I'm going to write a modified version of MakeGraph by mike3

[JmsNxn]

Wish me luck

 Done!

Lmao!

I'm a little stuck. But I have managed to make it so that instead of pixel by pixel run, we have \(n \times n\) pixels; with the origin as the sample point. So it makes it a little blocky. But it still displays the function, especially if you just "lower the resolution" by a small amount. I'm having a lot of trouble at writing the splining. But I think this is still an awesome tool. Despite what turns out. If mike3's program grows like \(O(N)\)--for \(N^2\) amount of pixels. I've at least got it to \(O(N^{1-\delta})\) for \(\delta >0\) but \(\delta \approx 0\). But the final graph is a bit more blocky depending on how big \(\delta\) is. It's really not that noticeable. But it's taking way faster to graph   Literally quartering graph time   With only a little bit of "lower resolution" on the graph

Once I figure out how to efficiently write splining, so without killing the \(O(N)\), we'll be coasting! Instead of setting each \(4 \times 4\) pixel block to the reference point, we spline towards the four corners of reference points... it'll look the exact same and run like \(O(N^{1-\delta})\) rather than \(O(N)\). It'll save us so much fucking time!

My solution so far is to calculate the reference point, and its derivative; and then consider each 4x4 pixel block as the reference point plus its derivative about this point. I'm having a little trouble getting the code tho. Lmao!

I think I nearly got it though. Where, we can "lower the resolution" which causes "increase the speed of the compiling of the graph".

EDIT:

Okay, I've added that the function mike3's program takes must be differentiable. This works fine for everything I want.

I just need to fix some dumb shit. But this makes "basically" mike3's graph. And if you want you can still just run his graph; program accounts for that. But if you want to sacrifice a bit of quality, you gain a bunch of speed!

I Fucking did it! Just let me do more test runs and shit. Don't want errors or anything. Just want to hit all the fringe cases....