[Pw_forum] example01 errors;conjugate-gradient algorithm;changing code functionals

Helen helen at fh.huji.ac.il
Wed May 30 13:48:07 CEST 2007


Hi,
I've been working a few days on the problem and have found that it occurs 
when ZGEMM is called in the ccalbec.f90 routine in the PW directory. The 
routine is called while calculating force_us_k. When the routine is called I 
get the error message:
forrtl: severe (174): SIGSEGV, segmentation fault occurred

Stack trace terminated abnormally.

I've checked all the input variables and they all seem to be the correct 
size and type, apart from perhaps npw which I thought should equal npwx? 
However I also tried running with npwx instead of npw in ZGEMM, whcih 
doesn't solve the problem. I've also checked that ZGEMM works on a little 
test routine I created so the ZGEMM is present and in principle works.

Any ideas what is causing this segmentation fault, maybe a memory leak?



A different question: the program is compiling and running on one computer, 
I tried running on a different computer and I get the error message

error while loading shared libraries: libguide.so: cannot open shared object 
file: No such file or directory

I tried to stop (unsuccessfully) dynamic linking by writing

./configure FFLAGS='-i-static'


or by writing

./configure FFLAGS='-static'



Any ideas of how to fix this problem?

Thnks for your help,

Helen


----- Original Message ----- 
From: "Axel Kohlmeyer" <akohlmey at cmm.chem.upenn.edu>
To: "Helen" <helen at fh.huji.ac.il>
Sent: Monday, May 28, 2007 6:14 PM
Subject: Re: [Pw_forum] example01 errors;conjugate-gradient 
algorithm;changing code functionals


> On Sun, 27 May 2007, Helen wrote:
>
> hello helen,
>
> HE> Thank you so much for your help
> HE> The output file si.scf.cg.out is indeed different from the reference 
> output
> HE> file. The output is identical up to the line
> HE>
> HE> convergence has been achieved
> HE>
> HE> Forces acting on atoms (Ry/au):
> HE>
> HE> and then the output of forces,stresses and writing to output data file 
> are
> HE> missing.
>
> which would mean, that your calculation was stopped and did not write
> the restart file, that is required for postprocessing.
>
> there are a number of possible reasons, most likely, the directory
> that you specified in the outdir parameter does not exist or is
> full.
>
> HE> I tried running other examples with the same problems.
>
> HE> My operating system is Linux i386
> HE> My cpu is Intel 3.2 GHz
> HE> My compiler is Intel (R) Fortran Compiler for 32-bit applications, 
> Version
> HE> 9.0, Intel (R) C Compiler for 32-bit applications, Version 9.0
>
> with the intel compilers, it is advisable to keep them updated to the
> latest patchlevel. however, your problem is not very likely related
> to a compiler problem. i wold investigate that when the more obvious
> explanations have been exhausted.
>
> HE> My libraries are
> HE> blas_libs='-L/opt/intel/mkl/8.0.1/lib/32 -lmkl_ia32 -lguide -lpthread'
> HE> lapack_libs='-lmkl_lapack'
> HE> fft_libs=' -lfftw '
>
> that looks reasonable. thanks for providing complete information
> on the compile. this is much appreciated.
>
> [...]
>
> HE> Is there anyway that when I run it will write where the code is 
> failing in
> HE> the error message?
>
> have a look at the Doc/INPUT_PW file for a list of common flags
> and their meaning. i would suggest setting verbosity='high' to
> narrow it down, but also, please check the validity of the entry
> for 'outdir' in your input.
>
> HE> Thank you again for your help,
> HE> Helen Eisenberg
> HE>
>
> cheers,
>   axel.
>
> -- 
> =======================================================================
> Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
>   Center for Molecular Modeling   --   University of Pennsylvania
> Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
> tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
> =======================================================================
> If you make something idiot-proof, the universe creates a better idiot.
> 




More information about the Pw_forum mailing list