SC|05 SC|05 Gateway to Discovery
About Interactive Schedule Programs Registration Exhibits Initiatives & Challenges News & Press Hotel & Travel

The HPC Analytics Challenge is open to everyone except those directly associated with the HPC Analytics Challenge Peer Review Committee. If in doubt, please make an inquiry to the HPC Analytics Challenge Chairs Don Jones (dr.jones@pnl.gov) and Paul Fussell (paul.fussell@boeing.com).

You may use any existing commercial products and applications or research prototypes. And, you may use and combine whatever tools you deem necessary.

 

Submission Information: What to Submit?

1. Notice of intent to participate in the HPC Analytics Challenge: due April 18, 2005

Register and submit your intent to participate to the HPC Analytics Challenge section at the conference submissions website: www.sc-submissions.org

Your submission stating your intent to participate in the HPC Analytics Challenge in SC|05 would include the following:

a. Entry title for the application
b. Contact information
c. Your organization
d. Dataset being used
e. Team leader information, technical point of contact, and team members
f. Upload an abstract (in PDF format) briefly describing your intended Challenge entry focus, dataset, and proposed output.

The uploaded abstract should be less than one page in length.

 

2. Application to participate in the HPC Analytics Challenge: due June 1, 2005. Notice of Intent submissions are not a prerequisite. Applicants who did not submit a Notice of Intent are still welcome to submit their full applications.

A web submission to the Challenge website (www.sc-submissions.org) applying to compete in the HPC Analytics Challenge including
a. Entry title for application
b. Name and contact details for the contact person for this Challenge entry
c. Information on Team leader, technical contact and team members
d. Upload an extended abstract (in PDF format). This extended abstract is to include:

  1. A brief statement of purpose of the Challenge entry
  2. A brief description of the methods to be used in the Challenge entry
  3. A short listing of the major tools or codes to be used in the Challenge entry
  4. A discussion of the dataset(s) used in the Challenge entry, including
    • Overview of the data set
    • Data set size
    • Complexity
    • How this dataset is to be used

e. Overview of methods to display results: Discussion of the methods, tools and/or codes used to display or visualize the results of the Challenge entry. Upload as a PDF-format file to the SC-Submissions web site (www.sc-submissions.org).

These applications will be held in confidence by the HPC Analytics Challenge Committee for the duration of the Challenge up to September 1, 2005.

The Challenge Committee will respond to the applications to participate by May 27, 2005.

 

3. Completed entry to the HPC Analytics Challenge including results: due August 1, 2005

A complete entry comprises

  1. A completed entry form
  2. A summary, and
  3. A video demonstration.

Entry form: The entry form is an updated Challenge application, including
a. Entry title for the application
b. Name and contact details for the contact person for your Challenge entry
c. Information on Team leader, technical contact and team members
d. Upload an extended abstract in PDF-format. The statement is to include

  1. A statement of purpose of the Challenge entry
    A description of the methods to be used in the Challenge entry
    A listing of the major tools or codes to be used in the Challenge entry
    A discussion of the dataset(s) used in the Challenge entry, including
  2. Details of the data set
  3. Data set size
  4. Complexity
  5. How this dataset has been used

e. Overview of methods to show results: Discussion of the methods, tools, or codes used to display or visualize the results of the Challenge entry. Upload as a PDF-format file to the SC-Submissions web site (www.sc-submissions.org).

Summary document: This would be a 2-page publication quality summary (in PDF format) that addresses each of the individual judging criteria listed below (see section on Judging Criteria). The summary should clearly identify the core problem being addressed. It should also provide an overview of the data and the analysis process, as well as summarize the strengths and weaknesses of the approach, and also provide at least one example screen shot. The 2-page summary (in PDF format) should adhere to the following formatting guidelines:

  1. The 2-page limit includes images, tables, code snippets, etc.
  2. Margins: 1-inch margin on all sides.
  3. Font-size: No smaller than 10 pts should be used.
  4. Spacing: Single line spacing for all sections.

Upload as an PDF-format file to the SC-Submissions web site (www.sc-submissions.org).

Winning entries may be included as-is in SC|05 printed materials.

Video Demonstration: The recorded video demonstration should focus on the interactive features of the completed entry and the process of analyzing and displaying the data. Demonstrations should be approximately five minutes in length and will not be considered beyond ten minutes in length. We ask for a recorded demonstration because interactivity is an important component for this Challenge and it would be difficult to evaluate only through a paper submission.

Provide a URL pointing to a web-readable movie, or a mailed in video or CD or DVD. The URL must be active through SC|05, Nov 18, 2005, and preferably for several weeks after that.

How to Submit:

The entry form and summary (in PDF format) are to be uploaded to the SC-Submissions web site (www.sc-submissions.org), and the video demonstration must be provided by either a URL pointing to a web-readable movie, or a mailed in video or CD or DVD. The overnight courier address for submitting the video, CD, or DVD is as follows:

ATN: Paul Fussell / HPC Analytics Challenge

c/o Doss Cook, MC 7L-21
Boeing Math and Computing
2810 160th Ave SE, Bldg 33-07
Bellevue, WA 98008-5462

425.865.3518

If sent via overnight mail, the video tape, CD, or DVD must be received no later than Wednesday, August 3, 2005.

Judging Criteria:

Submissions will be subject to a peer review by judges with analytics expertise. All submissions will be considered with the following three major dimensions in mind:

I. Analytics Techniques / Robustness / Flexibility

  • Flexibility of the application (how generic is your approach)
  • Quality of the written case study (description of the strengths AND weaknesses of the analytics used)
  • Novelty (whether the analysis and visualization methods are common approaches, well known versus more innovative and sophisticated approaches)

II. Performance & Scalability

  • How large is the data set? How well does (and would) the application perform and scale? Would the application struggle if the data set grew to a larger size?
  • Is the method linear, or superlinear/polynomial/exponential with time, and inversely with processors, network speed
  • What architecture components are used to solve the problem
  • Data complexity

III. Knowledge & Discovery (Results)

  • What insights were achieved
  • What are the results? What is the impact of these results?
  • How were the results verified?
  • What is the significance of the results? How meaningful?
  • Quality of the data analysis (what interesting insights were uncovered?)
  • Quality/sophistication/appropriateness of the visualization techniques to illustrate results
  • Usefulness of the application? Creative outcomes?