alexander-0x80 / beauty-of-roots Goto Github PK

Hi Alexander: thanks for your code.

I run roots.py on my computer and when the degree is larger than 20 it got very slow. for degree >= 22 it's expected to run a ｗhole day.

I saw you used a list and append the roots to it, and list will get reallocated many times, I think this is the bottleneck in the performance.

May be you should declare a numpy array of size degree * 2**degree, and assign roots to this array. It runs much faster on my computer (45mins for d=22)

Another tip: when assigning HSV colors to the img, you used two for loops and this can be simplified further. np.where( Z > 0) would find the elements in Z that are >0, and then one can do vectorized arithmetic on Z then. the hsv_to_rgb() function in colorsys does not support vectorized arithmetic, but the function with the same name in matplotlib.colors does, and the color map "afmhot" does exactly the coloring scheme you did in plot.py.