I had previously posted a thread: "floating point precision: is true double possible?", but it seems that the answer to that question was basically "no".

...so, in lieu of floating point math on true (64 bit) doubles, I'll need to resort to using fixed point math. This would require re-writing all of the normal floating-point functions (sine, tangent, exponentiation, etc.), but I'm sure this has been done before -- probably many times over. I'm curious if there are recommendations for fixed-point math libraries that have been successfully tested on the Arduino.

Alternately, an arbitrary precision library or library implementing 64-bit floating points would also be extremely useful. Any recommendations would be appreciated!