The ABINIT Users Mailing List ( CLOSED )

[Petr Sestak <sestak@kn.vutbr.cz>]

Hi Natalie and Aldo,

thank you for your help.
I have made this change in the code:

------------------------------------------------------------------------------------
    !  program to perform self-consistency loop

    TYPE (Anderson_Context) , POINTER :: AErho

    REAL(8) :: rmix,xocc,qf,small,range,zeff,delta,v1,v2,v3,v4,x,y
    REAL(8) :: qcal, rescale,cnvrg,emin,ecoul,eexc,etxc,eone,etot,ekin
    INTEGER :: icount,i,j,k,it,start,np,ierr,nroot
    INTEGER :: is,ip,id,jf,ig,io,l,nfix,ir,nzeff,loop,jierr
    REAL(8), ALLOCATABLE :: denout(:)
    INTEGER :: fcount=0

    v1=conv1;v2=conv2;v3=conv3;v4=conv4
    n=AEGrid%n; h=AEGrid%h
    rmix=rimix
    cnvrg=worst
    small=-1.d0 <====================================== added line
    IF (small.LT.0.d0.OR.small.GT.1.d0) small=small0

    WRITE(6,*) 'Density convergence parameter set to at least',cnvrg
    !
    ALLOCATE(denout(n),STAT=k)
    IF (k /= 0) THEN
       WRITE(6,*) 'Error in denout allocation ', n,k
       STOP
    ENDIF

    !    start iteration loop
---------------------------------------------------------------------------------------------------

After that, I have compiled the atompaw code and the compilation was
without an error.
I did some tests of the atompaw using all input files of Ti and O on the
web:
http://www.wfu.edu/~natalie/papers/pwpaw/periodictable/atoms/Ti/index.html
http://www.wfu.edu/~natalie/papers/pwpaw/periodictable/atoms/O/index.html
and everything is seems to be alright. I haven't seen any error. (tests
at our cluster => ifort)


Before I have seen the message from Natalie,
I have changed the compiler (ifort => gfortran), lapack and blas library
(intel MKL => standart libraries in OpenSUSE 11.1 distribution) and
atompaw code is also working (tests at my personal PC => gfortran).

Best regards,

Petr Sestak

 

Natalie Holzwarth wrote:
> Dear Petr,
>      I think that you have found a small error in the code due to the fact
> that the parameter "small" is used on line 290 of the aeatom.f90
> subroutine before it is defined.   If you add the line
>
> small=-1.d0 just before line 290 which reads:
> IF (small.LT.0.d0.OR.small.GT.1.d0) small=small0
>
> the compiler might be happier.  (If you don't use the run time checking
> parameter in the compilation you will not see this bug which explains why
> we didn't notice it before.)   Please let me know if this works and then I
> will fix the code on the web.
>
>                   Thanks, kindly,
>                         Natalie Holzwarth
>
> On Thu, 3 Dec 2009, Petr Sestak wrote:
>
>   
>> Dear Abinit users,
>>
>> I am trying to compile and use the AtomPAW code.
>> I have compiled the AtomPAW code (atompaw_2.2.patched.tgz) with this
>> make file:
>> -----------------------------------------------------------------------
>> LIBS = /opt/intel/mkl/8.0/lib/32/libmkl_lapack95.a \
>>        /opt/intel/mkl/8.0/lib/32/libmkl_blas95.a \
>>        /opt/intel/mkl/8.0/lib/32/libmkl_lapack.a \
>>        /opt/intel/mkl/8.0/lib/32/libmkl_ia32.a \
>>        /opt/intel/mkl/8.0/lib/32/libguide.a \
>>        /usr/lib/libpthread.so
>>
>> F90 = ifort
>> F90FLAGS = -O0 -g -xW -mp -C    #I tried default ifort optimization -O2
>> => no effect.
>> LDFLAGS = $(F90FLAGS)
>> #ifort => Version 10.1
>> ----------------------------------------------------------------------
>> or LIBS = -L/opt/intel/mkl/8.0/lib/32 -lmkl_lapack95 -lmkl_blas95
>> -lmkl_lapack. -lmkl_ia32 -lguide -lpthread
>> ----------------------------------------------------------------------
>>
>> With this makefile, the code was compiled without error. However, when I
>> tried use this binary file (atompaw < O_GGA_input) along with some input
>> from
>> http://www.wfu.edu/~natalie/papers/pwpaw/periodictable/atoms/NO/O_GGA_input
>> to check if everything is alright I got this error:
>> ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>> ./atompaw < O_GGA_input
>>  enter atomic symbol and atomic number
>>  exchange-correlation type -- LDA-PW(default) or GGA-PBE
>>  further optionally (space) "nonrelativistic/scalarrelativistic" keyword
>>  further optionally (space) "point-nucleus/finite-nucleus" keyword
>>  optionally (space) "loggrid/lineargrid" keyword if appropriate
>>      further optionally n (number of grid points)
>>                         r_max (max. grid radius)
>>                         r_match (exact value of r(n))
>>  Calculation for atomic number =            8
>>  enter maximum principal quantum numbers for s,p,d,f,g
>>    2   2   0   0   0
>>            3  orbitals will be calculated
>>   Below are listed the default occupations
>>  n  l     occupancy
>>  1  0      2.0000000E+00
>>  2  0      2.0000000E+00
>>  2  1      6.0000000E+00
>>   enter np l occ for all occupations for all revisions
>>    enter 0 0 0. to end
>>   Corrected occupations are:
>>  n  l     occupancy
>>  1  0      2.0000000E+00
>>  2  0      2.0000000E+00
>>  2  1      4.0000000E+00
>>
>>  nuclear charge    =            8
>>  electronic charge =    8.00000000000000
>>  net charge        =   0.000000000000000E+000
>>  InitGrid: -- linear         20001  2.500000000000000E-003
>>    50.0000000000000
>>            1           1           0   2.00000000000000
>>   -49.0000000000000
>>            2           2           0   2.00000000000000
>>   -6.25000000000000
>>            3           2           1   4.00000000000000
>>   -1.00000000000000
>>  qcal electrons =    8.00000005162370        8.00000000000000
>>  Perdew - Burke - Ernzerhof GGA
>> forrtl: severe (193): Run-Time Check Failure. The variable
>> 'aeatom_mp_scfloop_$SMALL' is being used without being defined
>> Image              PC        Routine            Line
>> Source
>> atompaw            083BB3E7  Unknown               Unknown  Unknown
>> atompaw            083BAA07  Unknown               Unknown  Unknown
>> atompaw            08381F9A  Unknown               Unknown  Unknown
>> atompaw            0835AFEA  Unknown               Unknown  Unknown
>> atompaw            0835BF9C  Unknown               Unknown  Unknown
>> atompaw            08056D1C  Unknown               Unknown  Unknown
>> atompaw            08056BCB  Unknown               Unknown  Unknown
>> atompaw            08082642  Unknown               Unknown  Unknown
>> atompaw            0804C641  Unknown               Unknown  Unknown
>> libc.so.6          B7DBB390  Unknown               Unknown  Unknown
>> atompaw            0804C5A1  Unknown               Unknown  Unknown
>> -----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>>
>> Does anybody know where I did a mistake? I tried to compile the older
>> version of AtomPAW (atompaw_2.1.tar.gz) but I have got the same error. I
>> tried make a compilation with dynamic linking and still no change. I
>> also don't know if  an error is in the compilation or in the input file.
>>
>> Best regards
>>
>> Petr Sestak
>>
>> ----------------------
>> Dr. Petr Sestak
>> Institute of Engineering Physics
>> Faculty of Mechanical Engineering
>> Brno University of Technology
>> Technicka 2896/2, Brno 616 69, Czech Republic
>>
>> Tel.:          +420 541142812
>> Mail:          sestak@fme.vutbr.cz
>>
>>
>>     
>
> N. A. W. Holzwarth                    email: natalie@wfu.edu
> Department of Physics                 www: http://www.wfu.edu/~natalie
> Wake Forest University                        voice: 336-758-5510
> Winston-Salem, NC 27109-7507            fax: 336-758-6142
> U. S. A.
>
>