The commonly available CPU chips implement 64-bit floating point arithmetic per IEEE 754, and the limits of the representation are as described in http://en.wikipedia.org/wiki/IEEE_754-2008 To do otherwise (as in Mathematica) means accepting an order of magnitude (or more) degradation in performance.
emax=1023 means the maximum exponent in base 2. This means the maximum number in base 10 is just under 10^e=.1024*10^.2, where e 1024*10^.2 308.255 The following is a divertissement on working with large exponents: http://www.jsoftware.com/jwiki/Essays/Large%20Exponent%20Real%20Arithmetic ----- Original Message ----- From: DIETER ENSSLEN <[email protected]> Date: Tuesday, September 29, 2009 12:15 Subject: Re: [Jgeneral] anyone know the math/computing reason for calling exponents an end at +- 1e328 To: General forum <[email protected]> > microsoft excel appears to have practically identical exponent limits > > thanks ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
