Re: [QE-users] Band unfolding in QE v6.6

I don't know Bandup, but if it is well structured it should be quite easy to adapt it to the new format of data files. Otherwise, you may try "unfold" by Pietro Bonfà: https://bitbucket.org/bonfus/unfold-x/src/master/ Paolo On Thu, Oct 8, 2020 at 4:15 AM William Hewett <

[QE-users] gather_grid function

Hi users, The gather_grid function is a 1D array with dimensions nr1x * nr2x * nr3x. Does it follow the convention written in the developer's manual -- written as a 3D array with dimensions nr1x, nr2x, nr3x, is the innermost loop x, then y, and the outermost loop z? Best regards, Andrew

Re: [QE-users] Is wfcdir still functional?

Dear Dr. Giannozzi, For QE input to be portable, not only wfcdir but also max_seconds should be read from environment variables. Time limit differs with batch queues. Best regards, ---Sender--- Takahiro Chiba Undergraduate at Hokkaido University takahiro_ch...@eis.hokudai.ac.jp -

[QE-users] Band unfolding in QE v6.6

Hi QE Users, I'm wanting to unfold a band structure calculation on a supercell, I've tried Bandup (https://github.com/band-unfolding/bandup) however it seems the current version of Bandup is not compatible with versions of QE post 6.3. My calculation requires QE 6.6. Does anyone know if it is

Re: [QE-users] Naming convention of PSlibrary pseudopotentials

l is used in pslibrary to distinguish the PP with larger core radii. It means lower accuracy. Andrea On Wed, 2020-10-07 at 13:45 +, 박민규 wrote: > Dear all, > > I have a small question. > > For example Ga in PSlibrary table, there are two pseudopotentials > that share almost identical

Re: [QE-users] Is wfcdir still functional?

Dear Mr. Marchand, Dr. Giannozzi, and all QE users, No "pwscf.wfcN" was found in "wfcdir" because he did calculation='scf' and disk_io='low', I suspect. When I do "relax" calculation with disk_io='low' option, QE 6.5 with intel MPI 18 on CentOS 6 does write "pwscf.wfcN" in "wfcdir", where N is

[QE-users] Naming convention of PSlibrary pseudopotentials

Dear all, I have a small question. For example Ga in PSlibrary table, there are two pseudopotentials that share almost identical name: Ga.rel-pbe-dn-kjpaw_psl.1.0.0.UPF Ga.rel-pbe-dnl-kjpaw_psl.1.0.0.UPF I understand the meaning of “d" and “n” based on the explanation of QE website.

Re: [QE-users] Is wfcdir still functional?

I think wfcdir is still used when wavefunctions are saved to disk during the scf calculation (option disk_io='high', that shouldn't be normally used) but not for writing the final data. Is there a good reason to use wfcdir? My opinion is that it should be deleted, together with outdir and

Re: [QE-users] Huge difference between Wall time and CPU time in electron-phonon calculation

If your cluster has a disk per node with a local (i.e., directly accessible by the node but not visible from other nodes) you may try to write data to that disk. It's a pain in the *** and may require some tweaking, but it might work. You need to ensure that the directories are there (they are

Re: [QE-users] Underestimated Adsorption Energy with optB88-vdW

Hello Thomas, 1/2 They haven't specified any parameters to my knowledge that would lead to such a large error. There may be slight difference in the thresholds of energetic convergence, but they should yield errors on a much smaller scale. 2/2 RE: box sizes - We did briefly fall foul of

Re: [QE-users] Huge difference between Wall time and CPU time in electron-phonon calculation

Dear Lorenzo and Michel, thank you for your reply. Yes, unfortunately I am using a small cluster and, for memory issues, I cannot use the SSD-based local scratch. Thus I am forced to use a non-parallel NFS. From your answers, I understand that, given the situation, the huge difference between

Re: [QE-users] Underestimated Adsorption Energy with optB88-vdW

Dear Gabriel, 1/2, are your sims. reproducing *all* of the ref. params? (eg. are cell parameters the same?) any missing params. would contribute by adding an *unknown* magnitude of error... 2/2, my guess would be that your evaluation of dE_a is wrong due to *different cell sizes*

[QE-users] Underestimated Adsorption Energy with optB88-vdW

Dear Users, Problem: I am performing calculations with the optB88-vdW functional to obtain adsorption energies of small aromatic molecules on metallic surface facets (Pt(111) and phenol for the cases discussed here). Unfortunately, the adsorption energies obtained from these calculations are

Re: [QE-users] Problem for testing k convergence

Dear Prof. Pietro and users, Yes, you have correctly pointed out and I have understood now. I want to check the variation with higher kmesh (8,9,10) and have found that the program is killed for 8*8*8 kmesh itself. The required static and dynamical RAM are 8.59 and 13.64 GB respectively (as

Re: [QE-users] Computing inner products involving orbitals

Thank you Matteo and Lorenzo! h_psi.f90 looks promising, and I will take a look at pw2wannier.f90 for seeing how to apply H to an orbital. Do you or others have any pointers on the inner product, either in reciprocal space or real space? ___ Quantum

Re: [QE-users] Computing inner products involving orbitals

You can find an example of how H|psi> is done (and the subroutines you have to call before to be sure that everything is in place) in PP/src/pw2wannier90.f90 I would recommend you compute just to begin with, and verify that it is \epsilon_i \delta_ij On 07/10/2020 09:15, Matteo Cococcioni

Re: [QE-users] Computing inner products involving orbitals

Dear Andrew, I would first compute | \psi'_l > = U_lm | \psi_m> using some linear algebra routine (as done in the code in many places by summing over plane waves), then compute H |\psi'_l > (e.g., using h_psi.f90), finally the product with < \psi'_n |. HTH Matteo Il giorno mer 7 ott 2020 alle