Double precision in fortran accurate up to 16 digits

Asked Apr 25 '16 at 18:49

Active Apr 25 '16 at 18:49

Viewed 498 times

I am using FORTRAN 95 using Silverfrost Plato where I am trying to make real double precision variables. Currently I use 'double precision' which gives accuracy up to 12 decimal digits. But I wanted to know if we could have digits accurate up to 16 digits. Interestingly, when I tried using 'real(kind=3)' there were some residual values showing up in the last few digits that ended up making the results go wrong.

I am unable to find a proper solution to this. Any help or links to relevant documentation would be greatly appreciated.

asked Apr 25 '16 at 18:49

Suga

2

What is the result of `selected_real_kind(16)`? – francescalus Apr 25 '16 at 19:00
2

Directly selecting `kind=3` is not portable. It will actually not compile at all for most compilers. If you have a problem with your code, show the code. Read http://stackoverflow.com/questions/838310/fortran-90-kind-parameter – Vladimir F Героям слава Apr 25 '16 at 20:26

Double precision in fortran accurate up to 16 digits

0 Answers0