My Community

Dear Team,

Trust this finds you well. I am having a convergence issue with my molecular dynamics calculation. When I perform relaxation for the same system, both electronic and ionic convergence was attained. I have a feeling it may require tweaking the ALGO parameter, however, I would like to seek insight as to what ALGO may be optimal for MD simulation before proceeding.

Dear Bolariso,

I had a quick look at your calculation. I think your electronic SCF calculation simply did not convergence as it needs more steps. Please do increase NELM to something like 200 and restart the calculation. Such slow convergence can be typical for large systems. If the system is metallic a larger smearing parameters ISMEAR / SIGMA can help. Also make sure that especially the first electronic minimization is converged before you calculate any MD step, since the MD step is calculated via the forces that are not properly converged in this case. Let me know if that helps.

In case you want me to try the calculation please do provide a KPOINTS file.

Best regards,
Alex

Dear Alex,

Thank you for your response. The calculation is performed using vasp_gam, so the KPOINTS are gamma centered. Would be glad to have your insight on the problem.

Automatic mesh
0
gamma
1 1 1
0.0 0.0 0.0

Thank you.

I tried your calculation and switching ALGO=Normal made it converge after 32 steps. Even though the DAV algo is slightly slower the overall speed of one SCF cycle is much faster like this. Please give this a try and let me know if this solves your problem.

Best,
Alex