[DFTB-Plus-User] DFTB-Plus-User Digest, Vol 27, Issue 7
Ben Hourahine
benjamin.hourahine at strath.ac.uk
Thu Apr 1 17:33:31 CEST 2010
Hello Dave,
Did you manage to find what was causing the problem? I've just tried the
input, and didn't get any convergence problems. If the code passes
make test
then another possibility might be that your SK files are damaged in some
way, the md5 sums for the pbc-0-2 files from the dftb.org page are :
c9420dc8afd32cc11cc47a541d2725cf pbc-0-2/C-C.skf
c00dd573ec2053927b551e2bb80aef79 pbc-0-2/C-H.skf
9eaca769b1317b77f0b5308fae8296bc pbc-0-2/H-C.skf
9a7fbda84f324126067d76b07f3d8a27 pbc-0-2/H-H.skf
Cheers
Ben
Dave Schall wrote:
> Okay, I am running "dftb+_s081217.i686-linux", I turned the scc back on,
> and I still don't get convergence. (and the system looks like it
> exploded). Would that hint at a bad initial coordinate input? How
> quickly did your's converge? Why would they be different?
>
> Thanks!
>
> Here is the end result of 100 cg steps and 100 scc steps:
>
> 95 -0.72192733E+02 0.21030738E-02 0.16794511E-01
> 96 -0.72194720E+02 -0.19864693E-02 0.13569232E-01
> 97 -0.72195137E+02 -0.41683246E-03 0.14380645E-01
> 98 -0.72192421E+02 0.27155471E-02 0.25102920E-01
> 99 -0.72192982E+02 -0.56062346E-03 0.16674836E-01
> 100 -0.72191969E+02 0.10125796E-02 0.17576861E-01
>
> Total Energy: -70.920240
> Total Mermin free energy: -70.920240
> WARNING!
> -> SCC is NOT converged, maximal iterations exceeded
> Maximal force component: 0.051841
>>> Charges saved for restart in charges.bin
> WARNING!
> -> !!! Geometry did NOT converge!
> !!! Geometry did NOT converge!
>
>
> Here's the input:
>
> #3x3 unit cell: Does not converge!
> Geometry = GenFormat {
> 72 S
> C H
> 1 1 -3.7979 -0.2582 -6.5783
> 2 1 -2.5319 0.2584 -5.8473
> 3 1 -2.5319 -0.2582 -4.3855
> 4 1 -3.7979 0.2584 -3.6545
> 5 2 -3.7979 -1.3798 -6.5783
> 6 2 -2.5319 1.3800 -5.8473
> 7 2 -2.5319 -1.3798 -4.3855
> 8 2 -3.7979 1.3800 -3.6545
> 9 1 -1.2659 -0.2582 -6.5783
> 10 1 0.0001 0.2584 -5.8473
> 11 1 0.0001 -0.2582 -4.3855
> 12 1 -1.2659 0.2584 -3.6545
> 13 2 -1.2659 -1.3798 -6.5783
> 14 2 0.0001 1.3800 -5.8473
> 15 2 0.0001 -1.3798 -4.3855
> 16 2 -1.2659 1.3800 -3.6545
> 17 1 1.2661 -0.2582 -6.5783
> 18 1 2.5321 0.2584 -5.8473
> 19 1 2.5321 -0.2582 -4.3855
> 20 1 1.2661 0.2584 -3.6545
> 21 2 1.2661 -1.3798 -6.5783
> 22 2 2.5321 1.3800 -5.8473
> 23 2 2.5321 -1.3798 -4.3855
> 24 2 1.2661 1.3800 -3.6545
> 25 1 -3.7979 -0.2582 -2.1927
> 26 1 -2.5319 0.2584 -1.4617
> 27 1 -2.5319 -0.2582 0.0001
> 28 1 -3.7979 0.2584 0.7310
> 29 2 -3.7979 -1.3798 -2.1927
> 30 2 -2.5319 1.3800 -1.4617
> 31 2 -2.5319 -1.3798 0.0001
> 32 2 -3.7979 1.3800 0.7310
> 33 1 -1.2659 -0.2582 -2.1927
> 34 1 0.0001 0.2584 -1.4617
> 35 1 0.0001 -0.2582 0.0001
> 36 1 -1.2659 0.2584 0.7310
> 37 2 -1.2659 -1.3798 -2.1927
> 38 2 0.0001 1.3800 -1.4617
> 39 2 0.0001 -1.3798 0.0001
> 40 2 -1.2659 1.3800 0.7310
> 41 1 1.2661 -0.2582 -2.1927
> 42 1 2.5321 0.2584 -1.4617
> 43 1 2.5321 -0.2582 0.0001
> 44 1 1.2661 0.2584 0.7310
> 45 2 1.2661 -1.3798 -2.1927
> 46 2 2.5321 1.3800 -1.4617
> 47 2 2.5321 -1.3798 0.0001
> 48 2 1.2661 1.3800 0.7310
> 49 1 -3.7979 -0.2582 2.1929
> 50 1 -2.5319 0.2584 2.9238
> 51 1 -2.5319 -0.2582 4.3857
> 52 1 -3.7979 0.2584 5.1166
> 53 2 -3.7979 -1.3798 2.1929
> 54 2 -2.5319 1.3800 2.9238
> 55 2 -2.5319 -1.3798 4.3857
> 56 2 -3.7979 1.3800 5.1166
> 57 1 -1.2659 -0.2582 2.1929
> 58 1 0.0001 0.2584 2.9238
> 59 1 0.0001 -0.2582 4.3857
> 60 1 -1.2659 0.2584 5.1166
> 61 2 -1.2659 -1.3798 2.1929
> 62 2 0.0001 1.3800 2.9238
> 63 2 0.0001 -1.3798 4.3857
> 64 2 -1.2659 1.3800 5.1166
> 65 1 1.2661 -0.2582 2.1929
> 66 1 2.5321 0.2584 2.9238
> 67 1 2.5321 -0.2582 4.3857
> 68 1 1.2661 0.2584 5.1166
> 69 2 1.2661 -1.3798 2.1929
> 70 2 2.5321 1.3800 2.9238
> 71 2 2.5321 -1.3798 4.3857
> 72 2 1.2661 1.3800 5.1166
> 0.0 0.0 0.0
> 7.5960000000000001 0.0 0.0
> 0.0 20.000000000000000 0.0
> 0.0 0.0 13.156727759999999
> }
> Driver = ConjugateGradient {
> MovedAtoms = Range { 1 -1 }
> MaxForceComponent [eV/AA] = 1.0e-2
> MaxSteps = 100
> # StepSize = 0.1
> OutputPrefix = "geo_end"
> }
> Hamiltonian = DFTB {
> SCC = Yes
> MaxAngularMomentum = {
> C = "p"
> H = "s"
> }
> SlaterKosterFiles = Type2FileNames {
> Prefix = "/home/schall2/Research/splines/slako/pbc-0-2/"
> Separator = "-"
> Suffix = ".skf"
> }
> KPointsAndWeights = SupercellFolding {
> 2 0 0
> 0 2 0
> 0 0 2
> 0.5 0.5 0.5
> }
> Eigensolver = Standard{}
> }
>
> On Fri, Mar 19, 2010 at 11:59 AM, <dftb-plus-user-request at dftb-plus.info
> <mailto:dftb-plus-user-request at dftb-plus.info>> wrote:
>
> Send DFTB-Plus-User mailing list submissions to
> dftb-plus-user at dftb-plus.info
> <mailto:dftb-plus-user at dftb-plus.info>
>
> To subscribe or unsubscribe via the World Wide Web, visit
> http://www.dftb-plus.info/mailman/listinfo/dftb-plus-user
> or, via email, send a message with subject or body 'help' to
> dftb-plus-user-request at dftb-plus.info
> <mailto:dftb-plus-user-request at dftb-plus.info>
>
> You can reach the person managing the list at
> dftb-plus-user-owner at dftb-plus.info
> <mailto:dftb-plus-user-owner at dftb-plus.info>
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of DFTB-Plus-User digest..."
>
>
> Today's Topics:
>
> 1. Convergence issues? (Dave Schall)
> 2. Re: Convergence issues? (B?lint Aradi)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Fri, 19 Mar 2010 11:39:36 -0400
> From: Dave Schall <schall2 at oakland.edu <mailto:schall2 at oakland.edu>>
> Subject: [DFTB-Plus-User] Convergence issues?
> To: dftb-plus-user at dftb-plus.info <mailto:dftb-plus-user at dftb-plus.info>
> Message-ID:
> <29919bb41003190839i1bd64f03r1263b15dc75723ea at mail.gmail.com
> <mailto:29919bb41003190839i1bd64f03r1263b15dc75723ea at mail.gmail.com>>
> Content-Type: text/plain; charset="iso-8859-1"
>
> Hello,
> I'm trying to do an energy minimization of slab-like planar
> structure and I
> cannot get convergence of steepest descents or cg method. In a
> separate, but
> possibly related issue, I can't get the scc calculation to converge
> either.
> I started with a 1x1 unit cell "slab" of the material, then a 2x2
> unit cell
> slab. These both converge in around 8 iterations. If I scale up to
> 3x3 cells
> or greater, I do not achieve convergence and the energy oscilates or
> actually goes up. Any suggestions for achieving convergence for energy
> minimization and / or scc? Ideally I'd like to scale up to a 10x5
> unit cell
> slab for my calculations.
>
> Attached are the input files for a 2x2 and 3x3 unit cell energy
> minimization
> calculations with scc off.
>
> Thanks in advance for your assistance!
>
> Dave Schall
>
> #2x2 unit cell: Converges just fine.
> Geometry = GenFormat {
> 32 S
> C H
> 1 1 -2.5319 -0.2582 -4.3855
> 2 1 -1.2659 0.2584 -3.6545
> 3 1 -1.2659 -0.2582 -2.1927
> 4 1 -2.5319 0.2584 -1.4617
> 5 2 -2.5319 -1.3798 -4.3855
> 6 2 -1.2659 1.3800 -3.6545
> 7 2 -1.2659 -1.3798 -2.1927
> 8 2 -2.5319 1.3800 -1.4617
> 9 1 0.0001 -0.2582 -4.3855
> 10 1 1.2661 0.2584 -3.6545
> 11 1 1.2661 -0.2582 -2.1927
> 12 1 0.0001 0.2584 -1.4617
> 13 2 0.0001 -1.3798 -4.3855
> 14 2 1.2661 1.3800 -3.6545
> 15 2 1.2661 -1.3798 -2.1927
> 16 2 0.0001 1.3800 -1.4617
> 17 1 -2.5319 -0.2582 0.0001
> 18 1 -1.2659 0.2584 0.7310
> 19 1 -1.2659 -0.2582 2.1929
> 20 1 -2.5319 0.2584 2.9238
> 21 2 -2.5319 -1.3798 0.0001
> 22 2 -1.2659 1.3800 0.7310
> 23 2 -1.2659 -1.3798 2.1929
> 24 2 -2.5319 1.3800 2.9238
> 25 1 0.0001 -0.2582 0.0001
> 26 1 1.2661 0.2584 0.7310
> 27 1 1.2661 -0.2582 2.1929
> 28 1 0.0001 0.2584 2.9238
> 29 2 0.0001 -1.3798 0.0001
> 30 2 1.2661 1.3800 0.7310
> 31 2 1.2661 -1.3798 2.1929
> 32 2 0.0001 1.3800 2.9238
> 0.0 0.0 0.0
> 5.0640000000000001 0.0 0.0
> 0.0 20.000000000000000 0.0
> 0.0 0.0 8.7711518399999999
> }
> Driver = ConjugateGradient {
> MovedAtoms = Range { 1 -1 }
> MaxForceComponent [eV/AA] = 1.0e-2
> MaxSteps = 100
> # StepSize = 0.1
> OutputPrefix = "geo_end"
> }
> Hamiltonian = DFTB {
> SCC = No
> MaxAngularMomentum = {
> C = "p"
> H = "s"
> # Eigensolver = Standard{}
> }
> SlaterKosterFiles = Type2FileNames {
> Prefix = "/home/schall2/Research/
> splines/slako/pbc-0-2/"
> Separator = "-"
> Suffix = ".skf"
> }
> KPointsAndWeights = SupercellFolding {
> 2 0 0
> 0 2 0
> 0 0 2
> 0.5 0.5 0.5
> }
> Eigensolver = Standard{}
> }
>
> #3x3 unit cell: Does not converge!
> Geometry = GenFormat {
> 72 S
> C H
> 1 1 -3.7979 -0.2582 -6.5783
> 2 1 -2.5319 0.2584 -5.8473
> 3 1 -2.5319 -0.2582 -4.3855
> 4 1 -3.7979 0.2584 -3.6545
> 5 2 -3.7979 -1.3798 -6.5783
> 6 2 -2.5319 1.3800 -5.8473
> 7 2 -2.5319 -1.3798 -4.3855
> 8 2 -3.7979 1.3800 -3.6545
> 9 1 -1.2659 -0.2582 -6.5783
> 10 1 0.0001 0.2584 -5.8473
> 11 1 0.0001 -0.2582 -4.3855
> 12 1 -1.2659 0.2584 -3.6545
> 13 2 -1.2659 -1.3798 -6.5783
> 14 2 0.0001 1.3800 -5.8473
> 15 2 0.0001 -1.3798 -4.3855
> 16 2 -1.2659 1.3800 -3.6545
> 17 1 1.2661 -0.2582 -6.5783
> 18 1 2.5321 0.2584 -5.8473
> 19 1 2.5321 -0.2582 -4.3855
> 20 1 1.2661 0.2584 -3.6545
> 21 2 1.2661 -1.3798 -6.5783
> 22 2 2.5321 1.3800 -5.8473
> 23 2 2.5321 -1.3798 -4.3855
> 24 2 1.2661 1.3800 -3.6545
> 25 1 -3.7979 -0.2582 -2.1927
> 26 1 -2.5319 0.2584 -1.4617
> 27 1 -2.5319 -0.2582 0.0001
> 28 1 -3.7979 0.2584 0.7310
> 29 2 -3.7979 -1.3798 -2.1927
> 30 2 -2.5319 1.3800 -1.4617
> 31 2 -2.5319 -1.3798 0.0001
> 32 2 -3.7979 1.3800 0.7310
> 33 1 -1.2659 -0.2582 -2.1927
> 34 1 0.0001 0.2584 -1.4617
> 35 1 0.0001 -0.2582 0.0001
> 36 1 -1.2659 0.2584 0.7310
> 37 2 -1.2659 -1.3798 -2.1927
> 38 2 0.0001 1.3800 -1.4617
> 39 2 0.0001 -1.3798 0.0001
> 40 2 -1.2659 1.3800 0.7310
> 41 1 1.2661 -0.2582 -2.1927
> 42 1 2.5321 0.2584 -1.4617
> 43 1 2.5321 -0.2582 0.0001
> 44 1 1.2661 0.2584 0.7310
> 45 2 1.2661 -1.3798 -2.1927
> 46 2 2.5321 1.3800 -1.4617
> 47 2 2.5321 -1.3798 0.0001
> 48 2 1.2661 1.3800 0.7310
> 49 1 -3.7979 -0.2582 2.1929
> 50 1 -2.5319 0.2584 2.9238
> 51 1 -2.5319 -0.2582 4.3857
> 52 1 -3.7979 0.2584 5.1166
> 53 2 -3.7979 -1.3798 2.1929
> 54 2 -2.5319 1.3800 2.9238
> 55 2 -2.5319 -1.3798 4.3857
> 56 2 -3.7979 1.3800 5.1166
> 57 1 -1.2659 -0.2582 2.1929
> 58 1 0.0001 0.2584 2.9238
> 59 1 0.0001 -0.2582 4.3857
> 60 1 -1.2659 0.2584 5.1166
> 61 2 -1.2659 -1.3798 2.1929
> 62 2 0.0001 1.3800 2.9238
> 63 2 0.0001 -1.3798 4.3857
> 64 2 -1.2659 1.3800 5.1166
> 65 1 1.2661 -0.2582 2.1929
> 66 1 2.5321 0.2584 2.9238
> 67 1 2.5321 -0.2582 4.3857
> 68 1 1.2661 0.2584 5.1166
> 69 2 1.2661 -1.3798 2.1929
> 70 2 2.5321 1.3800 2.9238
> 71 2 2.5321 -1.3798 4.3857
> 72 2 1.2661 1.3800 5.1166
> 0.0 0.0 0.0
> 7.5960000000000001 0.0 0.0
> 0.0 20.000000000000000 0.0
> 0.0 0.0 13.156727759999999
> }
> Driver = ConjugateGradient {
> MovedAtoms = Range { 1 -1 }
> MaxForceComponent [eV/AA] = 1.0e-2
> MaxSteps = 100
> # StepSize = 0.1
> OutputPrefix = "geo_end"
> }
> Hamiltonian = DFTB {
> SCC = No
> MaxAngularMomentum = {
> C = "p"
> H = "s"
> # Eigensolver = Standard{}
> }
> SlaterKosterFiles = Type2FileNames {
> Prefix = "/home/schall2/Research/splines/slako/pbc-0-2/"
> Separator = "-"
> Suffix = ".skf"
> }
> KPointsAndWeights = SupercellFolding {
> 2 0 0
> 0 2 0
> 0 0 2
> 0.5 0.5 0.5
> }
> Eigensolver = Standard{}
> }
>
>
> --
> J. David Schall, Assistant Professor
> Dept. of Mechanical Engineering, Oakland University
> 130 DHE, Rochester, MI 48309
> 248-370-2870
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL:
> http://www.dftb-plus.info/pipermail/dftb-plus-user/attachments/20100319/deefa1c6/attachment-0001.htm
>
> ------------------------------
>
> Message: 2
> Date: Fri, 19 Mar 2010 16:59:11 +0100
> From: B?lint Aradi <balint.aradi at bccms.uni-bremen.de
> <mailto:balint.aradi at bccms.uni-bremen.de>>
> Subject: Re: [DFTB-Plus-User] Convergence issues?
> To: User list for DFTB+ related questions
> <dftb-plus-user at dftb-plus.info
> <mailto:dftb-plus-user at dftb-plus.info>>
> Message-ID: <4BA39F4F.8000706 at bccms.uni-bremen.de
> <mailto:4BA39F4F.8000706 at bccms.uni-bremen.de>>
> Content-Type: text/plain; charset="utf-8"
>
> Dear Dave Schall,
>
> > I'm trying to do an energy minimization of slab-like planar
> structure and I
> > cannot get convergence of steepest descents or cg method. In a
> separate, but
> > possibly related issue, I can't get the scc calculation to
> converge either.
> > I started with a 1x1 unit cell "slab" of the material, then a 2x2
> unit cell
> > slab. These both converge in around 8 iterations. If I scale up to
> 3x3 cells
> > or greater, I do not achieve convergence and the energy oscilates or
> > actually goes up. Any suggestions for achieving convergence for energy
> > minimization and / or scc? Ideally I'd like to scale up to a 10x5
> unit cell
> > slab for my calculations.
> >
> > Attached are the input files for a 2x2 and 3x3 unit cell energy
> minimization
> > calculations with scc off.
>
> Actually I could not reproduce your problem. I took the 3x3 unit cell
> you sent, turned SCC on (which you should, as the pbc set is for SCC
> calculations) and got a converged geometry. (See attached input/output).
> I used dftb+_s081217, which I recommend to use.
>
> Best regards,
>
> B?lint
>
> --
> Dr. B?lint Aradi
> Bremen Center for Computational Materials Science, University of Bremen
> http://www.bccms.uni-bremen.de/en/~b_aradi
> <http://www.bccms.uni-bremen.de/en/%7Eb_aradi>
>
> -------------- next part --------------
> An embedded and charset-unspecified text was scrubbed...
> Name: output
> Url:
> http://www.dftb-plus.info/pipermail/dftb-plus-user/attachments/20100319/7ed23555/attachment.txt
> -------------- next part --------------
> An embedded and charset-unspecified text was scrubbed...
> Name: dftb_in.hsd
> Url:
> http://www.dftb-plus.info/pipermail/dftb-plus-user/attachments/20100319/7ed23555/attachment-0001.txt
> -------------- next part --------------
> A non-text attachment was scrubbed...
> Name: signature.asc
> Type: application/pgp-signature
> Size: 252 bytes
> Desc: OpenPGP digital signature
> Url :
> http://www.dftb-plus.info/pipermail/dftb-plus-user/attachments/20100319/7ed23555/attachment.pgp
>
> ------------------------------
>
> _______________________________________________
> DFTB-Plus-User mailing list
> DFTB-Plus-User at dftb-plus.info <mailto:DFTB-Plus-User at dftb-plus.info>
> http://www.dftb-plus.info/mailman/listinfo/dftb-plus-user
>
>
> End of DFTB-Plus-User Digest, Vol 27, Issue 7
> *********************************************
>
>
>
>
> --
> J. David Schall, Assistant Professor
> Dept. of Mechanical Engineering, Oakland University
> 130 DHE, Rochester, MI 48309
> 248-370-2870
>
--
Dr. B. Hourahine, SUPA, Department of Physics,
University of Strathclyde, John Anderson Building,
107 Rottenrow, Glasgow G4 0NG, UK.
+44 141 548 2325, benjamin.hourahine at strath.ac.uk
The University of Strathclyde is a charitable body,
registered in Scotland, with registration number SC015263
More information about the DFTB-Plus-User
mailing list