Truble with FERWE tag

[Jianguo Si]

Dear VASP Team,

Recently, I want to use VASP to preform some non-equilibrium physics calculations with FERWE tag. But unfortunately, I met some troubles.

I have done two tests as follwoings,

case-1: If I set the FERWE as

Code: Select all

FERWE = 48*0.9999 176*0 48*0.9999 176*0  48*0.9999 176*0 ...48*0.9999 176*0 48*0.9999 176*0  48*0.9999 176*0

for 224 bands and 64 K-points(set by explicit k-point mesh mode artificially), the code runs well.

case-2: If I set the FERWE as

Code: Select all

FERWE = 46*1 0.5116 0.5116 0.5116 0.0013 0.0013 0.0011 0.0011 171*0 46*1 0.8762 0.4285 0.2732 0.0526 0.0526 0.0001 172*0 47*1 0.9892 0.23330.0004 0.0004 0.0003 172*0 47*1 0.9999 0.071 0.0014 174*0 46*1 0.9973 0.9973 0.0141 0.0141 174*0 45*1 0.997 1 0.9983 0.0739 0.0029 0 0 0 0 0 0 0.0001 167*0 44*1 0.9979 1 1 0.9859 0.2352 0.0004 0.0004 0.001 0 0.0026 0 0 0.0001 167*0 42*1 0.9986 0.9986 0.9999 0.9988 0.876 0.4297 0.2732 0.0539 0.054 0.0003 0.0001 0 0.0001 0.0001 168*0 41*1 0.9999 0.9989 0.9952 0.9999 0.9996 0.8761 0.4288 0.2732 0.0573 0.0538 0.0003 0 0 0 0.0001 0.0001 167*0 43*1 0.9981 0.9946 0.9999	0.9996 0.9841 0.0366 0.0331 0.0006 0.0007 0.0002 0.0019 0.0001 169*0 45*1 0.9939 0.9998 0.998 0.0192 0.0003 0.0021 0 0 0 0 0.0002 168*0 44*1 0.9965 0.9974 0.9968 0.9983 0.0099 0.0056 174*0 44*1 0.9965 0.9974	0.9968 0.9983 0.0099 0.0056 174*0 45*1 0.9939 0.9998 0.998 0.0192 0.0003 0.0021 0 0 0 0 0.0002 168*0 43*1 0.9981 0.9946 0.9999 0.9996 0.9841 0.0366 0.0331 0.0006 0.0007 0.0002 0.0019 170*0 41*1 0.9999 0.9989 0.9952 0.9999 0.9996 0.8761 0.4288 0.2732 0.0574 0.0537 0.0003 0 0 0 0.0001 0.0001 167*0 43*1 0.9998 0.9994 1 0.9995 0.9882 0.2339 0.0004 0.0005 0.0009 0.0002 0.0007 0 0 0.0001 167*0 44*1 0.9999 0.9974 0.9999 0.9989 0.0157 0.0003 0.0011 0 0 0 0 0.0001 168*0 46*1 0.9997 0.9995 0.0002 0.0002 0.0008 173*0 43*1 0.9999 0.9999 1 0.9999 0.9999 0.0002 0.0002 174*0 46*1 0.9997 0.9995 0.0002 0.0002 0.0008 173*0 44*1 0.9999 0.9974 0.9999 0.9989 0.0157 0.0003 0.0011 0 0 0 0 0.0001 168*0 43*1 0.9998 0.9994 1 0.9995 0.9882 0.2339 0.0004 0.0005 0.0009 0.0002 0.0007 0 0 0.0001 167*0 43*1 0.999 0.999 1 0.9994 0.9862 0.0322 0.0322 0.0006 0.0005 0.0001 0.0002 170*0 44*1 0.9995 0.9992 0.9999 0.9994 0.0722 0.0018 0 0 0.0001 0 0 0.0001 168*0 44*1 0.9993 0.998 0.9987 0.9998 0.0065 0.0023 174*0 45*1 0.9999 0.9997 0.9999 0.0004 0.0002 0.0001 0 0 0.0001 170*0 45*1 0.9999 0.9997 0.9999 0.0004 0.0002 0.0001 0 0 0.0001 170*0 44*1 0.9993 0.998 0.9987 0.9998 0.0065 0.0023 174*0 44*1 0.9995 0.9992 0.9999 0.9994 0.0722 0.0018 0 0 0.0001 0 0 0.0001 168*0 44*1 0.9998 0.9992 0.9999 0.9997 0.0141 0.0003 0.0004 0 0.0001 171*0 .....  44*1 0.9998 0.9992 0.9999 0.9997 0.0141 0.0003 0.0004 0 0.0001 171*0 44*1 0.9997 0.9997 0.9993 0.9993 0.0123 0.0123 174*0 44*1

for same bands and Kpoints, the code aborted with "Error reading item FERWE from file INCAR. Error code was IERR= 5 ... ." I have checked the occupations carefully and confirmed that the number of occupations is right for 224 bands and 64 K-points. Why such setting is bad for VASP codes? Is there a maximum value for the input string? I want to fix this problem by revising the source codes, but I don't know which parameters is for this. Please kindly give me some suggestions. Thanks for all disscutions and reply!

Best & Regards!
Jianguo Si

[Jianguo Si]

Sorry for the wrong spelling of "trouble" in the title.

[Jianguo Si]

Dear all,

I have upload the input, output and the log files

Attachment.zip

. Please kindly give me some suggestions according to the files.

Regards!
Jianguo Si

[xin_zhao]

Well, for testing purposes, I used your POSCAR and POTCAR files but regenerated a KPOINTS file with only one k-point. I also edited the FERWE tag in your INCAR file, setting it as follows:

Code: Select all

FERWE = 46*1 0.5116 0.5116 0.5116 0.0013 0.0013 0.0011 0.0011 171*0

As expected, the VASP run failed. Then, I commented out the NBAND tag in the INCAR file, then this simple test ran successfully.

However, when I used your original KPOINTS and INCAR files with only the NBAND tag commented out, the same error persisted. I suspect the issue might be related to either the FERWE tag or the KPOINTS file.

Besides, in my opinion, this job setup is fundamentally flawed. I do not understand what you are trying to achieve. The orbital occupation settings do not even conserve the number of electrons in your system! Under these conditions, you aim to perform a phonon calculation. Even if this job were to run successfully, the results would be physically meaningless. Does this calculation setup have any theoretical basis or justification?

[Jianguo Si]

Well, for testing purposes, I used your POSCAR and POTCAR files but regenerated a KPOINTS file with only one k-point. I also edited the FERWE tag in your INCAR file, setting it as follows:

Code: Select all

FERWE = 46*1 0.5116 0.5116 0.5116 0.0013 0.0013 0.0011 0.0011 171*0

As expected, the VASP run failed. Then, I commented out the NBAND tag in the INCAR file, then this simple test ran successfully.

However, when I used your original KPOINTS and INCAR files with only the NBAND tag commented out, the same error persisted. I suspect the issue might be related to either the FERWE tag or the KPOINTS file.

Besides, in my opinion, this job setup is fundamentally flawed. I do not understand what you are trying to achieve. The orbital occupation settings do not even conserve the number of electrons in your system! Under these conditions, you aim to perform a phonon calculation. Even if this job were to run successfully, the results would be physically meaningless. Does this calculation setup have any theoretical basis or justification?

Thanks for your reply. The settings I used are for some simulations of momentum-dependent electronic excitation.