Dear Chris, > Is there a way to compute exp(x), where x is a "big" floating point > variable? there is currently no way in GMP, but this is possible with the MPFR library, which is built on top of GMP. See www.mpfr.org. Paul Zimmermann