LDA+U

[senersen]

Dear all,

I am trying to calculate eels from a KSS file calculated with abinit. However, I need to use LDA+U in my calculations. Therefore, I generate a KSS file using LDA+U. Then at a2y stage it complains about memory. So my question is;

1. Is it possible to calculate eels with LDA+U in yambo?
2. If it is possible, what might be the cause of my problem "insufficient virtual memory".

[Daniele Varsano]

Dear senersen,
please fill your signature with your name and affiliation, this is a rule of the forum. You can do it one for all filling the signature in your profile.

1. Is it possible to calculate eels with LDA+U in yambo?

In principle yes, but you need to be aware that when evaluating the dipole matrix element the non local part of the commutator will be missing and this can affect the intensity of the eels. If this contribution it is important or not I cannot tell you a priori.

2. If it is possible, what might be the cause of my problem "insufficient virtual memory".

I do not know what version of abinit you are using. If is the last release 8.x, I'm not sure that a2y is the right way to generate yambo database and you should resort to the e2y interface after having the abinit output in etsf format.
Anyway if you have memory problem (this can happen if you have very huge systems) you can use a2y -b (number of bands). This will fragment your wavefunction databases in separate files containing the number of bands you indicated.

Best,

Daniele