BRMIX: very serious problems the old and the new charge density differ

[li_weixue70]

Hello:
I am running VASP 6.3.0 with vtst and vaspsol++ compiled with the Intel one API Base and HPC toolkits on 2*AMD 7702 128C-256G system. My calculation system is at constant potential with CO adsorbed on the surface of Cu111, but I ran into this problem before calculating with solvation and constant potential, so it shouldn't have anything to do with those two components

• POTCAR contains this elements PAW_PBE Cu 05Jan2001 PAW_PBE C 08Apr2002 and PAW_PBE O 08Apr2002

• KPOINTS was set to 2 2 1 corresponding to 15A Cu cell

• Some of the key INCAR parameters are shown below

Code: Select all

 NCORE = 8 
 ENCUT = 400

Code: Select all

 ALGO = Fast 

Code: Select all

 EDIFF = 1E-4
 EDIFFG =-0.05 

and I get this error after 5 DAV calculation

Code: Select all

 DAV:   5    -0.898271934088E+05   -0.86583E+05   -0.13688E+03 24800   0.300E+01BRMIX: very serious problems
 the old and the new charge density differ
 old charge density:  1613.00000 new 1635.47232
    0.130E+02

Afterwards, the calculation was terminated because the charge was significantly changed and the corresponding solvation and constant potential parts were erroneous.

Regards,
xinze qi