Read Error when exporting wavefunctions
Mirjam Van Vroonhoven
mirjamv at theochem.kun.nl
Wed Jan 24 10:52:08 GMT 2001
I try to export Oxygen MRCI wavefunctions with the EXPORT card. For some
states, it works fine, but for other states I get the following error message:
EXPORT record 6000.2 to file
/scratch/mirjamv/Final/test_wavefunction_save/wavefunction.dat
rdabsf: Error in reading from file T1200004124.TMP (unit 12),
4096 words at word offset 38912
rdabsf: Error in reading from file T1200004124.TMP (unit 12),
4096 words at word offset 38912
(where I added two newlines to increase readability, the numbers are variable,
but the message is the same all the time)
It works fine for (for example) the b Singlet Sigma_g^+ of O_2, but it crashes
for the ground state X Triplet Sigma_g^-.
I used the following molpro input:
------- For X Triplet Sigma gerade minus ------
#!/bin/csh -f -e
time nice +19 molpro << EOF
***, O2 CI crashing of export
gprint,basis,orbitals
basis,VDZ
geometry={O1;O2,O1,R}
R = 2.2
int
rhf
wf,16,4,2
ci
state,1
save,6000.2
export,6000.2,wavefunction.dat
---
EOF
-----------------------------------------------
The input for the b Singlet Sigma gerade plus state was exactly the same,
except for the WF card in the Hartree - Fock, which read "wf,16,1,0". This job
worked fine.
The error seems to be generated in the routine export
($molpro/src/util/export.f), during the call to the subroutine lesw
($molpro/src/util/sreibw.f) which reads the requested record from temporary
file number 2 (unit 12). The routine lesw calls the routine rdabsf which is a
lowlevel IO routine. This one generates the error.
Can someone tell me what is wrong here? It is not a problem of the disk being
full or something. I run molpro on a SUN Ultra-Enterprise machine, with
Solaris 2.7.
Mirjam van Vroonhoven
PhD student in Theoretical Chemistry.
--
E-mail: mirjamv at theochem.kun.nl
Homepage: http://www.theochem.kun.nl/~mirjamv/
Phone: +31-24-365 30 33
Visiting Address: Department of Theoretical Chemistry
Toernooiveld 1, Room number N2035
6525 ED Nijmegen
More information about the Molpro-user
mailing list