search

OrderN / CONQUEST-release

Full public release of large scale and linear scaling DFT code CONQUEST
http://www.order-n.org/
MIT License
96 stars 25 forks source link

Missing potential and input file in pseudo-and-pao for O #66

Closed lionelalexandre closed 3 years ago

lionelalexandre commented 3 years ago

Oxygen atom potential and MakeIonFiles input are missing in the LDA and PBE directory. Maybe Dave have them somewhere.

davidbowler commented 3 years ago

Thank you for spotting this! The PBE and PBEsol files seem to be there, though as you say the LDA are missing.

smujahed commented 3 years ago

If I can jump on this bandwagon, there's no LDA for Ge.

davidbowler commented 3 years ago

Systematic check: compared to PBE, LDA is missing

Co, Cs, Ge, La, Lu, Mn, Ni, O

Any others I've missed?

davidbowler commented 3 years ago

It turns out that the missing elements have no entries in PseudoDojo for LDA with the standard accuracy. They do seem to exist with stringent accuracy.

@lionelalexandre @smujahed please check that I've not missed any more; I'll then generate the files and issue (another) hot fix.

lionelalexandre commented 3 years ago

Considering the 6 firsts rows, the At atom is also missing for all the functional

smujahed commented 3 years ago

It turns out that the missing elements have no entries in PseudoDojo for LDA with the standard accuracy. They do seem to exist with stringent accuracy.

@lionelalexandre @smujahed please check that I've not missed any more; I'll then generate the files and issue (another) hot fix.

@davidbowler I've just had a look. La and Lu definitely don't have LDA for standard accuracy. Co, Cs, Ge, Mn, Ni and O do seem to have standard accuracy potentials, but they don't have values for the tests (delta, delta1 and gbrv). I may be looking at a different version of ONCVPSP though.

Just for reference, I am looking at them on www.pseudo-dojo.org/index.html with options:

davidbowler commented 3 years ago

Considering the 6 firsts rows, the At atom is also missing for all the functional

Yes - it seems to be unavailable (though I'm not sure how much interest there is in it!)

davidbowler commented 3 years ago

Thanks, @lionelalexandre I hadn't seen that v0.3 was available - v0.4 is not (and there's no clear explanation).

smujahed commented 3 years ago

I think I've made a mistake, my apologies!

Inspecting the headers of the NC SR v0.4 LDA UPF files, files are marked as generated with v3.3.0 of ONCVPSP (at least for Si and As). Seems v0.3 UPFs are marked as using v3.2.3 of ONCVPSP... I'm confused by the version convention on the PseudoDojo website. Version numbering on PseudoDojo seems to be the version of the database rather than the version of ONCVPSP, my mistake!

davidbowler commented 3 years ago

It looks to me from what I can see that there are no potentials for La or Lu with LDA at either the standard or stringent levels. I don't understand why this is (they are present for PBE and PBEsol).

davidbowler commented 3 years ago

Fixed by #69