Error code while while running Coffee.

[Nicholas Winner]

Hello,

I am trying to use CoFFEE for correcting 2d monolayer defect energies.

I was able to run the supercell scaling part of th MoS2 example, and I have attempted to replicate it for my system. When I do this, I am able to get results for alpha=4,5,6,8 but when I get to 10, it fails.

The error is raised by the bicgstab method in the PS_main.py file, line 207 and 217. I retrieveed the error codes from bicgstab and found that both positive and neegative error codes are returned during my code execution. Positive means that the iterative solver did not converge, negative means "illegal input or breakdown"

I manually increased the max_iter to 20,000 and removed the issue of non-convergence, but illegal input still gets raised.

Below is my input file for alpha=10. It is the same as the alpha=8 input file (other than different cell size and slab center of course. If you could please take a look at my input and see if anything looks wrong, it would be much appreciated.

Best,

Nick

&CELL_PARAMETERS

Lattice_Vectors(normalized):

0.4999863 -0.86603331  0

0.4999863 0.86603331 0

0 0 1

Cell_dimensions bohr

67.3962 67.3962 67.3962

Ecut=15.0 Hartree

/

&DIELECTRIC_PARAMETERS Slab

Epsilon1_a1 = 39

Epsilon2_a1 = 1.0

Epsilon1_a2 = 39

Epsilon2_a2 = 1.0

Epsilon1_a3 = 3.4

Epsilon2_a3 = 1.0

Width = 6.85 bohr

Centre = 27.8875 bohr

Smoothness = 0.37807 bohr

Plot_eps = True

/

&GAUSSIAN_PARAMETERS:

Total_charge = -1

Sigma = 1.89035

Centre_a1 = 0.5

Centre_a2 = 0.5

Centre_a3 = 0.4773

/

~  

[Mit Naik]

Hi Nick,

Are you using the GitHub version of the code? It has some bug fixes related to this error. https://github.com/qtm-iisc/CoFFEE

Best,

Mit