I was working on an HTML5 application that draws fractal images defined by iterative functions such as the famous Mandelbrot fractal. http://en.wikipedia.org/wiki/Mandelbrot_set
I understand that Javascript uses only one data type for double precision floating point numbers. Due to this limitation, I cannot make my program as accurate as I would like. There are images and videos on the Internet that are fractal images so precise that they use hundreds of digits after the decimal point. See an example here: http://www.youtube.com/watch?v=foxD6ZQlnlU
I heard about high-precision calculation libraries for Javascript, but the ones I saw all use strings to represent numbers, which is not compatible with asm.js.
Is there a practical way to do high precision computing with asm.js? How high? I would say arbitrarily tall. Accuracy must be adjustable.
If I had to take a picture in the dark, I would suggest that since asm.js supports integer arithmetic, I could somehow make it work using arrays of integers to represent a large number of digits after the decimal point. I could represent numbers as strings and have a small function to split them into arrays that need to be passed to asm.js.