[Pw_forum] segmentation fault when using pwcond.x

Alexander Smogunov smogunov at sissa.it
Fri Dec 26 11:15:12 CET 2008


Dear Wang
Try to replace the subroutine PWCOND/init_orbitals.f90
by the one from espresso-4.0.3 distribution. 
This might fix the problem.
regards, Alexander

On Thursday 25 December 2008 19:29, Wang Qinjing wrote:
> Dear all:
>
> I am using pwcond.x to calculating the transimission of a system. The
> pwcond.x is compiled by openmpi 1.1.4. The pwcond.x was tested and
> confirmed get the same results as the example 12.
>
> I wrote the input file just follow the example 12. The necessary files from
> the scf calculation is in the directory ./outdir. The input file is as
> follows:
>
> --------------
>  &inputcond
>     outdir='./outdir',
>     prefixl='lead',
>     prefixs='scattering',
>     tran_file='trans.file',
>     ikind = 1,
>     energy0=3.d0,
>     denergy=0.d0,
>     ewind=4.d0,
>     epsproj=1.d-2,
>     nz1 = 1,
>  /
>     1
>     0.0  0.0  1.0
> 18
>   3.0
>   2.7
>   2.5
>   1.6
>   1.0
>   0.9
>   0.1
>   0.0
>  -0.1
>  -0.25
>  -1.15
>  -1.45
>  -1.9
>  -3.0
>  -4.0
>  -5.0
>  -6.0
>  -6.2
> ---------------------
>
> After I typed the command "pwcond.x < cond.in ", the last line of the
> standard out is as follows:
>
> -----------------------
>         k(    1) = (   0.0000000   0.0000000), wk =   1.0000000
>
>  ngper, shell number =          537         108
> Signal:11 info.si_errno:0(Success) si_code:1(SEGV_MAPERR)
> Failing at addr:0x98f7e70da0
> [0] func:/opt/mpi/ifort/openmpi-1.1.4/lib/libopal.so.0 [0x2a9588366a]
> [1] func:/lib64/tls/libpthread.so.0 [0x31c8e0c420]
> [2] func:pwcond.x(local_1_+0x110e) [0x478778]
> [3] func:pwcond.x(local_+0xab6) [0x474fdc]
> [4] func:pwcond.x(do_cond_+0x8ed) [0x45a565]
> [5] func:pwcond.x(MAIN__+0x48) [0x4584b4]
> [6] func:pwcond.x(main+0x2a) [0x44c2a2]
> [7] func:/lib64/tls/libc.so.6(__libc_start_main+0xdb) [0x31c851c4bb]
> [8] func:pwcond.x(free+0x52) [0x44c1ea]
> *** End of error message ***
> Segmentation fault
> -----------------------
>
> What may the problem be?
>
> Thank you in advance.
>
> Yours Wang
>
>
>
>
> ------------------------------------
> мУгь╬╣ё╛╦╢╣╘╢Ся╖нОюМо╣яп╬©иЗ
> Wang Qinjing, Physics Department,
> Fudan University



More information about the users mailing list