What is the fastest semi-arbitrary mathematical precision library?

Question

I use the long double in program C to calculate the 2D images of the Mandelbrot set, but I want to have higher accuracy for deeper magnification.

Is there any performance gain from a math library with arbitrary accuracy, which can limit the accuracy if necessary, and not jump with long double accuracy directly to arbitrary accuracy?

What is the fastest library of arbitrary math precision?

Answer 1

"Fastest" will depend somewhat on your platform and intended use.

MPFR library

GMP

This wiki article contains links to several libraries.

Answer 2

If you need higher accuracy, see qd at http://crd.lbl.gov/~dhbailey/mpdist/ .