workshop_ossc

A parallel agent-based model in R

The hands-on sessions are about going through the code together and understanding what each line approximately does.

Introduction - 5 minutes

  1. Assign someone to be the timekeeper & someone else with experience with R/RStudio to be the screen-sharer.
  2. Shortly (!) get to know each other.

Agent-based model - 20 minutes

  1. Open the file 01_example_sim.R in RStudio.
  2. Discuss and run the code until line 28.
  3. What is the proportion of happiness in the last subpopulation in result_2?
  4. Compile the C++ version of the ABM using the code on line 33.
  5. Now run the same ABM as in result_2 by using the C++ version (line 36). Do you notice the speedup?
  6. Run the rest of the file. Discuss what is happening.

Parallel programming - 20 minutes

  1. Open the file 02_parallel_eval.R in RStudio.
  2. Run the file until line 22.
  3. Run the following code: analysis_function(1). Discuss what this does.
  4. Now run & discuss the code until line 37. What does the plot show & what does the mean value represent?
  5. Next, discuss & run the code until line 54. Edit the number of threads so that there are 1 or 2 threads left on your machine. For example, if detectCores() returns 8, then line 48 should be n_threads <- 6 or n_threads <- 7.
  6. Run & discuss the remainder of the file.
This site is open source. Improve this page.