pardens close error

Problems running VASP: crashes, internal errors, "wrong" results.


Moderators: Global Moderator, Moderator

Post Reply
Message
Author
Danny
Full Member
Full Member
Posts: 201
Joined: Thu Nov 02, 2006 4:35 pm
License Nr.: 5-532
Location: Ghent, Belgium
Contact:

pardens close error

#1 Post by Danny » Mon Feb 04, 2008 2:26 pm

Code: Select all

"pardens.f90", line 1272: 1525-093 The CLOSE statement cannot be processed because the CLOSE statement is not allowed on unit 0, which is connected to standard error.  The program will recover by ignoring the CLOSE statement.

when calculating a partial charge density I got the above error. It is shown for all but one processor(ie if the job runs on 16 CPU's I get 15 messages). Can this be ignored as beeing a coding error where all cpu's try to close the same file, and after the succes of the first one, the others don't have anything to close anymore?

thx
Danny
Last edited by Danny on Mon Feb 04, 2008 2:26 pm, edited 1 time in total.

admin
Administrator
Administrator
Posts: 2921
Joined: Tue Aug 03, 2004 8:18 am
License Nr.: 458

pardens close error

#2 Post by admin » Wed Feb 06, 2008 8:50 am

file 0 is the job.stdout file, this file number should not be addressed to any of the files which are opened and closed by pardens. (please check which file-numbers are given to the parchg files by inserting a write statement to give you iuout after the iuout = NXTFRU() line
Last edited by admin on Wed Feb 06, 2008 8:50 am, edited 1 time in total.

Danny
Full Member
Full Member
Posts: 201
Joined: Thu Nov 02, 2006 4:35 pm
License Nr.: 5-532
Location: Ghent, Belgium
Contact:

pardens close error

#3 Post by Danny » Wed Feb 06, 2008 4:36 pm

a look in the code tell's me that value zero is not really excluded

Code: Select all

! before starting the whole stuff: search for a free I/O unit:
      io_begin

       iuout = NXTFRU()
       write(*,*) 'iuout= ',iuout,' !'
       [b]IF ((iuout <= -1) .OR. (iuout >= 99) ) THEN[/b]          CALL err_handle(io%iu6,1,"PARCHG.nb.nk","NONE","U")
          CALL err_handle(io%iu0,1,"PARCHG.nb.nk","NONE","U")
          IF (io%lopen) CALL WFORCE(io%iu6)
          STOP   ! on lack of output: stop execution
       END IF

      io_end

I've put in the write statement, and am now waiting for the results

Danny
Last edited by Danny on Wed Feb 06, 2008 4:36 pm, edited 1 time in total.

Danny
Full Member
Full Member
Posts: 201
Joined: Thu Nov 02, 2006 4:35 pm
License Nr.: 5-532
Location: Ghent, Belgium
Contact:

pardens close error

#4 Post by Danny » Wed Feb 06, 2008 7:42 pm

The results of the write statement are:
iuout= 2 !
while the error persists.

So could the cause be as I suggested in my first post? The job was run on 16 cpu's and there were 15 of these errors.

Danny
Last edited by Danny on Wed Feb 06, 2008 7:42 pm, edited 1 time in total.

Danny
Full Member
Full Member
Posts: 201
Joined: Thu Nov 02, 2006 4:35 pm
License Nr.: 5-532
Location: Ghent, Belgium
Contact:

pardens close error

#5 Post by Danny » Tue Feb 12, 2008 1:22 pm

up
Last edited by Danny on Tue Feb 12, 2008 1:22 pm, edited 1 time in total.

admin
Administrator
Administrator
Posts: 2921
Joined: Tue Aug 03, 2004 8:18 am
License Nr.: 458

pardens close error

#6 Post by admin » Tue Feb 12, 2008 3:14 pm

if the parcharge file unit is 2, you can ignore the error. The reason of this error certainly is that the STOP statement is in the subroutine which writes the partial charges itself, and that the stdout file is not closed properly before vasp stops..

[ Edited Tue Feb 12 2008, 04:19PM ]
Last edited by admin on Tue Feb 12, 2008 3:14 pm, edited 1 time in total.

Danny
Full Member
Full Member
Posts: 201
Joined: Thu Nov 02, 2006 4:35 pm
License Nr.: 5-532
Location: Ghent, Belgium
Contact:

pardens close error

#7 Post by Danny » Wed Feb 13, 2008 5:05 pm

ok, thx
Last edited by Danny on Wed Feb 13, 2008 5:05 pm, edited 1 time in total.

Post Reply