Complex numbers

[ziofil]

Is there a way to use complex numbers with re/im parts in quad precision that doesn't require you folks to explicitly support it as a dtype of its own?

[ngoldbaum]

No, I think we'd need to add a replacement for the Complex256 DType in NumPy that's backed by e.g. SLEEF.

I know @SwayamInSync was messing with the Mandelbrot fractal to test the quad precision DType, but I think he was doing everything with arrays storing real and imaginary parts.