Fractal Time Analysis Results

In my last post about fractals, I wrote up about how I wanted to analyze which seeds would take a long time to render and which would be fine.

The process goes like this:

real, im = -1, -1
times[IMG_SIZE][IMG_SIZE]
x, y = 0, 0

while x < IMG_SIZE
  y = 0
  while y < IMG_SIZE
    render
  
    times[x][y] = processing time
    y++
  x++

normalize times // min will be 0, max will be 255
output to bmp

So what you end up with is a greyscale image. The brighter it is, the longer it took to render.

The result is...well, amazing. Like really. See for yourself:



Does that look familiar to anyone?

I used a fairly small image size here (it was even smaller before). Problem is that rendering time is O(n²), so it takes 4 times as long when I double the size - originally I wanted to have an 800x800 image - turns out that'd take like a week for the thing to process! I might try tweaking the code to be multithreaded, that would probably half the time on my machine.