Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • Dec 02 17:36

    mfherbst on gh-pages

    build based on e2cc516 (compare)

  • Dec 02 16:56

    mfherbst on v0.4.2

    (compare)

  • Dec 02 16:55
    JuliaTagBot commented #358
  • Dec 02 16:25

    mfherbst on gh-pages

    build based on e2cc516 (compare)

  • Dec 02 16:07

    mfherbst on master

    Bump version: 0.4.1 → 0.4.2 (compare)

  • Dec 02 15:50

    mfherbst on gh-pages

    build based on 50d1a0f (compare)

  • Dec 02 15:28
    mfherbst commented #549
  • Dec 02 15:28
    mfherbst closed #549
  • Dec 02 15:28

    mfherbst on wannier_plotting

    (compare)

  • Dec 02 15:28

    mfherbst on master

    Add basic routine for wannier p… (compare)

  • Dec 02 15:27
    mfherbst edited #549
  • Dec 02 13:24
    LaurentVidal95 commented #549
  • Dec 01 20:44
    mfherbst opened #549
  • Dec 01 20:43

    mfherbst on wannier_plotting

    Untested routine for wannier pl… (compare)

  • Dec 01 20:40

    mfherbst on wannier_plotting

    Untested routine for wannier pl… (compare)

  • Dec 01 18:56

    mfherbst on gh-pages

    build based on c6bc80e (compare)

  • Dec 01 18:35

    mfherbst on master

    First steps towards an interfac… (compare)

  • Dec 01 18:35
    mfherbst closed #518
  • Dec 01 17:44

    mfherbst on gh-pages

    build based on a66ee37 (compare)

  • Dec 01 17:28
    mfherbst synchronize #518
Antoine Levitt
@antoine-levitt
I think just because we wanted examples/ to be a set of examples and tutorial did not quite fit the bill
Louis Ponet
@louisponet
haha I saw some stuff getting shuffled around here and there, I got the idea though except that
I see, yea that makes sense, I guess examples are a bit more cookbooky than a full on tutorial
although I don't think people would mind to have the tutorial also appear there.
Btw, the binder didn't work out for me
Antoine Levitt
@antoine-levitt
Oh, can you file an issue?
Louis Ponet
@louisponet
sure let me try again to see what the issue was, I think Plots was missing
btw, binder runs on a server or locally?
Antoine Levitt
@antoine-levitt
On a server I think?
Louis Ponet
@louisponet
Ah so I guess I can't run QE there :P
if I may one more question: How do I get the amazing setup of your github page? I couldn't really find it specified in any of the files or so
Antoine Levitt
@antoine-levitt
What do you mean setup?
The README.md?
Louis Ponet
@louisponet
Like the pretty theme sorry
Antoine Levitt
@antoine-levitt
It's all in the Readme.md
But there's no theme?
Louis Ponet
@louisponet
Wait, how do you then change the appearance of the gh-pages?
Antoine Levitt
@antoine-levitt
I have no idea
@mfherbst?
(@mfherbst is the answer to a lot of "how do you do X" questions)
Louis Ponet
@louisponet
haha okay! :D
Michael F. Herbst
@mfherbst
Sorry for the late reply ... first: Yes you can run QE on binder! Just install it via conda.
Second: The docs setup is the default Documenter.jl theme, so just use that Julia package and you should get something similar.
I don't recall having done any custom configuring there, but it's one of these things that you do and then forget ... so I might miss something
Antoine Levitt
@antoine-levitt
-sameroom portal
Sameroom
@sameroom-bot
<Sameroom> Your Portal URL is https://sameroom.io/oAhcmZZQ -- you can send the URL to someone on a different team to share this room. Note: you can connect more than two teams this way.
<Sameroom> I've connected 1 new room #gitter-bridge (DFTK) on Slack. See map
[Antoine Levitt, DFTK] test
Antoine Levitt
@antoine-levitt
retest
Jae-Mo Lihm
@jaemolihm
Hi, I'm a fan of DFTK and just came to ask one question about the recent issue JuliaMolSim/DFTK.jl#459. The problem of large dpsi is generally present in perturbation theory, and one solution in finite-temperature DFPT is to redefine dpsi so that drho stays the same but dpsi does not diverge (e.g. Eq.(71-72) of Baroni RMP (2001)). Does this help solve the issue? (I should say that I do not know AD/direct minimization very much.)
Sameroom
@sameroom-bot
[Antoine Levitt, DFTK] Huh I didn't know anyone read what I wrote! Thanks a lot for the suggestion, I was indeed thinking it should have come up in dfpt. I'll check out the ref you mentioned, thanks!!
Sameroom
@sameroom-bot
[Antoine Levitt, DFTK] out of curiosity how did you find out about DFTK?
Jae-Mo Lihm
@jaemolihm
I'm a PhD student working on DFT, and a few months ago I began a project on developing a Julia package related to tight-binding models and Wannier functions. (The code is not yet publicly available.) One of the first things I did was studying existing packages to learn how they are designed, and I found DFTK, maybe by searching "DFT" in discourse. I learned a lot from DFTK, partly because I was familiar with the general structure of a DFT code (I've used and contributed to QE), and also because some of the utilities (like k-mesh generation and symmetry) were also needed in my code.
Michael F. Herbst
@mfherbst
Cool. Great to hear you found our code useful and instructive! That's pretty much what we hoped to achieve: To lower the entrance barrier by keeping the code focused on mathematical and physical structure. By the way: We started working on an interface of DFTK to Wannier90 some time ago, but that work got a bit deferred at the moment (PhD student for the moment focusing on a different project). I guess that could be useful for you if we got that done at some point ...
umbriquse
@umbriquse
How appropriate is DFTK in simulating fluid interactions? Understandably water is pretty complicated, but what about liquid metals?
Sameroom
@sameroom-bot
[Antoine Levitt, DFTK] I would imagine for fluids you need pretty big MD simulations, for which classical MD is probably a better fit
[Antoine Levitt, DFTK] but I don't know
Michael F. Herbst
@mfherbst
I'd say it depends a bit on the size of the model and your material of interest. For molten sodium for example you
might have a chance, because you can get it down to like 1 electron per atom and a simple lda model
but I would say classical md / perhaps with a force field learned from DFT gets you closer in terms of cost
Sameroom
@sameroom-bot
<Sameroom> Whoops! You've exceeded your daily message limit on this Sameroom account (it will reset in 24 hours and 0 minutes). If you have too many Tubes for your budget, pause or delete some Tubes on the Manage page. If you dont have a subscription, visit pricing to upgrade to unlimited messaging.
Michael F. Herbst
@mfherbst
as always there is no black and white answer and I suggest you approach the problem step by step (ab inito md on a small case to get a feel for cost, classical md with a simple potential to get a feel for large-scale behaviour and then try to find a good middle ground)
Sameroom
@sameroom-bot
<Sameroom> Whoops! You've exceeded your daily message limit on this Sameroom account (it will reset in 23 hours and 57 minutes). If you have too many Tubes for your budget, pause or delete some Tubes on the Manage page. If you dont have a subscription, visit pricing to upgrade to unlimited messaging.
umbriquse
@umbriquse
Thank you for your answers! More specifically I was thinking more toward the accuracy of the simulation (e.g. the exchange potential).
Michael F. Herbst
@mfherbst
That depends quite a lot on the metal. If I remember correctly for sodium even LDA is quite good and I recall having seen studies with molten sodium clusters done with LDA. This is not my field of expertise, so I don't have a good overview, but I would imagine you can identify a decent functional to give you a reasonable compromise. One should add that it is known that common DFT functionals start giving larger errors at high temperature and that they generally fail to capture long-range effects. But for molten metals that's not too important usually.
Sameroom
@sameroom-bot
<Sameroom> Whoops! You've exceeded your daily message limit on this Sameroom account (it will reset in 13 hours and 31 minutes). If you have too many Tubes for your budget, pause or delete some Tubes on the Manage page. If you dont have a subscription, visit pricing to upgrade to unlimited messaging.
Michael F. Herbst
@mfherbst
Note to everyone: The gitter chat is discontinued. For future questions please ask in the #dftk channel at https://join.slack.com/t/juliamolsim/shared_invite/zt-tc060co0-HgiKApazzsQzBHDlQ58A7g
Sameroom
@sameroom-bot
<Sameroom> Whoops! You've exceeded your daily message limit on this Sameroom account (it will reset in 5 hours and 17 minutes). If you have too many Tubes for your budget, pause or delete some Tubes on the Manage page. If you dont have a subscription, visit pricing to upgrade to unlimited messaging.
Antoine Levitt
@antoine-levitt
yup, feel free to join the JuliaMolSim slack, it's full of nice people from the quickly-growing julia molecular simulation community!
Sameroom
@sameroom-bot
<Sameroom> Whoops! You've exceeded your daily message limit on this Sameroom account (it will reset in 4 hours and 43 minutes). If you have too many Tubes for your budget, pause or delete some Tubes on the Manage page. If you dont have a subscription, visit pricing to upgrade to unlimited messaging.