 

 

News and Announcements

 

Please note the following important information:

  • [9 July] List of accepted papers has been published
  • [7 July] Final paper submission deadline has been extended until: 18 July
  • [7 July] Workshop's date has been fixed for: 10 September

 

Organizational Information

 

The workshop is held in conjunction with the EuroMPI/ASIA conference, which happens in Kyoto, Japan on 9-12 September 2014. Please see the hosting conference's site for details.

Timeline:

  • Full paper submission due: 1 June
  • Author notification: 15 June
  • Final paper version submission:
    1 June
    18 July
  • Workshop's date: 10 September

 

 

Accepted Papers

 

Based on the selection done by the PC members - experts on the fields addressed by the workshop, the following papers have been accepted for the presentation at the EuroMPI'14 conference and for the publication in the post-conference (ACM and SIGHPC) proceedings:

  • Architecture-Awareness for Real-Time Big Data Systems
    Ian Gray, Neil Audsley, Yu Chan and Andy Wellings
  • Locality-Aware Process Mapping for High Performance Collective MPI-IO on FEFS with Tofu Interconnect
    Yuichi Tsujita, Atsushi Hori and Yutaka Ishikawa
  • Evaluation of Asynchronous MPI Communication in Map-Reduce System on the K Computer
    Motohiko Matsuda, Shinichiro Takizawa and Naoya Maruyama
  • Decoupling Architecture for All-to-all Computation
    Atsushi Hori, Kazumi Yoshinaga, Atsushi Tokuhisa, Yasumasa Joti, Kensuke Okada, Takashi Sugimoto, Mitsuhiro Yamaga, Takaki Hatsui, Makina Yabashi, Yuji Sugita, Yutaka Ishikawa and Nobuhiro Go
  • HPC in Big Data Age: An Evaluation Report for Java-Based
    Data-Intensive Applications Implemented with Hadoop and OpenMPI
    Alexey Cheptsov

 

 

Call for Papers

 

Many academic and industrial domains have been concerned by numerous research and innovation activities which target the  “big data” problem. Whereas the term “big data” itself is quite controversial and one may argue that this sort of problems has been known for several decades already, it has allowed the supercomputing community to revisit the major application areas of the modern HPC’s. Data-centric applications have started to gain the deserved level of attention in the HPC community as well.

Among the major problems, the lack of supports from the side of the existing parallelisation technologies, in particular for MPI, has prevented the usage of large-scale computing infrastructures in the typical data-centric application domains, such as Semantic Web, Information Retrieval, Knowledge Engineering, etc. However, quite a few activities have been done around the support of data-centric solutions in MPI. Prominent examples of the newly-emerged tools are Java bindings for MPI, Open MPI’s Hadoop integration framework MR+, etc. 

The goal of the workshop is to foster the discussion around the challenges in data-centric computing and envisioned solutions for them. Following this goal, the workshop will serve as a forum to present research activities related high-performance computing community in emerging data-centric application scenarios, such as coming from Semantic Web and other domains. The workshop addresses the current and future challenges related to the development of data-centric applications (in particular, by means of MPI as well as related technologies, such as MapReduce/Hadoop).

The main topics of interest are:

  • Applications

Presentation of "big-data" application use-cases  which are challenging in terms of computation as well as communication demands and storage requirements.

  • Parallelisation technologies

Discussion on data-centric parallelisation approaches (MPI, MapReduce, and others) and their application to the data-centric application domain.

  • Tools

Introduction of tools for supporting the development, debugging, optimization, or deployment and execution of data analytic applications on HPC infrastructures.

  • Supporting activities

Presentation of any other interesting research activities that target the big data application domain but do not specifically fall into any of the previous categories.

 

 

Organizers

 

Organization Committee:

  • Alexey Cheptsov - High Performance Computing Center Stuttgart (HLRS), Germany
  • Jose Gracia - High Performance Computing Center Stuttgart (HLRS), Germany
  • Jie Tao - Steinbuch Centre for Computing (SCC) of the Karlsruhe Institute of Technology (KIT), Germany

 

Program Committee:

  • Michael Resch - High Performance Computing Center Stuttgart (HLRS), Germany
  • Neil Audsley - University of York, UK
  • Ralph Castain - Intel Inc., USA
  • Philip Moore - Birmingham City University, UK
  • Heinz Kredel - University of Mannheim, Germany
  • Ian Gray - University of York, UK
  • Roberto R. Expósito - University of A Corunna, Spain
  • Eric Goodman, Sandia National Laboratories, USA
  • David Mizell - YarcData, USA
  • Ryusuke Egawa - Cybersciecnce Center of Tohoku University, Sendai, Japan
  • Alexander Vazhenin - University of Aizu, Japan
  • Alexey Cheptsov - High Performance Computing Center Stuttgart (HLRS), Germany
  • Jose Gracia - High Performance Computing Center Stuttgart (HLRS), Germany
  • Jie Tao - Steinbuch Centre for Computing (SCC) of the Karlsruhe Institute of Technology (KIT), Germany

 

 

 

 

Requirements to Contributions

 

We invite good quality papers on actual scientific and technological Big Data challenges and solutions. The papers should be formatted according to the ACM rulels and submitted in the pdf format. There are templates available for Word and Latex formats. The length of each paper must not exceed 6 pages.

 

 

Submission Procedure

 

Submissions should be done via the EasyChair system set up at the Link.

 

 

Contact Information

 

Please send your enquiries by email at the following address:
bigdatacomputing2014[at]easychair.org

 

 

Supporting Projects