Restarting with constant basis set

[tomic]

I am running some calculations with external optimiser I wrote and I need to keep the basis set constant during the relaxation run. So what I'm trying to do right now is to perform one electronic run, to get the WAVECAR file, and then set ISTART=2 for subsequent electronic runs. When I do that, I will typically be able to do one step using the basis from WAVECAR file, calculate structural change, and then the following step will fail with:

Code: Select all

ERROR: while reading WAVECAR, plane wave coefficients changed

Mind here, that I am using exactly same settings as in the run that produced WAVECAR file, and also the symmetry is off ISYM=0. Here's the INCAR file I've used for test runs on the diamond

Code: Select all

 ENCUT = 360.000000
 PREC = Normal
 ALGO = Very_Fast
 ISYM = 0
 NSIM = 4
 ISTART = 2
 NPAR = 4
 LCHARG = .FALSE.
 LWAVE = .FALSE.

Is there any specific detail about constant basis set that's not mentioned in the manual, because I cannot find anything that might be wrong.

[tomic]

I would also like to add that I have tried to manually specify FFT mesh (NGX, NGY, NGZ, NGXF, NGYF, NGZF) to ensure it remains unchanged but it made no difference. I still have the error.