Performance comparison in simulation of Mandelbrot set fractals using Numba
Gilbert Gutabaga Hungilo, Gahizi Emmanuel, Pranowo Pranowo · AIP conference proceedings · 2020
This paper aims to describe numerical simulation of the Mandelbrot Set Fractals by comparing performance of Python function which generate the Mandelbrot set fractals in three cases, using pure Python, Numba and Numba CUDA which run Python function in parallel. The fractal is simulated using iterated function system (IFS) based on number of iteration and size of generated fractals. The result shows that the Python function run in parallel using Numba CUDA is 1856 times faster than pure Python function on CPU and 14 times faster than Python function using Numba JIT compilation. The comparison measure used is the execution time of the Python function using %timeit Python magic command. Fractals have many application which needs faster computation like video games creation, flight simulation, and Cryptography for security purpose.