ABINIT Discussion Forums

Dear all,

I tried several times by change of dilatmx but always it displays this error message:


Difference of energy with previous step (new-old):
Absolute (Ha)=-4.72926E-01
Relative =-9.83569E-04
fconv : at Broyd/MD step 8, gradients have not converged yet.
max grad (force/stress) = 1.8840E-01 > tolmxf= 5.0000E-04 ha/bohr (free atoms)

Geometry Optimization Precondition: 0

--- !WARNING
src_file: mover.F90
src_line: 834
message: |
The new primitive vectors rprimd (an evolving quantity)
are too large with respect to the old rprimd and the accompanying dilatmx: 1.043683E+00
This large change of unit cell parameters is not allowed by the present value of dilatmx.
An adequate value would have been dilatmx_new= 1.045934E+00
Calculation continues with limited jump, by rescaling the projected move by the factor 8.558994E-01.

Best regards
mustafa

Dear Mustafa,
You can resubmit your job by increasing our dilatmx (1.06 or 1.08 or 1.10), this means that your cell relaxation is quite large. If you need dilatmx way beyond 1.1, this probably means that your initial cell is far from the ground state one, though other reasons can be at the source of the problem as bad convergence parameters (too low ecut or ngkpt) or bad compilation flags. First try to increase a bit your dilatmx.
Best wishes,
Eric