/pfn-intern-coding-tasks-2019-performance_optimization

My submitted program for PFN intern coding tasks (2019, Performance Optimization)

Primary LanguagePython

PFN intern coding tasks (2019, Performance Optimization)

Environment

  • Python 3.7.2
  • ply 3.11
  • numpy 1.16.2
  • scipy 1.2.1

How to run

  • python opt.py [N] | python interpreter.py [N]
    • run FFT
  • python opt.py [N] | python cost.py
    • show calcuration cost
  • python test_opt.py
    • run unit test

Result

I implemented Cooley-Tukey FFT and Rader's algorithm.

  • Sum of cost from N = 2 to 256 (calculated by cost.py)
    • naive.py (original): 16875645
    • opt.py: 3564224 (21.1%)
  • max error
    • 8.83390668259562e−13 (when N=212, initializer=sin)

License

Unknown, beucase this repository includes source code base on https://github.com/pfnet/intern-coding-tasks

I don't encourage you to include these source codes into other program.