high order number in computer simulation

Question

I wrote a Fortran code to solve PDE (like continuity equation) but the initial value of unknown are in order of 1.0e20 this imply my code to give NANE (not number) of infinity because it's multiplying or dividing big number what can I do to run simulation with such big number?
the equation are : Poisson equation and continuity like equations

Answer 1

You can use extended precision, real*8, or double precision (which are a 64-bit floating point representations) as the type instead of real (which is 32 bits). That will give an exponent range of at least 308 instead of the smaller range of 38.