Hi everyone, I have a fortran code with several subroutines in different .f files and an include file with all common variables. This code is in single precision. What is the easiest way to make it double precision.

Thanks MM

Through compiler options (e.g., -r8). No change to code needed, assuming everything is defined as "real" and not "real*4".

If things are specifically declared single precision, you'll have to edit every routine.

i think in that case, maybe a "search n replace" should do the job, although one has to be careful.

hi MM:

can you pls advise on how you compile the fortran program, in the case of you have different subroutines in different .f files??

So far, I inhclude all the subroutines (and mother program) in 1 single file, which is difficult for one to read through...

I am using the microsft visual fortran currently..

Hi kc That's easy, 1)Put each sub in a different .f file. 2)Open the mother file. 3)Compile mother file 4)Go to Project/Add to project/Files 5)Select all your subroutines files and click ok

Now you have all Subs and you can run your mother file

HI pc,

I didn't get you, There is no option to select in the time of compiling. Could you explain it more.

Thanks mm

Hi Pc, I didn't get you, there is no option to select in time of compiling. Could you explain more.

Thanks MM

What operating system and compiler are you using. I've found the relevant switches in all the mainstream UNIX compilers, plus Absoft (for Linux) and DVF (for Windows).

Nearly every Fortran compiler I've used (Linux/Unix systems) has an option to make all variables defined as "real" to be interpreted as "real*8" or "real*4" globally. It could be something like -r8 or -realsize=8, or something similar. Your compiler help should list such an option.