Prev: XLF V13.1 release for AIX -- fully compliant Fortran 2003 standard
Next: GCC/gfortran 4.5.0 release
From: rudra on 14 Apr 2010 05:31
Friends, I wrote the code:
module mprintinit
use extras
contains

subroutine printinit(cz,sptyp,reltyp,xctyp,confA,confB,alloy_rad,&
ityp,nsp,numprocs)

use parameters

implicit none
integer::bit,bit2,opfl,ityp,nsp,numprocs
character(8) :: date
character(10) :: time
character(5) :: zone
integer,dimension(8) :: values
character(2),dimension(2)::cz
character(10)::alloy_rad
real(8),dimension(2)::rz
character(32) :: sys
character(15) :: nis
character(8)::confA,confB
character(80)::sptyp,reltyp,xctyp,ermsg,brief
character(2)::sps
character(1),dimension(4)::orbnm
integer,dimension(2)::atz

orbnm(1)='s'
orbnm(2)='p'
orbnm(3)='d'
orbnm(4)='f'
open(19,file="err.log")!,status='replace')
write(19,*)"!=============ERROR FILE FOR ASR
RUN=============!"
open(9,file="asrrun.log",status='replace')
write(9,*)"!=============OUTPUT FILE FOR ASR
RUN=============!"


bit = operating_system(sys)
bit2 = operating_sys(nis)
call date_and_time(date,time,zone,values)
rz(1)=z_a;rz(2)=z_b
call atomname(rz,cz,atz)
call alloy_spec(sptyp,reltyp,xctyp,confA,confB)


do opfl=6,9,3
write(opfl,*)""
write(opfl, &
'(("Job submitted on:",i2,".",i2,".",i4,2x,"at
",i2,":",i2))') &
& values(3),values(2),values(1),values(5),values(6)
write(opfl,*) Trim(fil)
write(opfl, '("Running on system",2x,a32,"in
machine",a15,"on" &
& ,i2," node")') trim(sys),trim(nis),numprocs


write(opfl,'("Alloy system:: ",2(a2,"{",f4.2,"}"))') &
& cz(1),x,cz(2),y
write(opfl,'("Underlying alloy ::",a3)') lattice
write(opfl,'("Alloy Radius:: ",a10)')alloy_rad
write(opfl,*) "The Electronic Configuration"
write(opfl,"(1x,a2,'::',a2,'s',a2,1x,a2,'p',a2,1x,a2,'d',&
& a2,1x,a2,a1,a2)") cz(1),trim(confA(1:1)),trim(confA(5:5)),&
& trim(confA(2:2)),trim(confA(6:6)),trim(confA(3:3)), &
& trim(confA(7:7)),trim(confA(4:4)),orbnm(4),trim(confA(8:8))
write(opfl,"(1x,a2,'::',a2,'s',a2,1x,a2,'p',a2,1x,a2,'d',&
& a2,1x,a2,'f',a2)") cz(2),trim(confB(1:1)),trim(confB(5:5)),&
trim(confB(2:2)),trim(confB(6:6)),trim(confB(3:3)), &
trim(confB(7:7)),trim(confB(4:4)),trim(confB(8:8))

write(opfl,*) trim(reltyp)
write(opfl,*) trim(xctyp)

if(spn==1)then
write(opfl,*)"Running spin-independent calculation"
else if (spn==2)then
write(opfl,*)"Running spin-dependent calculation"
else
write(opfl,*) "Not allowed spintype!!IDIOT!!EXITING!"
STOP
endif

!Printing The input parameters
write(opfl,*)"The Input parameters:-"
! write(opfl,'("Number of orbital calc. :",t36,i1)') norb
write(opfl,'("Nearest neighbour in real space:",t36,i2)')
nrsite
write(opfl,'("Number of site in AS Map:",t36,i6)') nasite
write(opfl,'("Number of atom type: ",t36,i2)') ntype
write(opfl,'("Dimension of AS Map: ",t36,i2)') ntsite
write(opfl,'("Total number of n.n. in AS Map: ",t36,i6)')
nsite
write(opfl,'("Number of recursion step: ",t36,i2)')maxrec
write(opfl,'("Number of orbital(s+3p+5d+7f): ", t36,i2)')
lorbit
write(opfl,'("Number of seeds for fitting: ",t36,i2)')
ienum
write(opfl,'("Number of points extrapolated: ",t36,i2)')
nkp
write(opfl,'("SCF loop: ", t36,i3)') nscf
write(opfl,'("Atomic number of atom A: ",t36,i2)') z_a
write(opfl,'("Atomic number of atom B: ", t36,i2)') z_b
write(opfl,'("Core electron atom A: ", t36,i2)') c_a
write(opfl,'("Core electron atom B: ", t36,i2)') c_b
write(opfl,'("Valence electron atom A: ",t36,i2)') e_a
write(opfl,'("Valence electron atom B: ",t36,i2)') e_b
write(opfl,'("Energy Range: ", t36,2(f5.3,1x))') emin,emax
write(opfl,'("POTPAR for orb. S: ",t36,es9.2)') ALPS
write(opfl,'("POTPAR for orb. P: ", t36,es9.2)')ALPP
write(opfl,'("POTPAR for orb. D: ",t36,es9.2)') ALPD
write(opfl,'("POTPAR for orb. F: ",t36,es9.2)') ALPF
write(opfl,'("Medulang: ",t36,es9.2)') ruban
write(opfl,'("Mixing scheme parameter: ", t36,es9.2)') amix

end do

end subroutine

subroutine prspn(nsp)
if (nsp==1)then
sps="UP"
else
sps='DW'
endif
do opfl=6,9,3
write(opfl,'("WORKING FOR SPIN",1x,a2)') sps
end do
end subroutine

subroutine pratm(ityp)
do opfl=6,9,3
write(opfl,'("WORKING FOR ATOM",1x,i1)') ityp
end do
end subroutine

end module

which when compiled by gfortran, gives error:

$ make FC=gfortran
gfortran -c -O3 -fopenmp prinit.f90
prinit.f90:52.47:

& ,i2," node")') trim(sys),trim(nis),numprocs
1
Warning: Extension: Tab character in format at (1)
prinit.f90:125.7:

write(opfl,'("WORKING FOR ATOM",1x,i1)') ityp
1
Error: UNIT specification at (1) must be an INTEGER expression or a
CHARACTER variable
prinit.f90:124.11:

do opfl=6,9,3
1
Warning: Deleted feature: Loop variable at (1) must be integer
prinit.f90:114.11:

sps="UP"
1
Error: Can't convert CHARACTER(1) to REAL(4) at (1)
prinit.f90:116.11:

sps='DW'
1
Error: Can't convert CHARACTER(1) to REAL(4) at (1)
prinit.f90:119.14:

write(opfl,'("WORKING FOR SPIN",1x,a2)') sps
1
Error: UNIT specification at (1) must be an INTEGER expression or a
CHARACTER variable
prinit.f90:118.11:

do opfl=6,9,3
1
Warning: Deleted feature: Loop variable at (1) must be integer
make: *** [prinit.o] Error 1

though opfl is clearly stated as integer. what is the reason? I cant
say this is a bug as same behavior is noticed with sun95(sun95 does
not complain about "Error: Can't convert CHARACTER(1) to REAL(4) at
(1)")
I am using gfortran version 4.4
From: glen herrmannsfeldt on 14 Apr 2010 06:15
rudra <bnrj.rudra(a)gmail.com> wrote:
> Friends, I wrote the code:
> module mprintinit
> use extras
> contains

> subroutine printinit(cz,sptyp,reltyp,xctyp,confA,confB,alloy_rad,&
> ityp,nsp,numprocs)

> use parameters

> implicit none
> integer::bit,bit2,opfl,ityp,nsp,numprocs
> character(8) :: date
> character(10) :: time
> character(5) :: zone
> integer,dimension(8) :: values
> character(2),dimension(2)::cz
(big snip)

> end subroutine

Note all those declarations at the beginning of subroutine printinit.

Note no declarations at all in subroutine prspn.
In prspn, sps and opf1 are REAL variables.

> subroutine prspn(nsp)
> if (nsp==1)then
> sps="UP"
> else
> sps='DW'
> endif
> do opfl=6,9,3
> write(opfl,'("WORKING FOR SPIN",1x,a2)') sps
> end do
> end subroutine

(snip of the rest)

-- glen
From: Richard Maine on 14 Apr 2010 15:48
m_b_metcalf <michaelmetcalf(a)compuserve.com> wrote:

> You could try an 'implicit none' statement to help with this!

Actually, he has one, but it didn't help for the same reason that the
declaration of opfl didn't help - it was in the wrong subroutine.

I recommend putting implicit none in the top level scoping unit - that
is in the module scope in this case. That way it applies to all the
inner scopes as well (unless overridden). Then it would have helped...

Well maybe it would have helped. The subroutines with the problem were
awfully short anyway. I'm not sure that a message saying that opfl was
not declared would have been any more obvious than the messages saying
that it needed to be integer. There is only so much that error messages
can plausibly do for you. Otherwise good error messages aren't likely to
be able to teach Fortran in you are missing the basics. I swear it
almost looks like the OP thought that the declarations in one subroutine
applied to every subroutine in the module.

But in cases where the problem subroutines are larger and where you
don't otherwise get an error message at all, having the implicit none at
the top level can help.

--
Richard Maine | Good judgment comes from experience;
email: last name at domain . net | experience comes from bad judgment.
domain: summertriangle | -- Mark Twain
From: robin on 14 Apr 2010 21:53
"rudra" <bnrj.rudra(a)gmail.com> wrote in message news:3c462607-fc1b-429a-80ca-db9d94c872d4(a)z6g2000yqz.googlegroups.com...
| Friends, I wrote the code:
| module mprintinit
| use extras
| contains
|
| subroutine printinit(cz,sptyp,reltyp,xctyp,confA,confB,alloy_rad,&
| which when compiled by gfortran, gives error:

Believe the compiler.

Like the others have said, put in IMPLICIYT NONE in EVERY subroutine,
not just the first.

As well as that, it's a good idea to include the subroutine name
in each END statement.


From: rudra on 15 Apr 2010 12:45
Friends,
Thanks for your help and it is solved.


 | 
Pages: 1
Prev: XLF V13.1 release for AIX -- fully compliant Fortran 2003 standard
Next: GCC/gfortran 4.5.0 release