ABINIT Discussion Forums

Hi.

I sept almost one week to resolve this. I am having error to running nPT ensemble simulation using ionmov=13 and optcell=2.
But whenever I applied the temperature, unitcell expand to infinity.
I cause the bound error in the end.

bound: BUG -
Trouble finding boundary of G sphere for
kpt= 0.0000 0.0000 0.0000 and ng= 36 36 48
Action : check that kpt lies reasonably within first Brillouin zone;
else code bug, contact ABINIT group.
Action : contact ABINIT group.

I can not figure out what did i wrong.
Can you give me any suggestion? comment?

Thanks alot.
Soo

here is my input:

#SCF cycle parameters

tolvrs 1.d-5 nstep 50

#K-points and sym
occopt 3 #Fermi-Drac Semaring
kptopt 3 ngkpt 10 10 10
nsym 1 #No symmetry of structure.

#Molecular Dynamics parameters
#brvltt 2
ecutsm .5
dilatmx 1.5

ionmov 13 ntime 50
optcell 2
nnos 4
qmass 10 10 10 10
bmass 10
dtion 20
mditemp 300
mdftemp 300 #Constant temperature with 300K (Nuclie temperature)
tsmear 0.001 #Smearing temperature with 300K (electronic temperature.) this value doesn't need to be same wt mdtemp

#Cell and atoms definition
acell 8.0042377090E+00 8.0042377090E+00 1.0963690034E+01
#spgroup 194
#angdeg 90 90 120
#brvltt -1

chkprim 0
rprim 1.0012488 11.511474e-21 0
-0.50062 0.867107 0
8.384548e-17 1.45224633e-16 0.99968

natom 4 ntypat 2 typat 2*1 2*2 znucl 25 83
nband 36 #18*2*2 (including SOC)
xcart -7.6591950562E-29 2.2613566162E-27 4.3495498799E-12 #M
3.3555958205E-16 5.8120624509E-16 5.4801038518E+00 #M
3.7204934698E-15 4.6270200413E+00 2.7400519259E+00 #B
4.0071168996E+00 2.3135100207E+00 8.2201557777E+00 #B

need more output from your run. I suspect the thing expanded to infinity as you said, and once the cell is large enough, the reciprocal lattice vectors are tiny, and your dilatmx is large. At some point the G vectors for a new iteration can no longer be mapped on the old one.

Follow nettiquette viewtopic.php?f=20&t=251 if you want help

Hi jh502125!

Did you find out the solution for the problem, I got the same one and I tried kptopt = 1

In case others may run into the same problem, my way out of this crash is to lower dtion to say 0.2 or 0.1. This seems too small but it will not diverge dilatmx.

Cheers!
Vahid