Lab2 - Optimize a physics-based simulation for KNL and Intel CPUs


Take the Newtonian particle simulation

You can also read about the implementation here

  • Measure the speed of the serial version on a single core
  • Find bottleneck
  • Use evolutionary optimization to speed up the program


  • Identify and report all bottlenecks
  • Implement a parallel version by using OpenMP on a Windows based computer as well as on the Xeon Phi computer
  • Use Microsoft Visual Studio C++ on Windows and Intel Parallel Studion XE compiler (icpc)
  • Report speedup of both versions as a function of cores


Prepare a HTML page discussing the implementation and results. Show the code. Show all steps.