J.B. Langston

@trathborne __float128 I/O support is now committed. I am not getting the weird zoom issues where the center changes on Linux; only saw them on Windows.

Hmm, now I can't reproduce the zooming issues on Windows either... not sure what was going on earlier. Anyway, I guess 128-bit floating point can at least be considered an...

Thanks, I understand you don't have unlimited time. Even small contributions are welcome. I've gone years at a time without touching XaoS myself. Right now I'm on a roll though...

Saving quads has been fixed to use 34 digits of precision in the latest commit.

Just commenting it out in the project file is probably your best bet for now. By the way, when you comment out USE_LONG_DOUBLE, it uses double precision, not single. Single...

That won't work unfortunately. SIMD registers are not general-purpose floating point registers. True, they are 256-bit or even 512-bit, but they are used to pack multiple 64-bit or 32-bit numbers...

GPU support would require a near-total rewrite, I think. Also, with a GPU you don't really need the XaoS zooming algorithm to do real-time zooming; you could just brute force...

Ideally the functions that use number_t should be templatized using C++ templates ( https://en.cppreference.com/w/cpp/language/templates ) so that they can be instantiated at compile time to use whatever type we want,...

I would convert the fractal_context struct to a class and make all the functions that operate on that struct members of the class. Then you can instantiate multiple instances of...

Yes, I think so. I am not aware of any way to avoid copying the data. On Wed, Aug 5, 2020 at 3:19 PM Anurag Aggarwal wrote: > Thanks. I...