diolib.f RECL 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:

diolib.f RECL error

#1 Post by Danny » Mon Dec 17, 2007 3:41 pm

hello,

I've been running VASP on huygens for some time now, without to much problems, but when I wanted to do a partial charge density calculation I got this error:

Code: Select all

"diolib.f", line 177: 1525-183 The maximum record length of 9223372036854775806 for the unit 70 is out of the range of the scalar variable specified with the RECL= specifier in the INQUIRE statement.  A larger type kind should be used.  The program will stop.


everything in the calculation works fine, the WAVECAR is read correctly, KPOINTS are choosen,

Code: Select all

Start basisrun: Monday 17 December 2007, at 17:16:35
Einde basisrun: Monday 17 December 2007, at 17:16:35
Start STMrun: Monday 17 December 2007, at 17:16:35
 running on    1 nodes
 distr:  one band on    1 nodes,    1 groups
 vasp.4.6.31 08Feb07 complex
 POSCAR found :  3 types and  100 ions
 LDA part: xc-table for Ceperly-Alder, standard interpolation
 found WAVECAR, reading the header
 POSCAR, INCAR and KPOINTS ok, starting setup
 WARNING: wrap around errors must be expected
 FFT: planning ... 1
 reading WAVECAR
 the WAVECAR file was read sucessfully
Calculating partial charge density.
Open file   1 INCAR for input.
NDMOD =  -3
Energy range vs. Fermi energy from INCAR:    1.500   0.000
Efermi =   -0.236
No KPOINTS given: use all of them.

*****************************************************
* Parameters from pardens
*****************************************************

NBMOD is set to  -3
Selected energy range (EINT)   -0.2362   1.2638
Selected all k-points to calculate charge density.



but the moment of the actual calculation(when kpoint1:.... should appear) the job dies...
It seems to be a problem with the CHG file(I copied that from a static calculation) and something about an integer...but I thought it was defined as beeing 10byte(the number you see is only 8byte)???

What is the problem and how could this be solved?

thx
Danny
<span class='smallblacktext'>[ Edited ]</span>
Last edited by Danny on Mon Dec 17, 2007 3:41 pm, edited 1 time in total.

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

diolib.f RECL error

#2 Post by admin » Wed Dec 19, 2007 2:52 pm

please check whether that line in diolib.f reads
INQUIRE(UNIT=IUNIT,RECL=LRECL)
this behaved buggy on some machines (the local integer LRECL gave unreasonably high numbers) and has been fixed (2003) to
IF (ACC(1:6).EQ.'DIRECT')INQUIRE(UNIT=IUNIT,RECL=LRECL)
Last edited by admin on Wed Dec 19, 2007 2:52 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:

diolib.f RECL error

#3 Post by Danny » Wed Dec 19, 2007 3:24 pm

In the mean time I solved the problem by defining LRECL as an INTEGER*8 instead of just INTEGER...Is this an equivalent solution?

I'll now implement this solution.

thx
Danny
Last edited by Danny on Wed Dec 19, 2007 3:24 pm, edited 1 time in total.

Post Reply