[Wien] [EXTERNAL] Re: Bug in nn

[Parker, David S.]

Thank you for your timely attention to this matter. Unfortunately, the revised "nn.f"
Appears to call a function "angle" that is not contained in the code - see the compile.msg
Below. "angles" is a defined array, but the code will not compile presently. I am using
WIEN2K_19.1 As always, your help is much appreciated.

With best regards,

David Parker
Group Leader, Materials Theory
Lead, “Developing Substitutes”
Critical Materials Institute

ifort  -O -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML -traceback -assume buffered_io -I/software/tools/compilers/intel_2017/mkl/include
  -c nn.f
ifort -o ./nn struk.o variable_fields.o nn.o dirlat.o ord2.o reduce_alloc.o bvan.o  -O -FR -mp1 -w -prec_div -pc80 -pad -ip -DINTEL_VML -tra
ceback -assume buffered_io -I/software/tools/compilers/intel_2017/mkl/include  -L/software/tools/compilers/intel_2017/mkl/lib/ -lpthread -lm
 -ldl -liomp5 
nn.o: In function `MAIN__':
nn.f:(.text+0x4d0d): undefined reference to `angle_'
/usr/bin/ld: link errors found, deleting executable `./nn'
/usr/bin/sha1sum: ./nn: No such file or directory
make: *** [nn] Error 1

David Parker
Group Leader, Materials Theory
Lead, “Developing Substitutes”
Critical Materials Institute
 
 

On 12/14/20, 6:14 AM, "Wien on behalf of Peter Blaha" <wien-bounces at zeus.theochem.tuwien.ac.at on behalf of pblaha at theochem.tuwien.ac.at> wrote:

    You can call it a bug or a "feature", because we usually try to avoid 
    that something stupid can appear (huge output file) if a user specifies 
    unusual input (distances beyond a certain limit).

    Anyway, in order to not "delay your research" further, I've looked into 
    this issue.

    It has NOTHING to do with CXZ lattices, but with a limit in the number 
    of "neighboring shells". This is set to 100 shells and the code was 
    jumping out of the loop if this has been reached. This would happen in 
    all lattice types once 100 shells are reached, but of course in low 
    symmetry structures this is reached earlier.

    I changed it such that it is not jumping out of the loop, but will 
    continue to print.

    -- 

                                           P.Blaha
    --------------------------------------------------------------------------
    Peter BLAHA, Inst.f. Materials Chemistry, TU Vienna, A-1060 Vienna
    Phone: +43-1-58801-165300             FAX: +43-1-58801-165982
    Email: blaha at theochem.tuwien.ac.at    WIEN2k: http://www.wien2k.at
    WWW:   http://www.imc.tuwien.ac.at/TC_Blaha
    --------------------------------------------------------------------------

    Subject: FW: [Wien] [EXTERNAL] Re: Bug in nn
    From: "Parker, David S." <parkerds at ornl.gov>
    Date: 12/13/20, 5:06 PM
    To: "wien at zeus.theochem.tuwien.ac.at" <wien at zeus.theochem.tuwien.ac.at>
    Dear all:



    There is apparently a bug in nn for lattices of the CXZ and P1 symmetry 
    that is precluding

    Printing of extended distances in case.outputnn. The issue is not 
    related to distmax – I have

    Manually typed in values as large as 40 Bohr and the output is still 
    truncated to about 7.6 A

    (see case.outputnn and struct file below). This is significantly 
    delaying research results.



    Any help would be much appreciated.



    Thanks, David Parker