how to restart epw in the 5.4.1 version

General discussion around the EPW software

Moderator: stiwari

Post Reply
YSW
Posts: 23
Joined: Fri Jun 18, 2021 5:24 am
Affiliation: Shanghai University of Electric Pow

how to restart epw in the 5.4.1 version

Post by YSW »

Hi dear users and experts

I use the 5.4.1 epw version to calculate superconducting gap, and i find that the kmaps parameter is removed

1. If I want to resart epw, what parameter i need to set in epw-5.4.1
2. When should I use the input flag of "lacon=.true." ?

Best,
YSW

hlee
Posts: 415
Joined: Thu Aug 03, 2017 12:24 pm
Affiliation: The University of Texas at Austin

Re: how to restart epw in the 5.4.1 version

Post by hlee »

Dear YSW:
I use the 5.4.1 epw version to calculate superconducting gap, and i find that the kmaps parameter is removed

1. If I want to resart epw, what parameter i need to set in epw-5.4.1
Starting from EPW v.5.4.1, you don't need to care about kmaps. You can restart the calculations in the same way as in the previous versions except for not using kmaps.
2. When should I use the input flag of "lacon=.true." ?
With lacon=.true., you can perform the analytical continuation in the way described in the paper at https://journals.aps.org/prb/abstract/1 ... vB.37.4965 .
However, I found that there is (are) some issue(s) with the current implementation and I would suggest you not to use this feature.

Sincerely,

H. Lee

YSW
Posts: 23
Joined: Fri Jun 18, 2021 5:24 am
Affiliation: Shanghai University of Electric Pow

Re: how to restart epw in the 5.4.1 version

Post by YSW »

Dear Dr. H. Lee

Thanks for your reply, it's worked. But I have a another problem now.

I am getting this error:
Error in routine dynmat_asr (11):
inconsistent data (names)

I sure that my prefix.dvscf_q* and prefix.dyn_q* are right. And q2r.x can also be run successful.
I don't konw what cause this error?

best regards
YSW

YSW
Posts: 23
Joined: Fri Jun 18, 2021 5:24 am
Affiliation: Shanghai University of Electric Pow

Re: how to restart epw in the 5.4.1 version

Post by YSW »

Dear Dr. H. Lee

I'm sorry to bother you again, but I find the cause of this error.

The atomic order in ATOMIC_POSITIONS (crystal) in scf.in, which should be run before epw, should be same as ph.x.

best,
YSW

hpaudya1
Posts: 190
Joined: Tue Mar 21, 2017 7:11 pm
Affiliation:

Re: how to restart epw in the 5.4.1 version

Post by hpaudya1 »

Hi YSW,

It looks like your 'scf.in' (and 'nscf.in') used in the epw calculation is not consistent with the one that you used for the phonon calculation. The error you mentioned came from here;
https://gitlab.com/QEF/q-e/-/blob/devel ... r.f90#L361

This indicates that either number of atoms or atomic types are different.

Best,
Hari

YSW
Posts: 23
Joined: Fri Jun 18, 2021 5:24 am
Affiliation: Shanghai University of Electric Pow

Re: how to restart epw in the 5.4.1 version

Post by YSW »

Dear Dr. Hari

In my case,this error is caused by different atomic order of ATOMIC_POSITIONS (crystal) between scf.in (nscf.in) and ph.in.

e.g:
1.
ATOMIC_POSITIONS (crystal)
S 0.0000000000 0.0000000000 0.0000000000
H 0.5000000000 0.5000000000 0.0000000000
H 0.5000000000 0.0000000000 0.5000000000
H 0.0000000000 0.5000000000 0.5000000000
2.
ATOMIC_POSITIONS (crystal)
H 0.5000000000 0.5000000000 0.0000000000
S 0.0000000000 0.0000000000 0.0000000000
H 0.5000000000 0.0000000000 0.5000000000
H 0.0000000000 0.5000000000 0.5000000000

If (1) is used in scf.in(nscf.in), and (2) is used in ph.in.

I get the error:
Error in routine dynmat_asr (11):
inconsistent data (names)

Best,
YSW

Post Reply