Oddbean new post about | logout
▲ ▼
Simon Tatham | 1 years ago (raw) | root | parent | reply | flag 
 @c30feace @b4c50e1b @218dc8a6 hmm, if picking a different finite set of coefficients gives you a different fractal, what happens if you let a finite number of points wander continuously around the plane and make an animation of the results?
▲ ▼
John Carlos Baez | 1 years ago (raw) | root | parent | reply | flag 
 @088924ba - sounds fun, but it took Sam Derbyshire 4 days to generate this one picture using Mathematica, so to create a movie with many frames you'd have to lower the degree of the polynomial, only display a smaller region, get a better computer and/or be more clever about programming. 

This one picture was originally 5 gigabytes; this is a reduced version.

 @c30feace @218dc8a6

https://media.mathstodon.xyz/media_attachments/files/111/156/470/040/224/015/original/8ae93102c872d733.png
▲ ▼
74b1555a | 1 years ago (raw) | root | parent | reply | flag 
 @088924ba @c30feace @b4c50e1b I suspect one could relatively easily be more clever about programming. Has anyone tried to implement it for a GPU, for example? It feels intuitively like the sort of thing that could make good use of that style of parallel programming.
▲ ▼
Tomáš Kalvoda | 1 years ago (raw) | root | parent | reply | flag 
 @b4c50e1b @088924ba @218dc8a6 My impression is that Mathematica is not well known for its fast numerical computations. That's why I tried Julia, out of curiosity.

If I remember correctly, it was faster then days. Notice also, that I reached polynomials of higher degree. Anyway, gigabytes were indeed flying around 😂.