[Solved] problems compiling FEHM with intel visual fortran

[jbensabat]

Hello
I am trying to build FEHM with intel Visual fortran under windows.
I get a large number of errors of the same type around this portion of code

c add coding to save upwind position
if(iad.le.iad_up) then
fid=0.5
axyd=t8(neighc)
if(axyd.lt.0.0) fid=dnwgt
if(axyd.gt.0.0) fid=upwgt
t9(neighc)=fid
c
call setbit(nbits,neighc,upwind_l(iz4m1),fid)
c
else
if(bit(nbits,neighc,upwind_l(iz4m1))) then
t9(neighc)=1.0
else
t9(neighc)=0.0
endif
endif
the error arises because subroutine setbit accepts integers only but the array upwind_l if real.
I saw that in some routines any reference to setbit was commented
Is it ok to do so ?
thanks
jac

[daniellivingston]

Hi Jac, try commenting out bit and setbit:

I would say that if the new Intel compiler does not work with 'bit' and 'setbit' we comment out those. Macro 'itup' will be essentially disabled but very few of us have used it. I think no test cases do.

[jbensabat]

Hi and thank you for your support
I resolved this issue by checking out the flag "Check Routine Interfaces" in the "Diagnostics" option of the fortran workspace properties. I was able then to build FEHM. There are two types of issues: 1) related to the setbit subroutine - the dummy arguments are all integer (4 bytes) while the third argument in the calling is of type real*8. even if compilation goes thru I am not sure that results could be reliable; 2) there are many cases where the size of dummy arrays are larger than the ones in the calling sequence - but this is not a problem.

thanks
jac

[daniellivingston]

Great! Thanks for posting your solution. We'll include it in build documentation. I'm marking this as closed, then. Please re-open if any additional issues related to this topic appear.