Note: Due to insufficient submissions, this event has been cancelled.
High Performance Data Analysis and Visualization (HPDAV) 2015
An IPDPS 2015 Workshop
While the purpose of visualization and analysis is insight, realizing that objective requires solving complex problems related to crafting or adapting algorithms and applications to take advantage of evolving architectures, and to solve increasingly complex data understanding problems for ever larger and more complex data. These architectures, and the systems from which they are built, have increasingly deep memory hierarchies, increasing concurrency, decreasing relative per-core/per-node I/O capacity, lessening memory per core, are increasingly prone to failures, and face power limitations.
The purpose of this workshop is to bring together researchers, engineers, and architects of data-intensive computing technologies, which span visualization, analysis, and data management, to present and discuss research topics germane to high performance data analysis and visualization. Specifically, this workshop focuses on research topics related to adapting/creating algorithms, technologies, and applications for use on emerging computational architectures and platforms.
The workshop format includes traditional research papers (8pages) for in-depth topics, short papers (4 pages) for works in progress, and a panel discussion.
Paper TopicsWe invite papers on original, unpublished research in the following topic areas under the general umbrella of high performance visualization and analysis:
- Increasing concurrency at the node level, and at the systemwide level.
- Optimizations for improving performance, e.g., decreasing runtime, leveraging a deepening memory hierarchy, reducing data movement, reducing power consumption.
- Applications of visualization and analysis, where there is a strong thematic element related to being able to solve a larger or more complex problem because of algorithmic or design advances that take advantage of increasing concurrency, architectural features, etc.
- Data analysis and/or visualization systems/designs/architectures having an emphasis upon scalability, resilience, high-throughput/high-capacity, and that are able to take advantage of emerging architectures.
We anticipate a portion of the program to be dedicated to 20-minute research talks, and a portion to be dedicated to 10-minute short talks.
- Long papers: 8 pages, with additional pages for references if needed, to provide a full problem description, background and related work, methodology, and results.
- Short papers: 4 pages, for works in progress, vignettes, and topics of more limited scope.
- Latex and other templates: may be found via http://www.ipdps.org.
We solicit proposals for a panel, that would present position statements on topics related to HPDAV and would be of interest to a broad audience.
Guidelines for panel submissions:
- Content: Panel proposal statements should include the title of the panel, the names of the panelists, an overall panel statement about the focus and thesis of the panel, along with a brief position statement from each of the prospective panelists.
- Length: The panel proposal should be of sufficient length to convey the main objective for the panel, along with a clear statement about each panelist's position. The following guidelines are not strict, but may help give an idea of the level of detail: panel overview – 500 words; each panelist's statement – 200-400 words each.
- Format: please submit a single PDF containing all of the panel proposal content.
This workshop anticipates having one panel discussion, which would consist of 40 minutes of panelist presentations and 20 minutes of audience discussion.
Peer review process
All submissions – long papers, short papers, panel proposal – will undergo a peer-review process consisting of at least three reviewers.
All papers, both short and full-length, will be evaluated on the following criteria:
- Appropriateness for the scope of the workshop
- Technical soundness, novelty, and creativity
- Workshop submissions: 5 Jan 2015. All submissions – long papers, short papers, panel proposal – are due Monday, 5 Jan 2015, 23:59 Anywhere On Earth. Please submit your paper/panel proposal via the EDAS website used by IPDPS to one of the following three tracks: full papers, short papers, panel.
- Author notification: 9 Feb 2015. Authors of all submissions – long papers, short papers, and panel proposals – will be notified of the review via email results by 9 Feb 2015.
- Camera-ready copy: 27 Feb 2015. Authors of are expected to do revisions and produce camera-ready copy, which is due by 27 Feb 2015.
Presentation at the workshop
It is expected that each accepted submission will be presented at the workshop.
Wes Bethel, Lawrence Berkeley National Laboratory (organizer)
Randall Frank, Applied Research Associates
Kelly Gaither, Texas Advanced Computing Center
Berk Geveci, Kitware
Alex Gray, Skytree
Ken Joy, UC Davis
Pat McCormick, Los Alamos National Laboratory
Vijay Natarajan, Indian Institute of Science
Peter Nugent, Lawrence Berkeley National Laboratory
George Ostrouchov, Oak Ridge National Laboratory
Rob Ross, Argonne National Laboratory
John Shalf, Lawrence Berkeley National Laboratory
Dale Southard, NVIDIA
Craig Tull, Lawrence Berkeley National Laboratory
Venkat Vishwanath, Argonne National Laboratory John Wu, Lawrence Berkeley National Laboratory
Please contact the workshop organizer via email at ewbethel at lbl dot gov.