ECE 563 Projects, Spring 2013

Due dates (Fridays, 11:59 PM):
  • Project proposal draft, Week 2 (Jan 18)
  • Final Project proposal, Week 3 (Jan 25)
  • Intermediate report, Week 10 (Mar 22)
  • Final Project report, Last week of class (Apr 26)

    Submission format

    Reports must be in PDF format. Combine all files into a tar file and email to eigenman@purdue.edu

    File name: your_names-ece563-VERSION.tar (please use the name spelling of your names in all submissions)
    VERSION = draftproposal | proposal | intermediate | finalreport

    Include the original (usually the sequential) and the current, improved parallel program version in the submission. Include all files needed for making and executing the program. This is expecially important for the final project report (a student grader will make and run the programs to verify the result, based on the included make/run information.)

    Alternatively, you can submit a URL, at which you place all these files. All files "submitted" must remain unchanged at this site after the submission date. When your report is ready, send email to eigenman@purdue.edu stating that fact.

    Project descriptions should include

    In your final report, mention all of the following information. You may change the order, except for the first section (summary). If you don't have information to report, say "nothing to report" or "unknown".
  • Summary, including the goals you set for the project, a summary of the steps you took in the project and milestones you reached after each step.
  • Description of the application and algorithm(s)
  • Program description (language, #lines, #subroutines)
  • Names of developers and users of the code, if known.
  • Bibliographic references and web links to documents and papers that describe the application and the program, where available.
  • "make" information: machine used; compiler and compiler flags; libraries used, if any; run information (commands for setting up environment and providing input data).
  • Data sets: describe input data in terms of their meaning from an application angle and their properties (size, format)
  • Execution time(s) of the original program version. Indicate the machines and data sets used.
  • Memory footprint(s).
  • Parallel programming models used.
  • "Performance diaries:" describe all steps of applied transformations and settings of environment parameters as well as the execution time before and after each step. This should correspond to the project roadmap described at the beginning of the report.
  • Discussion of the results. Include a comparison with the estimates made at the beginning of your project and a discussion of discrepancies (discrepancies are not a negative. The important part is that you can explain them.) This is a critical part of the report. Provide enough evidence that you understand the performance results and their relationship to your programming/tuning steps.

    The reports should also express the effort that went into the project (indicate the average hours/week spent on the project, for each student, and the distribution of this time over the 15 weeks)

    Projects