Skip to main content
SciNet
  • Home
  • All Courses
  • Calendar
  • Certificates
  • SciNet
    Main Site Documentation my.SciNet
  • CCDB
  • More
Close
Toggle search input
English
English Français
You are currently using guest access
Log in
SciNet
Home All Courses Calendar Certificates SciNet Collapse Expand
Main Site Documentation my.SciNet
CCDB
Expand all Collapse all
  1. Dashboard
  2. PHY1610 - Winter 2022
  3. 8. Parallel Wave

8. Parallel Wave

Completion requirements
Opened: Monday, 28 March 2022, 12:00 AM
Due: Sunday, 3 April 2022, 11:59 PM

We're returning once more to the damped 1d wave equation, this time, to parallelize it.  Take the initial code from assignment 5 (so without blas), and take a new version of longwaveparams.txt below (chosen such that the calculation takes a bit longer).

Your task:

  • Parallelize the time-evolution part of this code (advance_wave) with OpenMP.  Adjust the Makefile accordingly.
  • Perform a scaling analysis of the parallelized code on a compute node, with the number of threads ranging from 1 to 16. Do so by writing either one or 16  job scripts.
  • Timing can be done in the job script using the linux 'time' command.
  • From the timings, compute the speed-up and plot the speed-up as a function of the number of threads.
  • Fit the speed-up results to Amdahl's law and deduce the serial fraction of the code.

Next:

  • Parallelize the conversion of binary numbers to text in the output_snapshot function with OpenMP.  
  • Perform the above scaling analysis again, produce a new speed-up vs. number of threads plot, and determine the new serial fraction.

Submit your code, Makefile, job script(s), git log, timing data and scaling plots for both cases.

  • longwaveparams.txt longwaveparams.txt
    28 March 2022, 12:50 AM
  • wave1dmod.zip wave1dmod.zip
    28 March 2022, 12:50 AM
Contact site support
You are currently using guest access (Log in)
Data retention summary


All content on this website is made available under the Creative Commons Attribution 4.0 International licence, with the exception of all videos which are released under the Creative Commons Attribution-NoDerivatives 4.0 International licence.
Powered by Moodle