The ABINIT Users Mailing List ( CLOSED )

[Lingzhu Kong <kongl@physics.rutgers.edu>]

Dear all,

I am restarting a calculation from wave function for a large system and got 
the following
errors"
hdr_check:  Wavefunction file is OK for direct restart of calculation
 wfsinp: inside loop, init ikpt0,isppol0=   1   1
 rwwf: ERROR -
  Reading option of rwwf. Trying to read
  a RF wf record of a wf file, unit=   1
  gave iostat=  39. Your file is likely not correct.
  Action: check your input wf file.  "

Then I used cut3d to analyze the wf file and  found
"npwarr= 2184721847218472185021847218502185021847"
while there should be space between numbers like " 21847 21847 ...."

If I somehow decrease npwarr such as using a small ecut, then the
restarted job runs OK. Clearly something overflows.

I looked into the code and saw in the hdr_io.F90 file
"write(unitfi, '(a,(10i5,8x))') ' npwarr=',hdr%npwarr(:)"
It seems that i5 is not long enough, which causes the above output of 
"npwarr=" to
be continuous numbers without space in between. but it also seems that this 
writing 
format should only affect the formatted file writing, not the wave function 
file, right?
I got stuck.

Any suggestions how to fix it? Thanks!

Lingzhu