Yes, I hit this too. Sorry I didn't report it - figured it was more
widespread, though on quick exam I did not see what the compiler did not
like (ie., may well just be a *.#! compiler bug). I was on RHEL 32 bit,
ifort 8.1.025, which in general I have liked better than 9, but partly
because of better compatibility with current MKL releases.
Regards - Bob
----- Original Message -----
From: "David Mathews" <David_Mathews.urmc.rochester.edu>
To: <amber-developers.scripps.edu>
Sent: Monday, March 27, 2006 6:37 PM
Subject: amber-developers: qmmm2/neb_gb;compile problem with evb_input.f
(ifort on linux)
> Hi All:
>
> Ross's fix will take of this bug with neb_gb. I don't know why
it
> wasn't spotted before; sorry.
>
> Also, the current CVS tree will not compile for me on Red Hat
> Enterprise Linux (32 bit) using ifort 8.
> I get:
> ifort -c -w95 -mp1 -O0 -tpp7 -FR -o evb_input.o _evb_input.f
> fortcom: Error: _evb_input.f, line 749: Unbalanced parentheses
> , emap(n)% lambda, n = 1, nbias )
> ---------------------------------------------^
> fortcom: Error: _evb_input.f, line 747: Syntax error, found ',' when
> expecting o
> ne of: :
> write(6,'((5x,2(a,i4),a,2x,a,f10.5))') ( 'emap(', emap(n)% ist,
','
> &
> ------------------------------------------------------^
> fortcom: Error: _evb_input.f, line 748: Syntax error, found ''' when
> expecting o
> ne of: ( <IDENTIFIER> <CHAR_CON_KIND_PARAM> <CHAR_NAM_KIND_PARAM>
> <CHARACTER_CON
> STANT> <INTEGER_CONSTANT> ...
> , emap(n)% jst, ') ::', 'lambda = ' &
> -----------------------------^
> compilation aborted for _evb_input.f (code 1)
>
>
> This is with "make serial" after "config -p4 ifort_ia32"
>
> Has anyone else encountered this problem?
>
> Thanks,
> Dave
>
>
> At 04:34 PM 3/27/2006, you wrote:
>>Divide by zero exception in this function:
>>
>>subroutine normalize(vector,dim)
>> !normalize a vector of components of a dim dimensional vector
>> !Very small vectors need to be considered zero for simulated
>> annealing.
>> !These are trapped below and not normalized.
>> implicit none
>> _REAL_ vector(*),vlength
>> integer dim, i
>>
>> vlength = 0.d0
>> do i = 1, dim
>> vlength = vlength + vector(i)*vector(i)
>> end do
>> vlength = 1.0d0/sqrt(vlength) !DIVIDE BY ZERO
>> if (vlength<1000.0d0) then
>> vector(1:dim) = vector(1:dim)*vlength
>> else
>> vector(1:dim) = 0.d0
>> end if
>>
>>end subroutine normalize
>>
>>at line 140 of pimd_force.f
>>
>>You get a divide by zero when all the components of vector are == 0.
>>Again, not sure why we're not seeing this on othe
> ore familiar with this part of the code to
>>determine whether the bug is that normalize can't handle the zero vector
>>or that the zero vector should never have been passed to normalize.
>>
>>John
>>
>
> ___________________________________________________
> David Mathews, MD, PhD
> Assistant Professor of Biochemistry & Biophysics
> and of Biostatistics & Computational Biology
> University of Rochester Medical Center
> Room 3-6830
> 601 Elmwood Avenue, Box 712
> Rochester, New York 14642
> http://rna.urmc.rochester.edu
>
>
>
>
>
>
Received on Wed Apr 05 2006 - 23:49:34 PDT
