I have previously successfully compiled a 64-bit version of FAST v8.16.00a-bjj using the GNU Fortran compiler on my Linux system and it works well. From what I have gathered, FAST is compiled in single precision by default. Since I would like to use BeamDyn to model turbine blades, FAST need be compiled in double precision, according to page 46 of the README file provided with the FAST source code:
If you are using BeamDyn, we suggest using an executable compiled in double precision (not single precision).
I tried to append -fdefault-real-8 to the FFLAGS compiler option in Compiling/makefile, and everything went well until an error occured:
TwrElRe = GetReynolds( VelHor, 2.0*TwrElRad, P%Wind%KinVisc )
Error: Type mismatch in argument ‘chordlen’ at (1); passed REAL(8) to REAL(4)
Has anyone successfully compiled the latest version of FAST v8 in double precision using gfortran? Am I missing something important? Any help is appreciated.