From eyvaz_isaev at yahoo.com Sun Oct 2 01:20:26 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Sat, 1 Oct 2005 16:20:26 -0700 (PDT) Subject: [Pw_forum] nbnd: correction In-Reply-To: <20050930125655.21950.qmail@web60322.mail.yahoo.com> Message-ID: <20051001232026.63034.qmail@web60313.mail.yahoo.com> Hi, I am sorry, in my recent replies there was an incorrect definition of bands number, nbnd. Actually, by deafult nbnd is defined as the largest of two numbers: nbnd=amax(nelec*(1+0.2)/2, nelec/2+4)) Bests, Eyvaz. __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From jasonsun98 at hotmail.com Sun Oct 2 04:28:38 2005 From: jasonsun98 at hotmail.com (sun jason) Date: Sun, 02 Oct 2005 02:28:38 +0000 Subject: [Pw_forum] about phonon frequencies In-Reply-To: <20051001054055.32319.38026.Mailman@democritos.sissa.it> Message-ID: Dear all, I found my calculated phonon frequencies are not accurate enough. for instance, when I vc-relax the cubic BN with LDA (B.pz-vbc.UPF and N.pz-vbc.UPF) the result is ------------- Final estimate of lattice vectors (input alat units) -0.491348786 0.000000000 0.491348786 0.000000000 0.491348786 0.491348786 -0.491348786 0.491348786 0.000000000 final unit-cell volume = 75.6348 (a.u.)^3 input alat = 6.8314 (a.u.) -------------- so I use 3.615*(0.491348786/0.5)=3.552 to calcalte phonon frequency. the result is 0.0000 0.0000 0.0000 1089.4410(TO) 1089.4410(TO) 1318.1950(LO) ---------they are too high if I vc-relax the lattice use GGA_PBE (B.pbe-n-van_ak.UPF and N.pbe-n-van_ak.UPF) the result is 3.618, a little large than experimental value. and the phonon frequencies are 0.0001 0.0001 0.0000 1029.5267 1029.5267 1270.6955 ---------they are too low so how can I get a more accurate phonon frequencies? any comment would be appreciate. best Jian SUN From vranjan at ncsu.edu Mon Oct 3 03:06:01 2005 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Sun, 2 Oct 2005 21:06:01 -0400 (EDT) Subject: [Pw_forum] pseudopotential In-Reply-To: <20051001054055.32319.38026.Mailman@democritos.sissa.it> References: <20051001054055.32319.38026.Mailman@democritos.sissa.it> Message-ID: <34160.152.14.74.99.1128301561.squirrel@webmail.ncsu.edu> Dear Users, I am looking for a pseudopotential for Pb with following : Perdew-Wang 91 gradient-corrected functional nonlinear core-correction semicore state d in valence Vanderbilt ultrasoft Is it there with someone ? The problem I am facing is as follows : I have several different types of atoms in a my calculation. I could find one common type (with above attributes) for all atoms, except Pb. For Pb, I am using Perdew-Zunger LDA type of pseudopotential. The programme stops with the following message : from readpp : error # 4 inconsistent DFT read I would really appreciate if some help could come. Thank you, Vivek Ranjan From giannozz at nest.sns.it Mon Oct 3 10:18:53 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 3 Oct 2005 10:18:53 +0200 Subject: [Pw_forum] About compiling cd.x and fpmd.x In-Reply-To: <008901c5c5a2$ddf19ca0$1d00a8c0@solarflare> References: <008901c5c5a2$ddf19ca0$1d00a8c0@solarflare> Message-ID: <200510031018.53956.giannozz@nest.sns.it> On Friday 30 September 2005 11:39, Huiqun Zhou wrote: > fpmd.x is the worst, it fails in both serial and parallel runs. there is a problem with a bad allocation. Replace FPMD/pseudo_base.f90 with the attached version. Concerning error in parallel execution, read what the manual says about files and file systems Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy -------------- next part -------------- A non-text attachment was scrubbed... Name: pseudo_base.f90 Type: text/x-fortran Size: 13460 bytes Desc: not available Url : /pipermail/attachments/20051003/6b58e136/attachment.bin From giannozz at nest.sns.it Mon Oct 3 10:25:17 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 3 Oct 2005 10:25:17 +0200 Subject: [Pw_forum] El-Ph funny error In-Reply-To: <433C1454.5060800@lg.ehu.es> References: <433C1454.5060800@lg.ehu.es> Message-ID: <200510031025.18007.giannozz@nest.sns.it> On Thursday 29 September 2005 18:20, Miguel Mart?nez Canales wrote: >from phq_init : error # 1 >wrong order of k points IF ( ABS( xq(1) - ( xk(1,ikq) - xk(1,ikk) ) ) > eps8 .OR. & ABS( xq(2) - ( xk(2,ikq) - xk(2,ikk) ) ) > eps8 .OR. & ABS( xq(3) - ( xk(3,ikq) - xk(3,ikk) ) ) > eps8 ) THEN WRITE( stdout, * ) ikk, ikq, nksq WRITE( stdout, * ) ( xq(ipol), ipol = 1, 3) WRITE( stdout, * ) ( xk(ipol,ikq), ipol = 1, 3) WRITE( stdout, * ) ( xk(ipol,ikk), ipol = 1, 3) CALL errore( 'phq_init', 'wrong order of k points', 1 ) END IF what does it print immediately before that? -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Mon Oct 3 11:39:39 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 3 Oct 2005 11:39:39 +0200 Subject: [Pw_forum] WARNING: something is going wrong In-Reply-To: <200509261533.22676.alberto.milani@chem.polimi.it> References: <200509261533.22676.alberto.milani@chem.polimi.it> Message-ID: <200510031139.39187.giannozz@nest.sns.it> On Monday 26 September 2005 15:33, Alberto Milani wrote: > WARNING : something is going wrong > resetting bfgs history > What does it exactly mean ? "exactly" I don't know, but basically the implementation of the bgfs algorithm may run into trouble when you are very close to convergence (look at the forces: they are exceedingly small). Nothing to worry about. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Mon Oct 3 15:10:10 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 3 Oct 2005 15:10:10 +0200 Subject: [Pw_forum] pp freezes In-Reply-To: <4337B814.5050007@Princeton.EDU> References: <43369F8D.9030302@Princeton.EDU> <4337B814.5050007@Princeton.EDU> Message-ID: <200510031510.10298.giannozz@nest.sns.it> On Monday 26 September 2005 10:57, Silviu Zilberman wrote: > It works however for the parallel code when using just one processor. > I think it fails somewhere in subroutine 'data_structure', it actually fails there. The pp.x code cannot (yet) perform the second step, i.e. writing to file in the required format for plotting, on more than one processor. It needs to be fixed > which is called from allocate_fft. Why do we need a second call > to allocate_fft, it is already done once in the first part of pp. it is already done only if you are executing the first part, i.e. extracting the required quantity from data files. You may however skip the first part and read the required quantity from an intermediate file. This is useful in some circumstances. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Mon Oct 3 15:14:16 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 3 Oct 2005 15:14:16 +0200 Subject: [Pw_forum] symmetry error In-Reply-To: <20050926071215.97712.qmail@web33605.mail.mud.yahoo.com> References: <20050926071215.97712.qmail@web33605.mail.mud.yahoo.com> Message-ID: <200510031514.16368.giannozz@nest.sns.it> On Monday 26 September 2005 09:12, Jigang Zhang wrote: > I am running a vc-relax calculation. The output > gives this error: it is not an error > from setup : error # -1 > Dynamics, you should have no symmetries it just means "you should not use symmetries in this particular case". If you do, and know what you are doing, it makes no harm Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Mon Oct 3 16:00:28 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 3 Oct 2005 16:00:28 +0200 Subject: [Pw_forum] Electron-Phonon: Two questions In-Reply-To: <4332C0F5.2090106@lg.ehu.es> References: <4332C0F5.2090106@lg.ehu.es> Message-ID: <200510031600.28117.giannozz@nest.sns.it> On Thursday 22 September 2005 16:34, Miguel Mart?nez Canales wrote: > - One of the calculations, although it yields results, gives me the > following error message: > > task # 0 > from set_kplusq : error # 0 > too many k points > > should I be worried? maybe: it is a very weird error. Could you please provide a test job? > - On the second non-selfconsistent run, with a dense grid [..] > sometimes kpoint counting surpasses k(9999) and k(****) follows. > Is this the source of the previous error ? most likely, no Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Mon Oct 3 16:12:18 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 3 Oct 2005 16:12:18 +0200 Subject: [Pw_forum] pseudopotential for studying high-pressure elements In-Reply-To: <6ac064b60509091741413bca77@mail.gmail.com> References: <6ac064b60509091741413bca77@mail.gmail.com> Message-ID: <200510031612.18187.giannozz@nest.sns.it> On Saturday 10 September 2005 02:41, Nichols A. Romero wrote: > Will the PP give meaning results once the atomic seperation is less > than r_c. In general, I would think not so would I. If the sum of the r_c is of the same order of the interatomic distance, the error might be minor, but if it is definitely larger, I wouldn't trust too much the results > So the next question is that for a fixed r_c what PP will be softess: > TM, RRKJ, Ultrasoft-PP, etc.? Ultrasoft are ultra, and not just soft, because they allow larger r_c, not because their pseudization produces intrinsically softer PPs. I know people working in high-pressure physics that do not trust ultrasoft PPs for this reason. I have heard that RRKJ pseudization is very effective in producing soft PPs, but it occasionally fails; TM always work Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From adeagbo at thp.Uni-Duisburg.DE Tue Oct 4 11:29:21 2005 From: adeagbo at thp.Uni-Duisburg.DE (Adeagbo Waheed Adeniyi) Date: Tue, 4 Oct 2005 11:29:21 +0200 (MESZ) Subject: [Pw_forum] pseudopotential for bismuth In-Reply-To: <20051004053633.7187.71287.Mailman@democritos.sissa.it> References: <20051004053633.7187.71287.Mailman@democritos.sissa.it> Message-ID: Dear PWSCF users, Does anybody have a pseudopotential for bismuth? Thanks Adeagbo =================================================== ADEAGBO WAHEED ADENIYI Institute of Physics, MC 328 University Duisburg-Essen Lotharstr. 1 47048 Duisburg Germany ==================================================== Phone: +49 203 379 1606 Fax: +49 203 379 3665 Present e-mail address: adeagbo at thp.uni-duisburg.de Permanent e-mail address: wadeagbo at mail.skannet.com http://www.thp.uni-duisburg.de ==================================================== "The intellect has little to do on the road to discovery. There comes a leap in consciousness, call it intuition or what you will, and the solution comes to you and you don't know how or why." -- AlbertEinstein From wmbmacam at lg.ehu.es Tue Oct 4 18:20:15 2005 From: wmbmacam at lg.ehu.es (=?UTF-8?B?TWlndWVsIE1hcnTDrW5leiBDYW5hbGVz?=) Date: Tue, 04 Oct 2005 18:20:15 +0200 Subject: [Pw_forum] El-Ph funny error In-Reply-To: <200510031025.18007.giannozz@nest.sns.it> References: <433C1454.5060800@lg.ehu.es> <200510031025.18007.giannozz@nest.sns.it> Message-ID: <4342ABBF.5060905@lg.ehu.es> Dear Dr. Gianozzi First, thanks to you and Dr.Isaev for your answers. > > IF ( ABS( xq(1) - ( xk(1,ikq) - xk(1,ikk) ) ) > eps8 .OR. & > ABS( xq(2) - ( xk(2,ikq) - xk(2,ikk) ) ) > eps8 .OR. & > ABS( xq(3) - ( xk(3,ikq) - xk(3,ikk) ) ) > eps8 ) THEN > WRITE( stdout, * ) ikk, ikq, nksq > WRITE( stdout, * ) ( xq(ipol), ipol = 1, 3) > WRITE( stdout, * ) ( xk(ipol,ikq), ipol = 1, 3) > WRITE( stdout, * ) ( xk(ipol,ikk), ipol = 1, 3) > CALL errore( 'phq_init', 'wrong order of k points', 1 ) > END IF > > what does it print immediately before that? > The outpur (from the last k point to the error is): k(16250) = ( 0.4200000 1.4200000 -0.2500000), wk = 0.0000000 pseudo 1 is Al (US) zval = 3.0 lmax= 1 lloc= 0 Version 0 0 0 of US pseudo code Using log mesh of 893 points The pseudopotential has 2 beta functions with: l(1) = 0 l(2) = 1 Q(r) pseudized with 8 coefficients, rinner = 0.900 0.900 0.900 Atomic displacements: There are 3 irreducible representations Representation 1 1 modes - To be done Representation 2 1 modes - To be done Representation 3 1 modes - To be done 1 2 8125 -0.125000000000000 -0.125000000000000 0.375000000000000 -2.00000000000000D-002 2.00000000000000D-002 0.270000000000000 -2.00000000000000D-002 2.00000000000000D-002 2.00000000000000D-002 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from phq_init : error # 1 wrong order of k points %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... If it helps, this tipically occurs in low symmetry points (i.e. outside 00x or xxx lines). Thanks again, and sorry for the late answer, Miguel -- ---------------------------------------- Miguel Mart?nez Canales Dto. F?sica de la Materia Condensada UPV/EHU Facultad de Ciencia y Tecnolog?a Apdo. 644 48080 Bilbao (Spain) Fax: +34 94 601 3500 Tlf: +34 94 601 5437 ---------------------------------------- 2005 International Year of Physics From giannozz at nest.sns.it Tue Oct 4 18:34:24 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 4 Oct 2005 18:34:24 +0200 Subject: [Pw_forum] El-Ph funny error In-Reply-To: <4342ABBF.5060905@lg.ehu.es> References: <433C1454.5060800@lg.ehu.es> <200510031025.18007.giannozz@nest.sns.it> <4342ABBF.5060905@lg.ehu.es> Message-ID: <200510041834.24814.giannozz@nest.sns.it> On Tuesday 04 October 2005 18:20, Miguel Mart?nez Canales wrote: > -0.125000000000000 -0.125000000000000 0.375000000000000 > -2.00000000000000D-002 2.00000000000000D-002 0.270000000000000 > -2.00000000000000D-002 2.00000000000000D-002 2.00000000000000D-002 so it is not a truncation problem. Could you please submit a tets job so that the error can be reproduced? Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From swblelia at sw.ehu.es Tue Oct 4 18:46:15 2005 From: swblelia at sw.ehu.es (Aritz Leonardo) Date: Tue, 04 Oct 2005 18:46:15 +0200 Subject: [Pw_forum] El-Ph funny error In-Reply-To: <4342ABBF.5060905@lg.ehu.es> References: <433C1454.5060800@lg.ehu.es> <200510031025.18007.giannozz@nest.sns.it> <4342ABBF.5060905@lg.ehu.es> Message-ID: <4342B1D7.7040804@sw.ehu.es> I think that it is a compiler error. Try to calculate the samething in such a way that the nscf calculation does not take more than 9999 points. I am sure that in this case you will obtain results. As you say, if the point has low symmetry and kpoints > 9999..... crashes. I am sorry that I cannot help you. I had the same problem and the way to solve it was using another compiler. Miguel Mart?nez Canales escribi?: > Dear Dr. Gianozzi > > First, thanks to you and Dr.Isaev for your answers. > >> >> IF ( ABS( xq(1) - ( xk(1,ikq) - xk(1,ikk) ) ) > eps8 .OR. & >> ABS( xq(2) - ( xk(2,ikq) - xk(2,ikk) ) ) > eps8 .OR. & >> ABS( xq(3) - ( xk(3,ikq) - xk(3,ikk) ) ) > eps8 ) THEN >> WRITE( stdout, * ) ikk, ikq, nksq >> WRITE( stdout, * ) ( xq(ipol), ipol = 1, 3) >> WRITE( stdout, * ) ( xk(ipol,ikq), ipol = 1, 3) >> WRITE( stdout, * ) ( xk(ipol,ikk), ipol = 1, 3) >> CALL errore( 'phq_init', 'wrong order of k points', 1 ) >> END IF >> >> what does it print immediately before that? >> > > The outpur (from the last k point to the error is): > > > k(16250) = ( 0.4200000 1.4200000 -0.2500000), wk = > 0.0000000 > > pseudo 1 is Al (US) zval = 3.0 lmax= 1 lloc= 0 > Version 0 0 0 of US pseudo code > > Using log mesh of 893 points > > The pseudopotential has 2 beta functions with: > > l(1) = 0 > l(2) = 1 > > Q(r) pseudized with 8 coefficients, rinner = 0.900 0.900 > 0.900 > > > > Atomic displacements: > There are 3 irreducible representations > > Representation 1 1 modes - To be done > > Representation 2 1 modes - To be done > > Representation 3 1 modes - To be done > 1 2 8125 > -0.125000000000000 -0.125000000000000 0.375000000000000 > -2.00000000000000D-002 2.00000000000000D-002 0.270000000000000 > -2.00000000000000D-002 2.00000000000000D-002 2.00000000000000D-002 > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > from phq_init : error # 1 > wrong order of k points > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > > stopping ... > > > If it helps, this tipically occurs in low symmetry points (i.e. > outside 00x or xxx lines). Thanks again, and sorry for the late answer, > > Miguel > From wmbmacam at lg.ehu.es Tue Oct 4 18:55:33 2005 From: wmbmacam at lg.ehu.es (=?UTF-8?B?TWlndWVsIE1hcnTDrW5leiBDYW5hbGVz?=) Date: Tue, 04 Oct 2005 18:55:33 +0200 Subject: [Pw_forum] El-Ph funny error In-Reply-To: <200510041834.24814.giannozz@nest.sns.it> References: <433C1454.5060800@lg.ehu.es> <200510031025.18007.giannozz@nest.sns.it> <4342ABBF.5060905@lg.ehu.es> <200510041834.24814.giannozz@nest.sns.it> Message-ID: <4342B405.8000304@lg.ehu.es> Dear Dr. Gianozzi, > so it is not a truncation problem. Could you please submit a tets > job so that the error can be reproduced? The batch I'm submitting should make it... although I will test it tonight. For this particular point, the lines you quoted are: -0.500000000000000 -0.250000000000000 1.00000000000000 -1.92307692307692D-002 1.92307692307692D-002 1.01923076923077 -1.92307692307692D-002 1.92307692307692D-002 1.92307692307692D-002 The job is: [...] # self-consistent calculation cat > al.scf.in << EOF &control calculation='scf' restart_mode='from_scratch', tstress = .true. tprnfor = .true. prefix='al', pseudo_dir = '$PSEUDO_DIR/', outdir='$TMP_DIR/' / &system ibrav= 2, celldm(1) =7.603, nat= 1, ntyp= 1, ecutwfc =40.0, occupations='smearing', smearing='gauss', degauss=0.04 / &electrons conv_thr = 1.0d-8 mixing_beta = 0.7 / ATOMIC_SPECIES Al 26.9815 Al.pbe-n-van.UPF ATOMIC_POSITIONS Al 0.00 0.00 0.00 K_POINTS {automatic} 26 26 26 1 1 1 EOF $ECHO " calculo autoconsistente...\c" $PW_COMMAND < al.scf.in > al.scf.out $ECHO " hecho" # non self-consistent calculation cat > al.nscf.in << EOF &control calculation='phonon' restart_mode='from_scratch', prefix='al', pseudo_dir = '$PSEUDO_DIR/', outdir='$TMP_DIR/' / &system ibrav= 2, celldm(1) =7.603, nat= 1, ntyp= 1, ecutwfc =40.0, occupations='smearing', smearing='gauss', degauss=0.04 / &electrons conv_thr = 1.0d-8 / &phonon xqq(1) = -0.500, xqq(2) = -0.250, xqq(3) = 1.000 / ATOMIC_SPECIES Al 26.9815 Al.pbe-n-van.UPF ATOMIC_POSITIONS Al 0.00 0.00 0.00 K_POINTS {automatic} 26 26 26 1 1 1 EOF $ECHO " calculo no autoconsistente para q=-0.500,-0.250,1.000..\c" $PW_COMMAND < al.nscf.in >> al.nscf.out $ECHO " hecho" # phonon calculation cat > al.ph.in << EOF phonons of Al at $qpoint &inputph tr2_ph=1.0d-12, prefix='al', amass(1)=26.9815, outdir='$TMP_DIR/', fildyn='dyn.$qpoint' fildvscf='aldv' elph=.true., trans=.true., / -0.500 -0.250 1.000 EOF $ECHO " calculo de lambda y gamma para q=$qpoint...\c" $PH_COMMAND < al.ph.in >> al.ph.$qpoint.out $ECHO " hecho" -- ---------------------------------------- Miguel Mart?nez Canales Dto. F?sica de la Materia Condensada UPV/EHU Facultad de Ciencia y Tecnolog?a Apdo. 644 48080 Bilbao (Spain) Fax: +34 94 601 3500 Tlf: +34 94 601 5437 ---------------------------------------- 2005 International Year of Physics "But you don't ask with respect. You don't offer friendship. You don't even think to call me Godfather." Don Vito Corleone From giannozz at nest.sns.it Tue Oct 4 19:03:00 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 4 Oct 2005 19:03:00 +0200 Subject: [Pw_forum] El-Ph funny error In-Reply-To: <4342B1D7.7040804@sw.ehu.es> References: <433C1454.5060800@lg.ehu.es> <4342ABBF.5060905@lg.ehu.es> <4342B1D7.7040804@sw.ehu.es> Message-ID: <200510041903.00303.giannozz@nest.sns.it> On Tuesday 04 October 2005 18:46, Aritz Leonardo wrote: > I am sorry that I cannot help you. I had the same problem > and the way to solve it was using another compiler. which compiler gave problems, and which one didn't? P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From swblelia at sw.ehu.es Tue Oct 4 19:15:22 2005 From: swblelia at sw.ehu.es (Aritz Leonardo) Date: Tue, 04 Oct 2005 19:15:22 +0200 Subject: [Pw_forum] El-Ph funny error In-Reply-To: <200510041903.00303.giannozz@nest.sns.it> References: <433C1454.5060800@lg.ehu.es> <4342ABBF.5060905@lg.ehu.es> <4342B1D7.7040804@sw.ehu.es> <200510041903.00303.giannozz@nest.sns.it> Message-ID: <4342B8AA.3090608@sw.ehu.es> http://www.democritos.it/pipermail/pw_forum/2005-January/001955.html http://www.democritos.it/pipermail/pw_forum/2004-October/001387.html Paolo Giannozzi escribi?: >On Tuesday 04 October 2005 18:46, Aritz Leonardo wrote: > > > >>I am sorry that I cannot help you. I had the same problem >>and the way to solve it was using another compiler. >> >> > >which compiler gave problems, and which one didn't? > >P. > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051004/3ac6ccf1/attachment.htm From swblelia at sw.ehu.es Tue Oct 4 19:20:17 2005 From: swblelia at sw.ehu.es (Aritz Leonardo) Date: Tue, 04 Oct 2005 19:20:17 +0200 Subject: [Pw_forum] funny error Message-ID: <4342B9D1.6000107@sw.ehu.es> http://www.democritos.it/pipermail/pw_forum/2005-January/001939.html From hqzhou at nju.edu.cn Wed Oct 5 04:35:21 2005 From: hqzhou at nju.edu.cn (Huiqun Zhou) Date: Wed, 5 Oct 2005 10:35:21 +0800 Subject: [Pw_forum] About compiling cd.x and fpmd.x References: <008901c5c5a2$ddf19ca0$1d00a8c0@solarflare> <200510031018.53956.giannozz@nest.sns.it> Message-ID: <002501c5c955$6ef832d0$1d00a8c0@solarflare> Thank you very much, Paolo, for your providing the fix! Everything seems OK now. I'll report performance figures on my machines and share with the list users. Regards, Huiqun Zhou ----- Original Message ----- From: "Paolo Giannozzi" To: Sent: Monday, October 03, 2005 4:18 PM Subject: Re: [Pw_forum] About compiling cd.x and fpmd.x > On Friday 30 September 2005 11:39, Huiqun Zhou wrote: > >> fpmd.x is the worst, it fails in both serial and parallel runs. > > there is a problem with a bad allocation. Replace > FPMD/pseudo_base.f90 with the attached version. > Concerning error in parallel execution, read what > the manual says about files and file systems > > Paolo > > -- > Paolo Giannozzi e-mail: giannozz at nest.sns.it > Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 > Piazza dei Cavalieri 7 I-56126 Pisa, Italy > From wmbmacam at lg.ehu.es Wed Oct 5 11:24:27 2005 From: wmbmacam at lg.ehu.es (=?UTF-8?B?TWlndWVsIE1hcnTDrW5leiBDYW5hbGVz?=) Date: Wed, 05 Oct 2005 11:24:27 +0200 Subject: [Pw_forum] Electron-Phonon: Two questions In-Reply-To: <200510031600.28117.giannozz@nest.sns.it> References: <4332C0F5.2090106@lg.ehu.es> <200510031600.28117.giannozz@nest.sns.it> Message-ID: <43439BCB.4010902@lg.ehu.es> Dear Dr. Gianozzi, Maybe this has something to do with the other el-ph thread. Anyway, with the test job I'm providing, I get the following nscf output. Am I using a kpoint grid that is too large? Job: [...] # self-consistent calculation cat > al.scf.in << EOF &control calculation='scf' restart_mode='from_scratch', tstress = .true. tprnfor = .true. prefix='al', pseudo_dir = '$PSEUDO_DIR/', outdir='$TMP_DIR/' / &system ibrav= 2, celldm(1) =7.603, nat= 1, ntyp= 1, ecutwfc =40.0, occupations='smearing', smearing='gauss', degauss=0.04 / &electrons conv_thr = 1.0d-8 mixing_beta = 0.7 / ATOMIC_SPECIES Al 26.9815 Al.pbe-n-van.UPF ATOMIC_POSITIONS Al 0.00 0.00 0.00 K_POINTS {automatic} 32 32 32 1 1 1 EOF $ECHO " running the scf calculation...\c" $PW_COMMAND < al.scf.in > al.scf.out $ECHO " done" cat > al.nscf.in << EOF &control calculation='phonon' restart_mode='from_scratch', prefix='al', pseudo_dir = '$PSEUDO_DIR/', outdir='$TMP_DIR/' / &system ibrav= 2, celldm(1) =7.603, nat= 1, ntyp= 1, ecutwfc =40.0, occupations='smearing', smearing='gauss', degauss=0.04 / &electrons conv_thr = 1.0d-8 / &phonon xqq(1) = -0.125, xqq(2) = -0.125, xqq(3) = -0.125 / ATOMIC_SPECIES Al 26.9815 Al.pbe-n-van.UPF ATOMIC_POSITIONS Al 0.00 0.00 0.00 K_POINTS {automatic} 32 32 32 1 1 1 EOF $ECHO " running the nscf calculation at q=$qpoint...\c" $PW_COMMAND < al.nscf.in >> al.nscf.out $ECHO " done" [...] It follows with a ph.x input file with elph.and.trans=.true. I won't submit this, as the error I get is in the nscf run. I was using just 2gb of ram. The nscf output is: Program PWSCF v.2.1.5 starts ... Today is 5Oct2005 at 3:31:22 Parallel version (MPI) Number of processors in use: 2 R & G space division: nprocp = 2 Ultrasoft (Vanderbilt) Pseudopotentials Current dimensions of program pwscf are: ntypx =10 npk =40000 lmax = 3 nchix = 6 ndmx = 2000 nbrx =14 nqfx = 8 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from set_kplusq : error # 22880 too many k points %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... Thanks again, Miguel Paolo Giannozzi wrote: > On Thursday 22 September 2005 16:34, Miguel Mart?nez Canales wrote: > > >>- One of the calculations, although it yields results, gives me the >>following error message: >> >> task # 0 >> from set_kplusq : error # 0 >> too many k points >> >>should I be worried? > > > maybe: it is a very weird error. Could you please provide a test job? > > >>- On the second non-selfconsistent run, with a dense grid [..] >>sometimes kpoint counting surpasses k(9999) and k(****) follows. >>Is this the source of the previous error ? > > > most likely, no > > Paolo > -- ---------------------------------------- Miguel Mart?nez Canales Dto. F?sica de la Materia Condensada UPV/EHU Facultad de Ciencia y Tecnolog?a Apdo. 644 48080 Bilbao (Spain) Fax: +34 94 601 3500 Tlf: +34 94 601 5437 ---------------------------------------- 2005 International Year of Physics From g.ballabio at cineca.it Wed Oct 5 11:37:23 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Wed, 5 Oct 2005 11:37:23 +0200 (MEST) Subject: [Pw_forum] symmetry error In-Reply-To: <200510031514.16368.giannozz@nest.sns.it> (from giannozz@nest.sns.it on Mon Oct 3 15:14:16 2005) References: <20050926071215.97712.qmail@web33605.mail.mud.yahoo.com> <200510031514.16368.giannozz@nest.sns.it> Message-ID: <1128505041l.4361l.2l@nb-ballabio.cineca.it> On 10/03/2005 03:14:16 PM, Paolo Giannozzi wrote: > On Monday 26 September 2005 09:12, Jigang Zhang wrote: > > > I am running a vc-relax calculation. The output > > gives this error: > > it is not an error > > > from setup : error # -1 > > Dynamics, you should have no symmetries > > it just means "you should not use symmetries in this particular > case". If you do, and know what you are doing, it makes no harm In this case, it would rather be a warning? Gerardo From giannozz at nest.sns.it Wed Oct 5 14:55:35 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 5 Oct 2005 14:55:35 +0200 Subject: [Pw_forum] Electron-Phonon: Two questions In-Reply-To: <43439BCB.4010902@lg.ehu.es> References: <4332C0F5.2090106@lg.ehu.es> <200510031600.28117.giannozz@nest.sns.it> <43439BCB.4010902@lg.ehu.es> Message-ID: <200510051455.35561.giannozz@nest.sns.it> On Wednesday 05 October 2005 11:24, Miguel Mart?nez Canales wrote: > Maybe this has something to do with the other el-ph thread. > Anyway, with the test job I'm providing, I get the following nscf > output. Am I using a kpoint grid that is too large? > from set_kplusq : error # 22880 > too many k points definitely so. The maximum allowable number of k-points is statically defined in variable npk, in Modules/parameters.f90. You have to increase it and to recompile. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From silviu at Princeton.EDU Wed Oct 5 23:55:36 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Wed, 05 Oct 2005 23:55:36 +0200 Subject: [Pw_forum] g95 compilation problem In-Reply-To: <200510030958.26516.p.giannozzi@sns.it> References: <20050927110733.39CC911280D@democritos.sissa.it> <1127922487.433abb37a1aaa@webmail.sissa.it> <433D9BDF.1050502@Princeton.EDU> <200510030958.26516.p.giannozzi@sns.it> Message-ID: <43444BD8.4020601@Princeton.EDU> Dear Paolo, I do have a working version of Q-Espresso on cygwin/g95, attached is the make.sys file. Unfortunately I am not familiar enough with the configure utility to modify the configure.ac file myself. The two major changes with respect to the original make.sys are: 1. The "-o" in the preprocessing step is not working properly and I had to use the redirection symbol instead. 2. There was a case problem. The original preprocessing is generating f90->F90. In cygwin it overwrites the original f90 file and all sort of strange things then happen. It seems to be a cygwin problem, ignoring the case of the extension. Therefore I changed the extension of the preprocessed file to F95 and added a rule for F95->object file. I get executables that seem to run properly but I have one problem: For some reason the path of (say) the pseudopotentials (PP) library is unrecognized by the code and it complains that it cannot fine the PP. The only way I found around it is to copy the PP to the local directory (symbolic link is not enough). The same problem exists with the working directory (outdir). For instance example 01, scf calculation for Ni using CG: &control calculation='scf' restart_mode='from_scratch', pseudo_dir = '/home/silviu/Work/O-sesame/pseudo/', outdir='/home/silviu/tmp/' prefix='ni' tprnfor = .true., tstress = .true. / ... output: Program PWSCF v.3.0 starts ... Today is 2Oct2005 at 23:28:19 Ultrasoft (Vanderbilt) Pseudopotentials Current dimensions of program pwscf are: ntypx = 10 npk = 40000 lmax = 3 nchix = 6 ndmx = 2000 nbrx = 14 nqfx = 8 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from readpp : error # 3 file /home/silviu/Work/O-sesame/pseudo/NiUS.RRKJ3.UPF not found %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... ~ However the file is there: $ ls /home/silviu/Work/O-sesame/pseudo/ Al.vbc.UPF H.fpmd.UPF NiUS.RRKJ3.UPF O_US.van As.gon.UPF H.vbc.UPF O.BLYP.UPF Pb.vdb.UPF C.meta.tm.UPF HUSPBE.RRKJ3 O.LDA.US.RRKJ3.UPF Ptrel.RRKJ3.UPF C.pz-rrkjus.UPF H_US.van O.pbe-rrkjus.UPF Si.recon CVS/ Hmeta.tm.UPF O.recon Si.vbc.UPF Cu.pz-d-rrkjus.UPF N.BLYP.UPF O.vdb.UPF SiPBE_nc.UPF Fe.pz-nd-rrkjus.UPF Ni.pbe-nd-rrkjus.UPF OPBE_nc.UPF Ti.vdb.UPF Any ideas where is the problem? Thanks, Silviu. Paolo Giannozzi wrote: >On Friday 30 September 2005 22:11, Silviu Zilberman wrote: > > > >>Thank you for your comments. I had to introduce a couple of other >>minor changes but now the code compiles. Running however seems >>still to be a problem, but I hope to make it work. >> >> > >Hi Silviu > >if you manage to prodce working executables, please let us >know how you did it. Note that g95 is improving very quickly: >try the very latest release > >Paolo > > > -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: make.sys.cygwin_g95 Url: /pipermail/attachments/20051005/5bcfb4cf/attachment.txt -------------- next part -------------- A non-text attachment was scrubbed... Name: silviu.vcf Type: text/x-vcard Size: 247 bytes Desc: not available Url : /pipermail/attachments/20051005/5bcfb4cf/attachment.vcf From alberto.milani at chem.polimi.it Thu Oct 6 10:28:51 2005 From: alberto.milani at chem.polimi.it (Alberto Milani) Date: Thu, 6 Oct 2005 10:28:51 +0200 Subject: [Pw_forum] Negative frequencies in phonon calculation Message-ID: <200510061028.52070.alberto.milani@chem.polimi.it> Dear all, I am calculating phonon frequencies for an infinite linear carbon chain and up to now I have obtained puzzling results. The two degenerate acoustic branches relative to transverse modes show large negative frequencies. At Gamma I found omegas of -100 cm-1. Then the curve begins to rise up to a value of 50 cm-1 but then a sharp discontinuity occurs and the frequencies fall down to -700 cm-1. In my calculation, I studied a well relaxed structure (Peierls distorted) and I tried also to increase the quality of the results by using all the tricks explained in the manual. My cell has parameters a=10, b=10 and c=2.5671, I am using Vanderbilt ultrasoft pseudopotential with PBE GGA and a K-POINTS grid of 1 1 70. By increasing ecutwfc and/or ecutrho (at the beginning I chose 35-350 Ry) I have obtained no better results and also by reducing convergence threshold both for SCF and phonons calculation nothing goes better. The same bad results are found by improving the k points grid (e.g. 10 10 70). In my calculation, I have directly computed the dynamical matrices without interpolating with q2r.x and matdyn.x codes and I have chosen 15 q-points in the irreducible Brillouin zone (the q vector has been chosen to be parallel to the linear chain axis). All the other branches (one longitudinal acoustic and three optical branches) are good and I don't understand what is going bad in the computation. Does anybody know what is the source of the problem? Thanks in advance. A. M. From jcconesa at icp.csic.es Thu Oct 6 11:06:10 2005 From: jcconesa at icp.csic.es (Jose C. Conesa) Date: Thu, 6 Oct 2005 11:06:10 +0200 Subject: [Pw_forum] Negative frequencies in phonon calculation In-Reply-To: <200510061028.52070.alberto.milani@chem.polimi.it> Message-ID: <43450522.154.4D6E5818@localhost> Hello, Assuming that there are no problems of numerical inaccuracy or inadequate convergence thresholds, negative frequencies at gamma point in a geometry-relaxed system usually mean that (within the system and the calculation method used) there is some distortion (decreasing the symmetry of the system) which leads to lower energy, i.e. you are at a saddle point in the energy hypersurface. You must move some atom out of the initial symmetry and optimize geometry again starting from that modified situation. If you can obtain the vector coordinates for the normal mode of vibration corresponding to the negative frequency, a displacement of the involved atoms, proportional to that normal mode, by e.g. a maximum of 0.2 angstrom will give to you a starting displaced geometry closer to the true energy minimum. Sometimes you need to do this several times if not all of the negative frequencies disappear after the first attempt. If, after obtaining a new geometry with no negative frequencies at the gamma point, you still have negative frequencies at other BZ points, this means that your system is unstable against some atomic displacements that multiply some of the primitive cell vectors by e.g. 2 or 3, leading to a supercell. Curing this situation may be more tedious. Good luck, Jose C. Conesa Instituto de Catalisis y Petroleoquimica, CSIC Campus de Cantoblanco 28049 Madrid - Spain Phone Nr. 34-91-5854766 Fax Nr. 34-91-5854760 From alberto.milani at chem.polimi.it Thu Oct 6 13:35:43 2005 From: alberto.milani at chem.polimi.it (Alberto Milani) Date: Thu, 6 Oct 2005 13:35:43 +0200 Subject: [Pw_forum] Negative frequencies in phonon calculation Message-ID: <200510061335.43936.alberto.milani@chem.polimi.it> Dear Jose, Thank for your help. I know well the problems about saddle points and negative frequencies because I usually calculate normal modes of molecules where this problem occurs many times. Here the problem is that I am studying an infinite linear chain and so I found strange that some kind of saddle point could be the source of the problem for such a system. Many other ab-initio calculation on polyyne finite chain have shown that these systems are indeed stable as linear chains (no negative frequencies are found). So I would exclude any kind of transversal deformartion from linearity. From this point of view, there is no possibility for any saddle point and the relaxed structure that I have obtained should be considered converged in a well defined minimum. In fact, no negative frequencies are found for the longitudinal phonons. So I think that the problem here is not the usual saddle instability. I wonder if the code fails in some other way in stabilizing the linear chain (e.g. choice of the functional, pseudopotential, ...). Thank you again for your attention. Yours, A. M. From degironc at sissa.it Thu Oct 6 13:40:05 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Thu, 06 Oct 2005 13:40:05 +0200 Subject: [Pw_forum] Negative frequencies in phonon calculation References: <200510061335.43936.alberto.milani@chem.polimi.it> Message-ID: <43450D15.8030706@sissa.it> Is the system optimized also for what concerns the chian lenght ? In an infinite system vanishing forces is not the only condition for equilibrium, also stress (chain tension in your case) should vanish Stefano de Girocoli Alberto Milani wrote: >Dear Jose, >Thank for your help. >I know well the problems about saddle points and negative frequencies because >I usually calculate normal modes of molecules where this problem occurs many >times. >Here the problem is that I am studying an infinite linear chain and so I found >strange that some kind of saddle point could be the source of the problem for >such a system. Many other ab-initio calculation on polyyne finite chain have >shown that these systems are indeed stable as linear chains (no negative >frequencies are found). So I would exclude any kind of transversal >deformartion from linearity. >From this point of view, there is no possibility for any saddle point and the >relaxed structure that I have obtained should be considered converged in a >well defined minimum. In fact, no negative frequencies are found for the >longitudinal phonons. >So I think that the problem here is not the usual saddle instability. I wonder >if the code fails in some other way in stabilizing the linear chain (e.g. >choice of the functional, pseudopotential, ...). >Thank you again for your attention. >Yours, >A. M. >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From alberto.milani at chem.polimi.it Thu Oct 6 14:09:07 2005 From: alberto.milani at chem.polimi.it (Alberto Milani) Date: Thu, 6 Oct 2005 14:09:07 +0200 Subject: [Pw_forum] Negative frequencies in phonon calculation Message-ID: <200510061409.07504.alberto.milani@chem.polimi.it> Dear Stefano, Thank for your advice. Here below I attach the final part of the output where stresses are computed for my relaxed system. It seems that the computation satisfy the 0.5 Kbar threshold on total stress (this threshold has been mentioned somewhere in the forum). Is it possible that for such a system these residual stress are still too large and produce nasty effects on the TA phonons? Thank you yours, A. M. the Fermi energy is -4.5276 ev ! total energy = -22.64532713 ryd estimated scf accuracy < 1.1E-09 ryd band energy sum = -7.45543717 ryd one-electron contribution = -66.15073880 ryd hartree contribution = 33.76544381 ryd xc contribution = -6.79656224 ryd ewald contribution = 16.53653010 ryd convergence has been achieved Forces acting on atoms (Ry/au): atom 1 type 1 force = 0.00000000 0.00000000 0.00000339 atom 2 type 1 force = 0.00000000 0.00000000 -0.00000339 Total force = 0.000005 Total SCF correction = 0.000065 entering subroutine stress ... total stress (ryd/bohr**3) (kbar) P= -0.21 -0.00000081 0.00000000 0.00000000 -0.12 0.00 0.00 0.00000000 -0.00000081 0.00000000 0.00 -0.12 0.00 0.00000000 0.00000000 -0.00000266 0.00 0.00 -0.39 kinetic stress (kbar) 845.28 0.00 0.00 0.00 845.28 0.00 0.00 0.00 573.85 local stress (kbar) 206.31 0.00 0.00 0.00 206.31 -0.00 0.00 -0.00 -7200.10 nonloc. stress (kbar) 251.24 0.00 0.00 0.00 251.24 0.00 0.00 0.00 288.31 hartree stress (kbar) 0.57 -0.00 0.00 -0.00 0.57 -0.00 0.00 -0.00 2866.06 exc-cor stress (kbar) -183.25 -0.00 0.00 -0.00 -183.25 0.00 0.00 0.00 -173.28 corecor stress (kbar) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ewald stress (kbar) -1120.28 0.00 0.00 0.00 -1120.28 0.00 0.00 0.00 3644.76 hubbard stress (kbar) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 bfgs converged in 10 scf cycles and 7 bfgs steps End of BFGS geometry calculation Final energy = -22.6453271292 ryd Saving the approximate inverse hessian CELL_PARAMETERS (alat) 1.000000000 0.000000000 0.000000000 0.000000000 1.000000000 0.000000000 0.000000000 0.000000000 0.256710000 ATOMIC_POSITIONS (angstrom) C 0.000000000 0.000000000 -0.632225902 C 0.000000000 0.000000000 0.632225902 Writing file cc.save for program phonon From giannozz at nest.sns.it Thu Oct 6 14:11:23 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 6 Oct 2005 14:11:23 +0200 Subject: [Pw_forum] Negative frequencies in phonon calculation In-Reply-To: <200510061335.43936.alberto.milani@chem.polimi.it> References: <200510061335.43936.alberto.milani@chem.polimi.it> Message-ID: <200510061411.23276.giannozz@nest.sns.it> On Thursday 06 October 2005 13:35, Alberto Milani wrote: > I wonder if the code fails in some other way in stabilizing the > linear chain (e.g. choice of the functional, pseudopotential, ...). the choice of the functional in linear chains is for sure delicate. Once upon a time I did some calculations in conjugate chains, only to discover that there is no hope to obtain a decent bond alternation with either LDA or GGA. Also forget about decent interchain interactions. Anyway, this is unlikely to be relevant to your problem. You may try two things: - repeat the calculation with norm-conserving pseudopotentials, just to check if there is something wrong in phonon calculations with ultrasoft PPs - make a frozen-phonon calculation for one of the modes that displays a large negative frequency, if possible (i.e. if you do not need a too large supercell) Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From alberto.milani at chem.polimi.it Thu Oct 6 14:29:40 2005 From: alberto.milani at chem.polimi.it (Alberto Milani) Date: Thu, 6 Oct 2005 14:29:40 +0200 Subject: [Pw_forum] Negative frequencies in phonon calculation Message-ID: <200510061429.40874.alberto.milani@chem.polimi.it> Dear Paolo, Thank you very much for your advice. I will try with norm-conserving pseudopotentials and with frozen-phonon calculation. I will soon report my results... Thanks A. M. From felipe at titan.ipicyt.edu.mx Thu Oct 6 15:29:25 2005 From: felipe at titan.ipicyt.edu.mx (FELIPE VALENCIA) Date: Thu, 6 Oct 2005 08:29:25 -0500 (CDT) Subject: [Pw_forum] PBE pseudopotentials for alkali metals. In-Reply-To: <43444BD8.4020601@Princeton.EDU> Message-ID: Dear all, Has anybody got working PBE pseudopotentials for Na, K and Li? I would greatly appreciate if you send them to me! Thanks and regards, Felipe Valencia From vranjan at uark.edu Fri Oct 7 16:18:51 2005 From: vranjan at uark.edu (Vivek Ranjan) Date: Fri, 07 Oct 2005 10:18:51 -0400 Subject: [Pw_forum] problem with neb calculation Message-ID: <1f8c0c81f91b2d.1f91b2d1f8c0c8@uark.edu> Hi, Just curious. Nothing worked till I included " -bmaxdata:512000000" in LDFLAGS, ie. LDFLAGS=-q64 $(LIBS) $(LIBOBJS) -bmaxdata:512000000 works. Thank you, Vivek ----- Original Message ----- From: Eyvaz Isaev Date: Friday, September 30, 2005 4:32 pm Subject: Re: [Pw_forum] problem with neb calculation > Hi, > > It looks like the next message: > > ** ON ENTRY TO ZHPEV PARAMETER NUMBER 1 HAD AN > ILLEGAL VALUE > > There is Paolo's comments on this: > >ESSL contains a version of ZHPEV that is > incompatible > > with ZHPEV provided by LAPACK. If __AIX is defined, > > the former is used. > > Load ESSL before LAPACK. > > Try to follow this advice, i.e. recompile PWSCF with > options -lessl -llapack. > > Bests, > Eyvaz. > > --- Vivek Ranjan wrote: > > > Dear Users, > > > > I ran the attached input file on three different > > machines with : (i) Xeon > > processors, (ii) IBM p690, (iii) another machine > > with IBM p690. I could > > run the programme successfully on the first two, but > > the run on the third > > machine stopped with the following statement : > > > > > > ------------------------------ iteration 1 > > ------------------------ > > tcpu = 0.1; self-consistency for image > > 1 > > ** ON ENTRY TO DSYGV PARAMETER NUMBER 2 HAD AN > > ILLEGAL VALUE > > > > > > > -------------------------------------------------------------------- > > > > Usually everything is similar on the second and the > > third machines > > mentioned above. As a result the same makefile > > works well on both of > > them. I would also like to mention that I have > > successfully performed many > > calculations using PWSCF on all three machines. So, > > my PWSCF compilation > > seems to be alright. > > > > Any hints ? > > > > Thank you, > > > > Vivek > > > > > __________________________________________________________________________ > > web : http://www.geocities.com/vivekranjanweb > > > __________________________________________________________________________ > > Vivek Ranjan Email : > > vranjan at chips.ncsu.edu > > Center for High Performance Simulation : > > vranjan at ncsu.edu > > Department of Physics > > NC State University Phone : > > +1-919-513-0613 (Office) > > Campus Box 7518 Fax : > > +1-919-513-4804 > > Raleigh, NC 27695 > > > -------------------------------------------------------------------- > ------ > > > > > __________________________________ > Yahoo! Mail - PC Magazine Editors' Choice 2005 > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- A non-text attachment was scrubbed... Name: vranjan.vcf Type: text/x-vcard Size: 374 bytes Desc: Card for Vivek Ranjan Url : /pipermail/attachments/20051007/8f4b735b/attachment.vcf From eyvaz_isaev at yahoo.com Fri Oct 7 16:43:45 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 7 Oct 2005 07:43:45 -0700 (PDT) Subject: [Pw_forum] problem with neb calculation In-Reply-To: <1f8c0c81f91b2d.1f91b2d1f8c0c8@uark.edu> Message-ID: <20051007144345.62427.qmail@web60312.mail.yahoo.com> Hi, Yes, so strange. Does it depend on bmaxdata size? Bests, Eyvaz. --- Vivek Ranjan wrote: > Hi, > > Just curious. Nothing worked till I included " > -bmaxdata:512000000" in LDFLAGS, ie. > > LDFLAGS=-q64 $(LIBS) $(LIBOBJS) -bmaxdata:512000000 > > works. > > Thank you, > Vivek > > > ----- Original Message ----- > From: Eyvaz Isaev > > > Date: Friday, September 30, 2005 4:32 pm > Subject: Re: [Pw_forum] problem with neb calculation > > > Hi, > > > > It looks like the next message: > > > > ** ON ENTRY TO ZHPEV PARAMETER NUMBER 1 HAD AN > > ILLEGAL VALUE > > > > There is Paolo's comments on this: > > >ESSL contains a version of ZHPEV that is > > incompatible > > > with ZHPEV provided by LAPACK. If __AIX is > defined, > > > the former is used. > > > Load ESSL before LAPACK. > > > > Try to follow this advice, i.e. recompile PWSCF > with > > options -lessl -llapack. > > > > Bests, > > Eyvaz. > > > > --- Vivek Ranjan wrote: > > > > > Dear Users, > > > > > > I ran the attached input file on three different > > > machines with : (i) Xeon > > > processors, (ii) IBM p690, (iii) another machine > > > with IBM p690. I could > > > run the programme successfully on the first two, > but > > > the run on the third > > > machine stopped with the following statement : > > > > > > > > > ------------------------------ iteration > 1 > > > ------------------------ > > > tcpu = 0.1; self-consistency for > image > > > 1 > > > ** ON ENTRY TO DSYGV PARAMETER NUMBER 2 HAD > AN > > > ILLEGAL VALUE > > > > > > > > > > > > -------------------------------------------------------------------- > > > > > > Usually everything is similar on the second and > the > > > third machines > > > mentioned above. As a result the same makefile > > > works well on both of > > > them. I would also like to mention that I have > > > successfully performed many > > > calculations using PWSCF on all three machines. > So, > > > my PWSCF compilation > > > seems to be alright. > > > > > > Any hints ? > > > > > > Thank you, > > > > > > Vivek > > > > > > > > > __________________________________________________________________________ > > > web : http://www.geocities.com/vivekranjanweb > > > > > > __________________________________________________________________________ > > > Vivek Ranjan Email : > > > vranjan at chips.ncsu.edu > > > Center for High Performance Simulation : > > > vranjan at ncsu.edu > > > Department of Physics > > > NC State University Phone : > > > +1-919-513-0613 (Office) > > > Campus Box 7518 Fax : > > > +1-919-513-4804 > > > Raleigh, NC 27695 > > > > > > -------------------------------------------------------------------- > > ------ > > > > > > > > > > __________________________________ > > Yahoo! Mail - PC Magazine Editors' Choice 2005 > > http://mail.yahoo.com > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > begin:vcard > n:Ranjan;Vivek > fn:Vivek Ranjan > tel;cell:+1-479-236-3673 > tel;fax:+1-479-575-4580 > tel;home:+1-479-571-3159 > tel;work:+1-479-575-5596 > url:www.uark.edu/misc/aaron5/index.html > org:University of Arkansas;Department of Physics > adr:;;226, Physics Building, University of > Arkansas;Fayetteville;AR;72701;USA > version:2.1 > email;internet:vranjan at uark.edu > title:Dr. > end:vcard > > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From proffess at yandex.ru Fri Oct 7 17:06:43 2005 From: proffess at yandex.ru (Sergey Lisenkov) Date: Fri, 7 Oct 2005 19:06:43 +0400 (MSD) Subject: [Pw_forum] problem with neb calculation In-Reply-To: <20051007144345.62427.qmail@web60312.mail.yahoo.com> References: <20051007144345.62427.qmail@web60312.mail.yahoo.com> Message-ID: <43468F03.000005.25318@soapbox.yandex.ru> >Hi, > >Yes, so strange. Does it depend on bmaxdata size? For me it seems to be very strange. On our IBM PPC970 cluster (OS Linux) we cannot use -qbmaxdata option, but I reproduce a problem reported by Vivek. Sergey From g.ballabio at cineca.it Fri Oct 7 17:53:45 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Fri, 7 Oct 2005 17:53:45 +0200 (MEST) Subject: [Pw_forum] problem with neb calculation In-Reply-To: <1f8c0c81f91b2d.1f91b2d1f8c0c8@uark.edu> (from vranjan@uark.edu on Fri Oct 7 16:18:51 2005) References: <1f8c0c81f91b2d.1f91b2d1f8c0c8@uark.edu> Message-ID: <1128700424l.27742l.0l@nb-ballabio.cineca.it> On 10/07/2005 04:18:51 PM, Vivek Ranjan wrote: > Just curious. Nothing worked till I included " > -bmaxdata:512000000" in LDFLAGS, That's curious indeed. The -bmaxdata option is required to go beyond 256 MB of memory if you compile in 32-bit mode, but in 64-bit mode it shouldn't be necessary. Maybe your compiler is configured with a small memory limit for 64-bit mode too? Gerardo From vranjan at uark.edu Fri Oct 7 19:41:58 2005 From: vranjan at uark.edu (Vivek Ranjan) Date: Fri, 07 Oct 2005 13:41:58 -0400 Subject: [Pw_forum] problem with neb calculation Message-ID: <2020bdf2020a3a.2020a3a2020bdf@uark.edu> No. It does not depend on the datasize. Its a very small calculation, first one of the three in example 17. I figured another colleague could run on the same machine and the difference between my and his make.sys was in bmaxdata. ----- Original Message ----- From: Eyvaz Isaev Date: Friday, October 7, 2005 10:43 am Subject: Re: [Pw_forum] problem with neb calculation > Hi, > > Yes, so strange. Does it depend on bmaxdata size? > > Bests, > Eyvaz. > > --- Vivek Ranjan wrote: > > > Hi, > > > > Just curious. Nothing worked till I included " > > -bmaxdata:512000000" in LDFLAGS, ie. > > > > LDFLAGS=-q64 $(LIBS) $(LIBOBJS) -bmaxdata:512000000 > > > > works. > > > > Thank you, > > Vivek > > > > > > ----- Original Message ----- > > From: Eyvaz Isaev > > > > > > Date: Friday, September 30, 2005 4:32 pm > > Subject: Re: [Pw_forum] problem with neb calculation > > > > > Hi, > > > > > > It looks like the next message: > > > > > > ** ON ENTRY TO ZHPEV PARAMETER NUMBER 1 HAD AN > > > ILLEGAL VALUE > > > > > > There is Paolo's comments on this: > > > >ESSL contains a version of ZHPEV that is > > > incompatible > > > > with ZHPEV provided by LAPACK. If __AIX is > > defined, > > > > the former is used. > > > > Load ESSL before LAPACK. > > > > > > Try to follow this advice, i.e. recompile PWSCF > > with > > > options -lessl -llapack. > > > > > > Bests, > > > Eyvaz. > > > > > > --- Vivek Ranjan wrote: > > > > > > > Dear Users, > > > > > > > > I ran the attached input file on three different > > > > machines with : (i) Xeon > > > > processors, (ii) IBM p690, (iii) another machine > > > > with IBM p690. I could > > > > run the programme successfully on the first two, > > but > > > > the run on the third > > > > machine stopped with the following statement : > > > > > > > > > > > > ------------------------------ iteration > > 1 > > > > ------------------------ > > > > tcpu = 0.1; self-consistency for > > image > > > > 1 > > > > ** ON ENTRY TO DSYGV PARAMETER NUMBER 2 HAD > > AN > > > > ILLEGAL VALUE > > > > > > > > > > > > > > > > > > -------------------------------------------------------------------- > > > > > > > > Usually everything is similar on the second and > > the > > > > third machines > > > > mentioned above. As a result the same makefile > > > > works well on both of > > > > them. I would also like to mention that I have > > > > successfully performed many > > > > calculations using PWSCF on all three machines. > > So, > > > > my PWSCF compilation > > > > seems to be alright. > > > > > > > > Any hints ? > > > > > > > > Thank you, > > > > > > > > Vivek > > > > > > > > > > > > > > __________________________________________________________________________ > > > > web : http://www.geocities.com/vivekranjanweb > > > > > > > > > > __________________________________________________________________________ > > > > Vivek Ranjan Email : > > > > vranjan at chips.ncsu.edu > > > > Center for High Performance Simulation : > > > > vranjan at ncsu.edu > > > > Department of Physics > > > > NC State University Phone : > > > > +1-919-513-0613 (Office) > > > > Campus Box 7518 Fax : > > > > +1-919-513-4804 > > > > Raleigh, NC 27695 > > > > > > > > > > -------------------------------------------------------------------- > > > ------ > > > > > > > > > > > > > > > __________________________________ > > > Yahoo! Mail - PC Magazine Editors' Choice 2005 > > > http://mail.yahoo.com > > > _______________________________________________ > > > Pw_forum mailing list > > > Pw_forum at pwscf.org > > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > begin:vcard > > n:Ranjan;Vivek > > fn:Vivek Ranjan > > tel;cell:+1-479-236-3673 > > tel;fax:+1-479-575-4580 > > tel;home:+1-479-571-3159 > > tel;work:+1-479-575-5596 > > url:www.uark.edu/misc/aaron5/index.html > > org:University of Arkansas;Department of Physics > > adr:;;226, Physics Building, University of > > Arkansas;Fayetteville;AR;72701;USA > > version:2.1 > > email;internet:vranjan at uark.edu > > title:Dr. > > end:vcard > > > > > > > > > __________________________________ > Yahoo! Mail - PC Magazine Editors' Choice 2005 > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- A non-text attachment was scrubbed... Name: vranjan.vcf Type: text/x-vcard Size: 374 bytes Desc: Card for Vivek Ranjan Url : /pipermail/attachments/20051007/32f45a58/attachment.vcf From yuwen_66 at yahoo.com Mon Oct 10 05:09:44 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Sun, 9 Oct 2005 20:09:44 -0700 (PDT) Subject: [Pw_forum] questions about kpoints.x tool Message-ID: <20051010030944.85513.qmail@web51006.mail.yahoo.com> Dear PW users, I am experiencing trouble with the k point generation tool kpoints.x. Your kind help with the following question will be much appreciated. 1. Does anybody know why the k points given in the two output files(info and the user-input outfile) are different? 2. It seems this tool only output the k points in the irreducible part of the BZ. Is there a method to get all the k points in the whole BZ? 3. And what's the difference between "write all k?" and not? Many thanks, W. YU __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ From yuwen_66 at yahoo.com Mon Oct 10 05:23:31 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Sun, 9 Oct 2005 20:23:31 -0700 (PDT) Subject: [Pw_forum] questions about kpoints.x tool Message-ID: <20051010032332.91505.qmail@web51004.mail.yahoo.com> I am sorry for asking some silly questions in the former post with the same title. Now I only need to know the anser to question 1. 1. Does anybody know why the k points given in the two output files(info and the user-input outfile) are different? Thanks, W. YU __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ From degironc at sissa.it Mon Oct 10 09:31:36 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Mon, 10 Oct 2005 09:31:36 +0200 Subject: [Pw_forum] questions about kpoints.x tool References: <20051010032332.91505.qmail@web51004.mail.yahoo.com> Message-ID: <434A18D8.3070407@sissa.it> k-points are generate from a specified regular grid in the BZ, removing symmetry-equivalent points. The first output file contains the symmetry-unequivalent k-points and their weights (it's the set to be used in the calculation) while file "info" specifies for each point in the original regular grid to which other point it is equivalent. It's generated for checking purpose. stefano W. YU wrote: >I am sorry for asking some silly questions in the >former post with the same title. Now I only need to >know the anser to question 1. > >1. Does anybody know why the k points given in the two >output files(info and the user-input outfile) are >different? > >Thanks, > >W. YU > > > >__________________________________ >Yahoo! Music Unlimited >Access over 1 million songs. Try it free. >http://music.yahoo.com/unlimited/ >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From adeagbo at thp.Uni-Duisburg.DE Mon Oct 10 13:48:37 2005 From: adeagbo at thp.Uni-Duisburg.DE (Adeagbo Waheed Adeniyi) Date: Mon, 10 Oct 2005 13:48:37 +0200 (MESZ) Subject: [Pw_forum] q2r.x on rhomboheral In-Reply-To: <20051010053509.11236.59320.Mailman@democritos.sissa.it> References: <20051010053509.11236.59320.Mailman@democritos.sissa.it> Message-ID: Dear PWSCF user, I came across the error message below while running q2r.x for rhomboheral structure of Bi after I have completed all the phonon calculations. ====================================================== nqs= 1 q= 0.00000000 0.00000000 0.00000000 reading dyn.mat. from file dyn3 nqs= 6 q= 0.35355340 0.20412410 0.57735030 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from init : error # 1 q not allowed %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... ================================================= I generated 2x2x2 grids of k-points using the k.points.x utililty by suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points 1 0.0000000 0.0000000 0.0000000 1.00 2 0.3535534 -0.2041241 0.2886751 3.00 3 0.3535534 0.2041241 0.5773503 3.00 4 0.0000000 0.0000000 0.8660254 1.00 which I used as q-vectors in the phonon calculation. I will appreciate any help to the problem. Thanks From L.Boeri at fkf.mpg.de Mon Oct 10 13:50:55 2005 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Mon, 10 Oct 2005 13:50:55 +0200 (CEST) Subject: [Pw_forum] charge density and pot file Message-ID: Dear PW-SCF users, I would be interested in obtaining the potential/density file in G-space instead of r-space. Has anybody faced/solved the same problem before? Could anybody indicate me what part of the code I should modify to obtain it? Thanks a lot in advance for your help, Lilia Boeri From degironc at sissa.it Mon Oct 10 14:01:08 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Mon, 10 Oct 2005 14:01:08 +0200 Subject: [Pw_forum] q2r.x on rhomboheral References: <20051010053509.11236.59320.Mailman@democritos.sissa.it> Message-ID: <434A5804.40100@sissa.it> Are you specifying the grid 2x2x2 in the q2r.x input file ? is the value of celldm(3)=1.0 appropriate for your crystal structure ? stefano Adeagbo Waheed Adeniyi wrote: > Dear PWSCF user, > > I came across the error message below while running q2r.x for > rhomboheral structure of Bi after I have completed all the phonon > calculations. > > ====================================================== > nqs= 1 > q= 0.00000000 0.00000000 0.00000000 > reading dyn.mat. from file > dyn3 > > nqs= 6 > q= 0.35355340 0.20412410 0.57735030 > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > from init : error # 1 > q not allowed > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > > stopping ... > > ================================================= > > I generated 2x2x2 grids of k-points using the k.points.x utililty by > suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points > > 1 0.0000000 0.0000000 0.0000000 1.00 > 2 0.3535534 -0.2041241 0.2886751 3.00 > 3 0.3535534 0.2041241 0.5773503 3.00 > 4 0.0000000 0.0000000 0.8660254 1.00 > > > which I used as q-vectors in the phonon calculation. > > I will appreciate any help to the problem. > > > Thanks > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum From yay451 at mail.usask.ca Mon Oct 10 21:13:45 2005 From: yay451 at mail.usask.ca (Yansun Yao) Date: Mon, 10 Oct 2005 13:13:45 -0600 Subject: [Pw_forum] Question on based-centered monoclinic cell input Message-ID: <1128971625.434abd69aba9e@webmail.usask.ca> Dear Pwscf users, I have a question on the input defining a based-centered monoclinic cell (ibrav=13). If I input the relevant information as described in the documentation, I did not get the correct crystal matrix. I inspect the routine latgen.f90 and found that the program produce the fllow lattice which I do not quite understand In unit of alat, the latgen routine generates cos(ab) 0 sin(ab) cos(ab) 0 -sin(ab) b/a c/a 0 For a based cnetered cell, I expect the two vectors on the plane should be 1/2(a+b) and 1/2(a-b)? Is the program expected a different setting? If so, what is it? Any explanation and help is greatly appreciated. Thank you! Yansun University of Saskatechwan From yuwen_66 at yahoo.com Tue Oct 11 04:56:34 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Mon, 10 Oct 2005 19:56:34 -0700 (PDT) Subject: [Pw_forum] meaning of "folds in point" Message-ID: <20051011025634.86606.qmail@web51012.mail.yahoo.com> Dear Stefeno, Thank you for answering my question. One more question, I found there are four columns in the output of kpoints.x. I don't know the meaning of the last column called "folds in point" and are there any use for these numbers? Best regards, W. YU __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ From adeagbo at thp.Uni-Duisburg.DE Tue Oct 11 10:01:02 2005 From: adeagbo at thp.Uni-Duisburg.DE (Adeagbo Waheed Adeniyi) Date: Tue, 11 Oct 2005 10:01:02 +0200 (MESZ) Subject: [Pw_forum] q2r.x on rhomboheral In-Reply-To: <20051011053511.17297.65160.Mailman@democritos.sissa.it> References: <20051011053511.17297.65160.Mailman@democritos.sissa.it> Message-ID: Thank you Dr Stefano, It is the first time I encounter the error when I execute q2r.x in particular for Rhombohedral structure which I am doing for the first time. I didn't have any problem with other systems (cubic/fcc) I have calculated so far. Whenever I supplied any value for celldim(3) for Rhombohedral (ibrav=5) I always get the following when I execute kpoint.x for 2x2x2, ======================== 4 1 0.0000000 0.0000000 0.0000000 1.00 2 0.3535534 -0.2041241 0.2886751 3.00 3 0.3535534 0.2041241 0.5773503 3.00 4 0.0000000 0.0000000 0.8660254 1.00 =========================================== These are the set of values I used as q-vectors in the calculation of phonons. May be my input file is wrong. I have no idea =============== &CONTROL calculation = 'phonon' , restart_mode = 'from_scratch' , outdir = 'tmp/' , pseudo_dir = '/home4/hes11/hes110/espresso/pseudo/' , prefix = 'bi' , tstress = .true. , tprnfor = .true. , / &SYSTEM ibrav = 5, celldm(1) = 8.9263,celldm(4)=0.53954 nat = 2, ntyp = 1, ecutwfc = 80.0 , ecutrho = 360.0 , occupations='smearing' degauss = 0.02, smearing = 'fermi-dirac' nspin = 1, starting_magnetization(1) = 0.0 / &ELECTRONS electron_maxstep = 60, conv_thr = 1.0D-7 , startingpot = 'atomic' , startingwfc = 'random' , mixing_mode = 'plain' , diagonalization = 'david' , / &phonon xqq(1) = 0.3535534, xqq(2) =-0.2041241, xqq(3) =0.2886751 / ATOMIC_SPECIES Bi 208.98038 Bi.pbe_nc.UPF ATOMIC_POSITIONS crystal Bi 0.249929420 0.249929420 0.249929420 Bi 0.750070580 0.750070580 0.750070580 K_POINTS automatic 16 16 16 1 1 1 ================================================================ > Message: 2 > Date: Mon, 10 Oct 2005 13:48:37 +0200 (MESZ) > From: Adeagbo Waheed Adeniyi > To: pw_forum at pwscf.org > Subject: [Pw_forum] q2r.x on rhomboheral > Reply-To: pw_forum at pwscf.org > > Dear PWSCF user, > > I came across the error message below while running q2r.x for > rhomboheral structure of Bi after I have completed all the phonon > calculations. > > ====================================================== > nqs= 1 > q= 0.00000000 0.00000000 0.00000000 > reading dyn.mat. from file > dyn3 > > nqs= 6 > q= 0.35355340 0.20412410 0.57735030 > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from init : error # 1 > q not allowed > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > stopping ... > > ================================================= > > I generated 2x2x2 grids of k-points using the k.points.x utililty by > suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points > > 1 0.0000000 0.0000000 0.0000000 1.00 > 2 0.3535534 -0.2041241 0.2886751 3.00 > 3 0.3535534 0.2041241 0.5773503 3.00 > 4 0.0000000 0.0000000 0.8660254 1.00 > > > which I used as q-vectors in the phonon calculation. > > I will appreciate any help to the problem. > > > Thanks > --__--__-- > > Message: 4 > Date: Mon, 10 Oct 2005 14:01:08 +0200 > From: Stefano de Gironcoli > To: pw_forum at pwscf.org > Subject: Re: [Pw_forum] q2r.x on rhomboheral > Reply-To: pw_forum at pwscf.org > > Are you specifying the grid 2x2x2 in the q2r.x input file ? > is the value of celldm(3)=1.0 appropriate for your crystal structure ? > stefano > > Adeagbo Waheed Adeniyi wrote: > >> Dear PWSCF user, >> >> I came across the error message below while running q2r.x for >> rhomboheral structure of Bi after I have completed all the phonon >> calculations. >> >> ====================================================== >> nqs= 1 >> q= 0.00000000 0.00000000 0.00000000 >> reading dyn.mat. from file >> dyn3 >> >> nqs= 6 >> q= 0.35355340 0.20412410 0.57735030 >> >> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >> >> from init : error # 1 >> q not allowed >> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >> >> >> stopping ... >> >> ================================================= >> >> I generated 2x2x2 grids of k-points using the k.points.x utililty by >> suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points >> >> 1 0.0000000 0.0000000 0.0000000 1.00 >> 2 0.3535534 -0.2041241 0.2886751 3.00 >> 3 0.3535534 0.2041241 0.5773503 3.00 >> 4 0.0000000 0.0000000 0.8660254 1.00 >> >> >> which I used as q-vectors in the phonon calculation. >> >> I will appreciate any help to the problem. >> >> >> Thanks >> >> From degironc at sissa.it Tue Oct 11 10:57:18 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Tue, 11 Oct 2005 10:57:18 +0200 Subject: [Pw_forum] meaning of "folds in point" References: <20051011025634.86606.qmail@web51012.mail.yahoo.com> Message-ID: <434B7E6E.6090806@sissa.it> in the output the first column is an integer numbering the points, when it is said that point 33 folds in points 7 it means that point 33 is equivalent for symmetry reasons to point 7 so that one does not need to compute it but can rather obtain its contribution to band energy, density etc from the results of points 7 (with appropriate weight). This is just the way special points operate... in case of doubts please refere to some original literature. A. Baldereschi, Phys. Rev. B 7, 5212 (1973). D.J. Chadi e M.L. Cohen, Phys. Rev.B 8, 5747 (1973). H.J. Monkhorst e J.D. Pack, Phys. Rev. B 13, 5188 (1976). best, stefano W. YU wrote: >Dear Stefeno, > >Thank you for answering my question. > >One more question, I found there are four columns in >the output of kpoints.x. I don't know the meaning of >the last column called "folds in point" and are there >any use for these numbers? > >Best regards, > >W. YU > > > > >__________________________________ >Yahoo! Music Unlimited >Access over 1 million songs. Try it free. >http://music.yahoo.com/unlimited/ >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From degironc at sissa.it Tue Oct 11 11:06:40 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Tue, 11 Oct 2005 11:06:40 +0200 Subject: [Pw_forum] q2r.x on rhomboheral References: <20051011053511.17297.65160.Mailman@democritos.sissa.it> Message-ID: <434B80A0.4090803@sissa.it> Adeagbo Waheed Adeniyi wrote: > Thank you Dr Stefano, > It is the first time I encounter the error when I execute q2r.x in > particular for Rhombohedral structure which I am doing for the first > time. > I didn't have any problem with other systems (cubic/fcc) I have > calculated so far. > > Whenever I supplied any value for celldim(3) for Rhombohedral (ibrav=5) I guess the problem is that ibrav=5 requires the value of celldm(4) and not celldm(3) moreover the supplied value should be the same as the one used in the scf calculation. stefano > > I always get the following when I execute kpoint.x for 2x2x2, > ======================== > 4 > 1 0.0000000 0.0000000 0.0000000 1.00 > 2 0.3535534 -0.2041241 0.2886751 3.00 > 3 0.3535534 0.2041241 0.5773503 3.00 > 4 0.0000000 0.0000000 0.8660254 1.00 > =========================================== > These are the set of values I used as q-vectors in the calculation of > phonons. > May be my input file is wrong. I have no idea > =============== > &CONTROL > calculation = 'phonon' , > restart_mode = 'from_scratch' , > outdir = 'tmp/' , > pseudo_dir = '/home4/hes11/hes110/espresso/pseudo/' > , > prefix = 'bi' , > tstress = .true. , > tprnfor = .true. , > / > &SYSTEM > ibrav = 5, > celldm(1) = 8.9263,celldm(4)=0.53954 > nat = 2, > ntyp = 1, > ecutwfc = 80.0 , > ecutrho = 360.0 , > occupations='smearing' > degauss = 0.02, > smearing = 'fermi-dirac' > nspin = 1, > starting_magnetization(1) = 0.0 > / > &ELECTRONS > electron_maxstep = 60, > conv_thr = 1.0D-7 , > startingpot = 'atomic' , > startingwfc = 'random' , > mixing_mode = 'plain' , > diagonalization = 'david' , > / > &phonon > xqq(1) = 0.3535534, xqq(2) =-0.2041241, xqq(3) =0.2886751 > / > ATOMIC_SPECIES > Bi 208.98038 Bi.pbe_nc.UPF > ATOMIC_POSITIONS crystal > Bi 0.249929420 0.249929420 0.249929420 > Bi 0.750070580 0.750070580 0.750070580 > K_POINTS automatic > 16 16 16 1 1 1 > ================================================================ > From giannozz at nest.sns.it Tue Oct 11 12:26:04 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 11 Oct 2005 12:26:04 +0200 Subject: [Pw_forum] Question on based-centered monoclinic cell input In-Reply-To: <1128971625.434abd69aba9e@webmail.usask.ca> References: <1128971625.434abd69aba9e@webmail.usask.ca> Message-ID: <200510111226.04603.giannozz@nest.sns.it> On Monday 10 October 2005 21:13, Yansun Yao wrote: > I have a question on the input defining a based-centered monoclinic > cell (ibrav=13) [..] In unit of alat, the latgen routine generates > > cos(ab) 0 sin(ab) > cos(ab) 0 -sin(ab) > b/a c/a 0 > For a based centered cell, I expect the two vectors on the > plane should be 1/2(a+b) and 1/2(a-b)? > Is the program expected a different setting? If so, what > is it? the primitive lattice vectors used by the code are explained at the end of file Doc/INPUT_PW. The primitive lattice vectors for the base-centered monoclinic case look actually quite weird. Maybe nobody ever used them. If you have a clear idea of what they should look like, please let us know. Note that you can input your lattice vectors by specifying ibrav=0 and a card CELL_PARAMETERS Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From afloris at physik.fu-berlin.de Tue Oct 11 16:52:53 2005 From: afloris at physik.fu-berlin.de (Andrea Floris) Date: Tue, 11 Oct 2005 16:52:53 +0200 (CEST) Subject: [Pw_forum] projwfc.f90 In-Reply-To: <20051010053509.11236.59320.Mailman@democritos.sissa.it> References: <20051010053509.11236.59320.Mailman@democritos.sissa.it> Message-ID: Dear all, I'd like to have a clearer (or a less confused) idea about what EXACTLY projwfc.f90 does (how exactly the projections are done, necessity of orthogonalize the atomic wf present in the pseudopotentials, overcompletness of this atomic "basis", if the projection of psi(n,k) is done onto some Block superposition atomic states phi(atom,l,m) at that k.. ecc ecc). Relative to this, I ask: does it exists some documentation (papers, some notes or whatever) on which the program projwfc.f90 (and the routines that it calls) is based? I tried to go through the code, but in that part is really not the best way to understand what's going on.. Thank you , Andrea From giannozz at nest.sns.it Tue Oct 11 17:07:15 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 11 Oct 2005 17:07:15 +0200 Subject: [Pw_forum] charge density and pot file In-Reply-To: References: Message-ID: <200510111707.16019.giannozz@nest.sns.it> On Monday 10 October 2005 13:50, Lilia Boeri wrote: > I would be interested in obtaining the potential/density > file in G-space instead of r-space. Has anybody faced/solved > the same problem before? it is not really a new problem: converting from G- to r-space is done all the time in the code. You need to know how data is stored and organized, though, or else you may adopt the "monkey technique": find a piece of code that does what you need and copy it. A good place to look at and to modify could be routine PP/chdens.f90, where you do not need to consider the k=0 and the parallel case. Basically you need to copy the (real) charge density in real space to a complex array, make the fft, then re-order the array: psic(:) = DCMPLX (rhor(:), 0.d0) call cft3 (psic, nr1, nr2, nr3, nrx1, nrx2, nrx3, - 1) allocate (rhog( ngm)) rhog (:) = psic (nl (:) ) rhog(i), i=1, ngm, should contain the Fourier component for G = g(:,i) (2pi/a units) whose norm is gg(i) ( (2pi/a)^2 units) Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Tue Oct 11 17:33:46 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 11 Oct 2005 17:33:46 +0200 Subject: [Pw_forum] projwfc.f90 In-Reply-To: References: <20051010053509.11236.59320.Mailman@democritos.sissa.it> Message-ID: <200510111733.47027.giannozz@nest.sns.it> On Tuesday 11 October 2005 16:52, Andrea Floris wrote: > does it exists some documentation (papers, some notes or whatever) > on which the program projwfc.f90 (and the routines that it calls) is > based? some time ago I wrote some notes (latex file below) about Loewdin charges. Hope this helps Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy -- \documentclass[a4paper,12pt]{article} \begin{document} \section{L\"owdin and Mulliken population analysis} We have $| \phi_\mu\rangle$ atomic (or atomic-like) orbitals, not necessarily orthonormal, and we want to write KS orbitals for our system: $|\psi_\alpha\rangle$, as sums over said atomic orbitals: $$ |\psi_\alpha\rangle = \sum_\mu c_\mu^{(\alpha)} | \phi_\mu \rangle. $$ The generalized orthonormality relations for KS orbitals is written as: $$ \langle\psi_\alpha| \hat S |\psi_\beta\rangle = \delta_{\alpha\beta}. $$ where $\hat S$ is an operator defined in the US PP framework. The charge density $\rho({\bf r})$ is given by $$ \rho({\bf r}) = \sum_{\mu,\nu} P_{\mu\nu} \left (\phi^*_\nu \hat S \phi_\mu\right) ({\bf r}) $$ where $$ P_{\mu\nu} = \sum_\alpha c_\mu^{(\alpha)} c_\nu^{(\alpha)*} $$ define an operator $\hat P$, and $$ \left (\phi^*_\nu \hat S \phi_\mu\right) ({\bf r}) = \phi^*_\nu({\bf r}) \phi_\mu({\bf r}) + \sum_{lm} \langle\phi_\nu|\beta_l\rangle q_{lm}({\bf r}) \langle\beta_m|\phi_\mu\rangle $$ where the $\beta$'s and $q$'s are components of the US PP. \subsection{Mulliken population analysis} We write the total number of electrons $N$ as: $$ N = \int \rho({\bf r}) d{\bf r} = \mbox{Tr} \hat P\hat S = \sum_\mu (\sum_\nu P_{\mu\nu} S_{\nu\mu}) \equiv \sum_\mu q_\mu $$ where $q_\mu$ is the {\em Mulliken charge} associated to state $\mu$ : $$ q_\mu = \sum_\nu \sum_\alpha c_\mu^{(\alpha)} c_\nu^{(\alpha)*} S_{\nu\mu} $$ and $$ S_{\nu\mu} = \langle\phi_\nu|\hat S|\phi_\mu\rangle = \int \phi^*_\nu({\bf r}) \phi_\mu({\bf r}) d{\bf r} + \sum_{lm} \langle\phi_\nu|\beta_l\rangle Q_{lm} \langle\beta_m|\phi_\mu\rangle $$ with $$ Q_{lm}=\int q_{lm}({\bf r}) d{\bf r}. $$ In general, this matrix is not diagonal, even with NC PP. The coefficients $c_\mu^{(\alpha)}$ are obtained by inverting the linear system: $$ \langle \phi_\nu | \hat S | \psi_\alpha\rangle = \sum_\mu c_\mu^{(\alpha)} S_{\nu\mu} $$ that is $$ c_\mu^{(\alpha)} = (\hat S^{-1})_{\mu\nu} \langle \phi_\nu | \hat S | \psi_\alpha\rangle $$ and finally $$ q_\mu = \sum_{\alpha\nu} \langle \psi_\alpha | \hat S | \phi_\mu\rangle (\hat S^{-1})_{\mu\nu} \langle \phi_\nu | \hat S | \psi_\alpha\rangle. $$ \subsection{L\"owdin population analysis} The total number of electrons $N$ can be alternatively written as $$ N = \mbox{Tr} \left[\hat S^{1/2}\hat P \hat S^{1/2}\right] = \sum_\mu \tilde q_\mu. $$ where $\tilde q_\mu$ is called {\em L\"owdin charge} associated to state $\mu$. Let us introduce an auxiliary set of atomic orbitals $\tilde\phi$ via : $$ |\tilde\phi_\mu\rangle = \sum_\nu (\hat S^{-1/2})_{\nu\mu}| \phi_\nu\rangle, \qquad | \phi_\mu\rangle = \sum_\nu (\hat S^{1/2})_{\nu\mu}|\tilde\phi_\nu\rangle, $$ for which the generalized orthonormality relation $$ \langle\tilde\phi_\mu|\hat S| \tilde\phi_\nu\rangle = \delta_{\mu\nu} $$ holds. The KS orbitals for our systems can be rewritten as $$ |\psi_\alpha\rangle = \sum_\mu c_\mu^{(\alpha)} | \phi_\mu \rangle = \sum_\mu \tilde c_\mu^{(\alpha)} | \tilde\phi_\mu \rangle $$ where $$ \tilde c_\mu^{(\alpha)} = \sum_\nu (\hat S^{1/2})_{\nu\mu} c_\nu^{(\alpha)}. $$ By comparison with the above expression of $\tilde q_\mu$ we get $$ \tilde q_\mu = \sum_\alpha |\tilde c_\mu^{(\alpha)}|^2 = |\langle \tilde\phi_\mu | \hat S | \psi_\alpha\rangle|^2 $$ Reference: {\em Modern Quantum Chemistry}, A. Szabo and N. Ostlund (Dover, NY 1996), p. 153 \end{document} From L.Boeri at fkf.mpg.de Tue Oct 11 17:41:01 2005 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Tue, 11 Oct 2005 17:41:01 +0200 (CEST) Subject: [Pw_forum] charge density and pot file In-Reply-To: <200510111707.16019.giannozz@nest.sns.it> References: <200510111707.16019.giannozz@nest.sns.it> Message-ID: dear Paolo, thanks a lot for your response. What confused me a little bit was actually the format in which the variables are stored. In particular, I have problems in locating which part of the code attaches and index "ig" to a particular plane wave: i.e. how is the index i related to ix, iy,iz? thanks a lot in advance, Lilia On Tue, 11 Oct 2005, Paolo Giannozzi wrote: > On Monday 10 October 2005 13:50, Lilia Boeri wrote: > > > I would be interested in obtaining the potential/density > > file in G-space instead of r-space. Has anybody faced/solved > > the same problem before? > > it is not really a new problem: converting from G- to r-space is > done all the time in the code. You need to know how data is > stored and organized, though, or else you may adopt the "monkey > technique": find a piece of code that does what you need and > copy it. A good place to look at and to modify could be routine > PP/chdens.f90, where you do not need to consider the k=0 > and the parallel case. Basically you need to copy the (real) > charge density in real space to a complex array, make the fft, > then re-order the array: > > psic(:) = DCMPLX (rhor(:), 0.d0) > call cft3 (psic, nr1, nr2, nr3, nrx1, nrx2, nrx3, - 1) > allocate (rhog( ngm)) > rhog (:) = psic (nl (:) ) > > rhog(i), i=1, ngm, should contain the Fourier component for > G = g(:,i) (2pi/a units) whose norm is gg(i) ( (2pi/a)^2 units) > > Paolo > > -- > Paolo Giannozzi e-mail: giannozz at nest.sns.it > Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 > Piazza dei Cavalieri 7 I-56126 Pisa, Italy > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From giannozz at nest.sns.it Tue Oct 11 19:00:10 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 11 Oct 2005 19:00:10 +0200 Subject: [Pw_forum] charge density and pot file In-Reply-To: References: <200510111707.16019.giannozz@nest.sns.it> Message-ID: <200510111900.10851.giannozz@nest.sns.it> On Tuesday 11 October 2005 17:41, Lilia Boeri wrote: > In particular, I have problems in locating which part of the code > attaches an index "ig" to a particular plane wave: i.e. how is the > index i related to ix, iy,iz? it depends on what ix, iy, iz are ... There are two different ways of storing G-components. One is as list of components for wavevector g_i (i=1,ngm): g_i = ( G(1,i), G(2,i), G(3,i) ) * 2pi/a, where variable G is contained in module gvect. The other way is the "FFT grid" that is used to perform FFT. The latter is related to the former by a set of indices NL: rho(g_i) = rho_fft(nl(i)) Unless you need to do something special, this should be more or less all you need Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From konstantin_kudin at yahoo.com Wed Oct 12 01:18:56 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Tue, 11 Oct 2005 16:18:56 -0700 (PDT) Subject: [Pw_forum] Question on stress in a system with constraints In-Reply-To: <200510111900.10851.giannozz@nest.sns.it> Message-ID: <20051011231856.75295.qmail@web52004.mail.yahoo.com> Hi there, I have a basic question on the stress tensor. Presumably, the stress tensor computed in the CP implies continuous stretching of the underlying system. Is this correct? Is the derivative with respect to % of stretch, or is it to extra length in Bohrs, for example? What are the units of the stress in CP ? If one projected out certain forces (for constrained coordinates), then the stress needs to be corrected for the fact that certain forces are no longer there. Basically, the original stress was computed as if all the forces were present, now, however, some of them are missing. This means that if the system were to be stretched, the constrained coordinates would still be constrained. My question is how I can correct the stress in a system with constraints if this needs to be done. Thanks! Kostya __________________________________ Start your day with Yahoo! - Make it your home page! http://www.yahoo.com/r/hs From marzari at MIT.EDU Wed Oct 12 02:48:14 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Tue, 11 Oct 2005 20:48:14 -0400 Subject: [Pw_forum] Question on stress in a system with constraints In-Reply-To: <20051011231856.75295.qmail@web52004.mail.yahoo.com> References: <20051011231856.75295.qmail@web52004.mail.yahoo.com> Message-ID: <434C5D4E.9030706@mit.edu> Hi Kostya, a quick comment - Don Hamann has written a fairly extensive PRB this year discussing many of these issues: Vol 72, 350105 (2005). CP stresses are (I hope) a derivative with respect to the strain tensor - ie.e they do not take into account that atoms can relax in response to the stress (the paradigmatic case is the response to a strain in the 111 direction in silicon - the internal strain parameter measures how the distance between the two atoms in the unit cell changes in response to the symmetry-breaking stress). So, you have the bare stress, calculated by CP and/or PWSCF, but you want the renormalized one "dressed" by the relaxations (mediated by the inverse of the dynamical matrix, and by the coupling between displacements and strains). The constraint will allow you to renormalize appropriately the bare stress, if you have all the response functions from DFPT (and their correct long-wavelength limit) but it might be easier to do it by finite differeces of the energy along the strain direction, while constraining the atoms. By the way - the dressing of a perturbation by the ionic relaxations is very relevant for piezoelectricity (e.g. Stefano de Gironcoli 1989 PRL) or for the interactions in substitutional alloy (PRL 72 4001 (1994) - some of the issues with the long wavelength limit are discussed there). Best, nicola Konstantin Kudin wrote: > Hi there, > > I have a basic question on the stress tensor. > > Presumably, the stress tensor computed in the CP implies continuous > stretching of the underlying system. Is this correct? Is the derivative > with respect to % of stretch, or is it to extra length in Bohrs, for > example? > What are the units of the stress in CP ? > > If one projected out certain forces (for constrained coordinates), > then the stress needs to be corrected for the fact that certain forces > are no longer there. Basically, the original stress was computed as if > all the forces were present, now, however, some of them are missing. > This means that if the system were to be stretched, the constrained > coordinates would still be constrained. > > My question is how I can correct the stress in a system with > constraints if this needs to be done. > > Thanks! > Kostya > > > > __________________________________ > Start your day with Yahoo! - Make it your home page! > http://www.yahoo.com/r/hs > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From giannozz at nest.sns.it Wed Oct 12 09:51:50 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 12 Oct 2005 09:51:50 +0200 Subject: [Pw_forum] Question on stress in a system with constraints In-Reply-To: <20051011231856.75295.qmail@web52004.mail.yahoo.com> References: <20051011231856.75295.qmail@web52004.mail.yahoo.com> Message-ID: <200510120951.50518.giannozz@nest.sns.it> On Wednesday 12 October 2005 01:18, Konstantin Kudin wrote: > What are the units of the stress in CP ? internal units used to be Ry/bohr^3, but I think that in recent versions of the codethey are Hartree/bohr^3 ; what is printed is GPa Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From adeagbo at thp.Uni-Duisburg.DE Wed Oct 12 11:54:47 2005 From: adeagbo at thp.Uni-Duisburg.DE (Adeagbo Waheed Adeniyi) Date: Wed, 12 Oct 2005 11:54:47 +0200 (MESZ) Subject: [Pw_forum] q2r.x on rhomboheral In-Reply-To: References: <20051011053511.17297.65160.Mailman@democritos.sissa.it> Message-ID: Thank you Dr. Stefano. I will try it again. But (ibrav=5) always ask for celldim(3) NOT celldim(4). I will try celldim(4) instead celldim(3) by ignoring celldim(3) it asks for. Thanks. Adeagbo Adeagbo Waheed Adeniyi wrote: > Thank you Dr Stefano, > It is the first time I encounter the error when I execute q2r.x in > particular for Rhombohedral structure which I am doing for the first > time. > I didn't have any problem with other systems (cubic/fcc) I have > calculated so far. > > Whenever I supplied any value for celldim(3) for Rhombohedral (ibrav=5) I guess the problem is that ibrav=5 requires the value of celldm(4) and not celldm(3) moreover the supplied value should be the same as the one used in the scf calculation. stefano On Tue, 11 Oct 2005, Adeagbo Waheed Adeniyi wrote: > Thank you Dr Stefano, > It is the first time I encounter the error when I execute q2r.x in > particular for Rhombohedral structure which I am doing for the first > time. > I didn't have any problem with other systems (cubic/fcc) I have calculated so > far. > > Whenever I supplied any value for celldim(3) for Rhombohedral (ibrav=5) > I always get the following when I execute kpoint.x for 2x2x2, > ======================== > 4 > 1 0.0000000 0.0000000 0.0000000 1.00 > 2 0.3535534 -0.2041241 0.2886751 3.00 > 3 0.3535534 0.2041241 0.5773503 3.00 > 4 0.0000000 0.0000000 0.8660254 1.00 > =========================================== > These are the set of values I used as q-vectors in the calculation of > phonons. > May be my input file is wrong. I have no idea > =============== > &CONTROL > calculation = 'phonon' , > restart_mode = 'from_scratch' , > outdir = 'tmp/' , > pseudo_dir = '/home4/hes11/hes110/espresso/pseudo/' > , > prefix = 'bi' , > tstress = .true. , > tprnfor = .true. , > / > &SYSTEM > ibrav = 5, > celldm(1) = 8.9263,celldm(4)=0.53954 > nat = 2, > ntyp = 1, > ecutwfc = 80.0 , > ecutrho = 360.0 , > occupations='smearing' > degauss = 0.02, > smearing = 'fermi-dirac' > nspin = 1, > starting_magnetization(1) = 0.0 > / > &ELECTRONS > electron_maxstep = 60, > conv_thr = 1.0D-7 , > startingpot = 'atomic' , > startingwfc = 'random' , > mixing_mode = 'plain' , > diagonalization = 'david' , > / > &phonon > xqq(1) = 0.3535534, xqq(2) =-0.2041241, xqq(3) =0.2886751 > / > ATOMIC_SPECIES > Bi 208.98038 Bi.pbe_nc.UPF > ATOMIC_POSITIONS crystal > Bi 0.249929420 0.249929420 0.249929420 > Bi 0.750070580 0.750070580 0.750070580 > K_POINTS automatic > 16 16 16 1 1 1 > ================================================================ > >> Message: 2 >> Date: Mon, 10 Oct 2005 13:48:37 +0200 (MESZ) >> From: Adeagbo Waheed Adeniyi >> To: pw_forum at pwscf.org >> Subject: [Pw_forum] q2r.x on rhomboheral >> Reply-To: pw_forum at pwscf.org >> >> Dear PWSCF user, >> >> I came across the error message below while running q2r.x for >> rhomboheral structure of Bi after I have completed all the phonon >> calculations. >> >> ====================================================== >> nqs= 1 >> q= 0.00000000 0.00000000 0.00000000 >> reading dyn.mat. from file >> dyn3 >> >> nqs= 6 >> q= 0.35355340 0.20412410 0.57735030 >> >> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >> from init : error # 1 >> q not allowed >> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >> >> stopping ... >> >> ================================================= >> >> I generated 2x2x2 grids of k-points using the k.points.x utililty by >> suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points >> >> 1 0.0000000 0.0000000 0.0000000 1.00 >> 2 0.3535534 -0.2041241 0.2886751 3.00 >> 3 0.3535534 0.2041241 0.5773503 3.00 >> 4 0.0000000 0.0000000 0.8660254 1.00 >> >> >> which I used as q-vectors in the phonon calculation. >> >> I will appreciate any help to the problem. >> >> >> Thanks > >> --__--__-- >> >> Message: 4 >> Date: Mon, 10 Oct 2005 14:01:08 +0200 >> From: Stefano de Gironcoli >> To: pw_forum at pwscf.org >> Subject: Re: [Pw_forum] q2r.x on rhomboheral >> Reply-To: pw_forum at pwscf.org >> >> Are you specifying the grid 2x2x2 in the q2r.x input file ? >> is the value of celldm(3)=1.0 appropriate for your crystal structure ? >> stefano >> >> Adeagbo Waheed Adeniyi wrote: >> >>> Dear PWSCF user, >>> >>> I came across the error message below while running q2r.x for >>> rhomboheral structure of Bi after I have completed all the phonon >>> calculations. >>> >>> ====================================================== >>> nqs= 1 >>> q= 0.00000000 0.00000000 0.00000000 >>> reading dyn.mat. from file >>> dyn3 >>> >>> nqs= 6 >>> q= 0.35355340 0.20412410 0.57735030 >>> >>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >>> >>> from init : error # 1 >>> q not allowed >>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >>> >>> >>> stopping ... >>> >>> ================================================= >>> >>> I generated 2x2x2 grids of k-points using the k.points.x utililty by >>> suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points >>> >>> 1 0.0000000 0.0000000 0.0000000 1.00 >>> 2 0.3535534 -0.2041241 0.2886751 3.00 >>> 3 0.3535534 0.2041241 0.5773503 3.00 >>> 4 0.0000000 0.0000000 0.8660254 1.00 >>> >>> >>> which I used as q-vectors in the phonon calculation. >>> >>> I will appreciate any help to the problem. >>> >>> >>> Thanks >>> >>> > From adeagbo at thp.Uni-Duisburg.DE Wed Oct 12 12:03:45 2005 From: adeagbo at thp.Uni-Duisburg.DE (Adeagbo Waheed Adeniyi) Date: Wed, 12 Oct 2005 12:03:45 +0200 (MESZ) Subject: [Pw_forum] q2r.x on rhomboheral In-Reply-To: References: <20051011053511.17297.65160.Mailman@democritos.sissa.it> Message-ID: Thank you Dr. Stefano. I will try it again. But (ibrav=5) always ask for celldim(3) NOT celldim(4). I will try celldim(4) instead celldim(3) by ignoring celldim(3) it asks for. Thanks. It doesn't matter any value one supplies as celldim(3) or celldim(4), these sets of values are always produced for 2 2 2 4 1 0.0000000 0.0000000 0.0000000 1.00 2 0.3535534 -0.2041241 0.2886751 3.00 3 0.3535534 0.2041241 0.5773503 3.00 4 0.0000000 0.0000000 0.8660254 1.00 *************************************************** * * * Welcome to the special points world! * *_________________________________________________* * 1 = cubic p (sc ) 8 = orthor p (so ) * * 2 = cubic f (fcc) 9 = orthor 1-(fco) * * 3 = cubic i (bcc) 10 = orthor f (fco) * * 4 = hex & trig p 11 = orthor i (bco) * * 5 = trigonal r 12 = monoclinic p * * 6 = tetrag p (st ) 13 = monocl 1-(fcm) * * 7 = tetrag i (bct) 14 = triclinic p * *************************************************** bravais lattice >> 5 filout [mesh_k] >> k222tri enter celldm(3) >> 0.53954 mesh: n1 n2 n3 >> 2 2 2 mesh: k1 k2 k3 (0 no shift, 1 shifted) >> 0 0 0 write all k? [f] >> # of k-points == 4 of 8 > Adeagbo > > > Adeagbo Waheed Adeniyi wrote: > > >> Thank you Dr Stefano, > >> It is the first time I encounter the error when I execute q2r.x in > >> particular for Rhombohedral structure which I am doing for the first > >> time. > >> I didn't have any problem with other systems (cubic/fcc) I have > >> calculated so far. > >> > >> Whenever I supplied any value for celldim(3) for Rhombohedral > (ibrav=5) > > I guess the problem is that ibrav=5 requires the value of celldm(4) > and > not celldm(3) > > moreover the supplied value should be the same as the one used in the > > scf calculation. > > > stefano > > > On Tue, 11 Oct 2005, Adeagbo Waheed Adeniyi wrote: > >> Thank you Dr Stefano, >> It is the first time I encounter the error when I execute q2r.x in >> particular for Rhombohedral structure which I am doing for the first >> time. >> I didn't have any problem with other systems (cubic/fcc) I have calculated >> so far. >> >> Whenever I supplied any value for celldim(3) for Rhombohedral (ibrav=5) >> I always get the following when I execute kpoint.x for 2x2x2, >> ======================== >> 4 >> 1 0.0000000 0.0000000 0.0000000 1.00 >> 2 0.3535534 -0.2041241 0.2886751 3.00 >> 3 0.3535534 0.2041241 0.5773503 3.00 >> 4 0.0000000 0.0000000 0.8660254 1.00 >> =========================================== >> These are the set of values I used as q-vectors in the calculation of >> phonons. >> May be my input file is wrong. I have no idea >> =============== >> &CONTROL >> calculation = 'phonon' , >> restart_mode = 'from_scratch' , >> outdir = 'tmp/' , >> pseudo_dir = '/home4/hes11/hes110/espresso/pseudo/' >> , >> prefix = 'bi' , >> tstress = .true. , >> tprnfor = .true. , >> / >> &SYSTEM >> ibrav = 5, >> celldm(1) = 8.9263,celldm(4)=0.53954 >> nat = 2, >> ntyp = 1, >> ecutwfc = 80.0 , >> ecutrho = 360.0 , >> occupations='smearing' >> degauss = 0.02, >> smearing = 'fermi-dirac' >> nspin = 1, >> starting_magnetization(1) = 0.0 >> / >> &ELECTRONS >> electron_maxstep = 60, >> conv_thr = 1.0D-7 , >> startingpot = 'atomic' , >> startingwfc = 'random' , >> mixing_mode = 'plain' , >> diagonalization = 'david' , >> / >> &phonon >> xqq(1) = 0.3535534, xqq(2) =-0.2041241, xqq(3) =0.2886751 >> / >> ATOMIC_SPECIES >> Bi 208.98038 Bi.pbe_nc.UPF >> ATOMIC_POSITIONS crystal >> Bi 0.249929420 0.249929420 0.249929420 >> Bi 0.750070580 0.750070580 0.750070580 >> K_POINTS automatic >> 16 16 16 1 1 1 >> ================================================================ >> >>> Message: 2 >>> Date: Mon, 10 Oct 2005 13:48:37 +0200 (MESZ) >>> From: Adeagbo Waheed Adeniyi >>> To: pw_forum at pwscf.org >>> Subject: [Pw_forum] q2r.x on rhomboheral >>> Reply-To: pw_forum at pwscf.org >>> >>> Dear PWSCF user, >>> >>> I came across the error message below while running q2r.x for >>> rhomboheral structure of Bi after I have completed all the phonon >>> calculations. >>> >>> ====================================================== >>> nqs= 1 >>> q= 0.00000000 0.00000000 0.00000000 >>> reading dyn.mat. from file >>> dyn3 >>> >>> nqs= 6 >>> q= 0.35355340 0.20412410 0.57735030 >>> >>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >>> from init : error # 1 >>> q not allowed >>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >>> >>> stopping ... >>> >>> ================================================= >>> >>> I generated 2x2x2 grids of k-points using the k.points.x utililty by >>> suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points >>> >>> 1 0.0000000 0.0000000 0.0000000 1.00 >>> 2 0.3535534 -0.2041241 0.2886751 3.00 >>> 3 0.3535534 0.2041241 0.5773503 3.00 >>> 4 0.0000000 0.0000000 0.8660254 1.00 >>> >>> >>> which I used as q-vectors in the phonon calculation. >>> >>> I will appreciate any help to the problem. >>> >>> >>> Thanks >> >>> --__--__-- >>> >>> Message: 4 >>> Date: Mon, 10 Oct 2005 14:01:08 +0200 >>> From: Stefano de Gironcoli >>> To: pw_forum at pwscf.org >>> Subject: Re: [Pw_forum] q2r.x on rhomboheral >>> Reply-To: pw_forum at pwscf.org >>> >>> Are you specifying the grid 2x2x2 in the q2r.x input file ? >>> is the value of celldm(3)=1.0 appropriate for your crystal structure ? >>> stefano >>> >>> Adeagbo Waheed Adeniyi wrote: >>> >>>> Dear PWSCF user, >>>> >>>> I came across the error message below while running q2r.x for >>>> rhomboheral structure of Bi after I have completed all the phonon >>>> calculations. >>>> >>>> ====================================================== >>>> nqs= 1 >>>> q= 0.00000000 0.00000000 0.00000000 >>>> reading dyn.mat. from file >>>> dyn3 >>>> >>>> nqs= 6 >>>> q= 0.35355340 0.20412410 0.57735030 >>>> >>>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >>>> >>>> from init : error # 1 >>>> q not allowed >>>> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% >>>> >>>> >>>> stopping ... >>>> >>>> ================================================= >>>> >>>> I generated 2x2x2 grids of k-points using the k.points.x utililty by >>>> suppliying celldm(3) = 1.00 . I obtained these set of 4 k-points >>>> >>>> 1 0.0000000 0.0000000 0.0000000 1.00 >>>> 2 0.3535534 -0.2041241 0.2886751 3.00 >>>> 3 0.3535534 0.2041241 0.5773503 3.00 >>>> 4 0.0000000 0.0000000 0.8660254 1.00 >>>> >>>> >>>> which I used as q-vectors in the phonon calculation. >>>> >>>> I will appreciate any help to the problem. >>>> >>>> >>>> Thanks >>>> >>>> >> > =================================================== ADEAGBO WAHEED ADENIYI Institute of Physics, MC 328 University Duisburg-Essen Lotharstr. 1 47048 Duisburg Germany ==================================================== Phone: +49 203 379 1606 Fax: +49 203 379 3665 Present e-mail address: adeagbo at thp.uni-duisburg.de Permanent e-mail address: wadeagbo at mail.skannet.com http://www.thp.uni-duisburg.de ==================================================== "The intellect has little to do on the road to discovery. There comes a leap in consciousness, call it intuition or what you will, and the solution comes to you and you don't know how or why." -- AlbertEinstein From giannozz at nest.sns.it Wed Oct 12 13:32:16 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 12 Oct 2005 13:32:16 +0200 Subject: [Pw_forum] q2r.x on rhomboheral In-Reply-To: References: <20051011053511.17297.65160.Mailman@democritos.sissa.it> Message-ID: <200510121332.16096.giannozz@nest.sns.it> On Wednesday 12 October 2005 11:54, Adeagbo Waheed Adeniyi wrote: > But (ibrav=5) always ask for celldim(3) NOT celldim(4). oops...this is wrong. Please try the attached version of kpoint.f . Note that celldm(4) < 1 (it is the cosine of an angle) Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy -------------- next part -------------- A non-text attachment was scrubbed... Name: kpoints.f Type: text/x-fortran Size: 8559 bytes Desc: not available Url : /pipermail/attachments/20051012/dedb1b2e/attachment.bin From giannozz at nest.sns.it Wed Oct 12 17:19:33 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 12 Oct 2005 17:19:33 +0200 Subject: [Pw_forum] Question on based-centered monoclinic cell input In-Reply-To: <1128971625.434abd69aba9e@webmail.usask.ca> References: <1128971625.434abd69aba9e@webmail.usask.ca> Message-ID: <200510121719.33992.giannozz@nest.sns.it> On Monday 10 October 2005 21:13, Yansun Yao wrote: > I have a question on the input defining a based-centered > monoclinic cell (ibrav=13). I am afraid that the ibrav=13 case was not really correct. Please try if the attached version of "latgen.f90" produces more sensible lattice vectors. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy -------------- next part -------------- A non-text attachment was scrubbed... Name: latgen.f90 Type: text/x-fortran Size: 7627 bytes Desc: not available Url : /pipermail/attachments/20051012/53306e69/attachment.bin From konstantin_kudin at yahoo.com Wed Oct 12 17:28:04 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Wed, 12 Oct 2005 08:28:04 -0700 (PDT) Subject: [Pw_forum] Question on stress in a system with constraints In-Reply-To: <434C5D4E.9030706@mit.edu> Message-ID: <20051012152804.8020.qmail@web52013.mail.yahoo.com> Dear Nicola and Paolo, Thanks for the comments! However, I do not think that one needs the response functions from DFPT to remove constraints from the stress. What happens is that for homogeneous strain it is probably assumed that the fractional coordinates in the cell remain the same, however, the lattice vectors change, and so all the atomic Cartesian coordinates are updated. Is that how the stress actually defined? With constraints, the change in the lattice vectors should also update the fractional coordinates of some atoms, leading to extra derivatives which include the usual atomic forces of these atoms times the change in the fractional coordinates. Kostya --- Nicola Marzari wrote: > > > Hi Kostya, > > > a quick comment - Don Hamann has written a fairly extensive PRB this > year discussing many of these issues: Vol 72, 350105 (2005). > > CP stresses are (I hope) a derivative with respect to the strain > tensor - ie.e they do not take into account that > atoms can relax in response to the stress (the paradigmatic case is > the > response to a strain in the 111 direction in silicon - the internal > strain parameter measures how the distance between the two atoms > in the unit cell changes in response to the symmetry-breaking > stress). > So, you have the bare stress, calculated by CP and/or PWSCF, but you > want the renormalized one "dressed" by the relaxations (mediated by > the > inverse of the dynamical matrix, and by the coupling between > displacements and strains). The constraint will allow you to > renormalize appropriately the bare stress, if you have all the > response > functions from DFPT (and their correct long-wavelength limit) > but it might be easier to do it by finite differeces of the energy > along > the strain direction, while constraining the atoms. > > By the way - the dressing of a perturbation by the ionic relaxations > is very relevant for piezoelectricity (e.g. Stefano de Gironcoli > 1989 PRL) or for the interactions in substitutional alloy > (PRL 72 4001 (1994) - some of the issues with the long wavelength > limit > are discussed there). > > Best, > > nicola > > > > > Konstantin Kudin wrote: > > > Hi there, > > > > I have a basic question on the stress tensor. > > > > Presumably, the stress tensor computed in the CP implies > continuous > > stretching of the underlying system. Is this correct? Is the > derivative > > with respect to % of stretch, or is it to extra length in Bohrs, > for > > example? > > What are the units of the stress in CP ? > > > > If one projected out certain forces (for constrained coordinates), > > then the stress needs to be corrected for the fact that certain > forces > > are no longer there. Basically, the original stress was computed as > if > > all the forces were present, now, however, some of them are > missing. > > This means that if the system were to be stretched, the constrained > > coordinates would still be constrained. > > > > My question is how I can correct the stress in a system with > > constraints if this needs to be done. > > > > Thanks! > > Kostya > > > > > > > > __________________________________ > > Start your day with Yahoo! - Make it your home page! > > http://www.yahoo.com/r/hs > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > -- > --------------------------------------------------------------------- > Prof Nicola Marzari Department of Materials Science and Engineering > 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA > tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ From marzari at MIT.EDU Wed Oct 12 17:35:36 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Wed, 12 Oct 2005 11:35:36 -0400 Subject: [Pw_forum] Question on stress in a system with constraints In-Reply-To: <20051012152804.8020.qmail@web52013.mail.yahoo.com> References: <20051012152804.8020.qmail@web52013.mail.yahoo.com> Message-ID: <434D2D48.10207@mit.edu> Hi Kostya, I think that you do not need the response functions if you want to calculate the bare stress (i.e. derivative with respect to strain, in which fractional coordinates remain the same). If you want to have the bare stress plus constraints, you need to know at least the tensor that couples strain and forces. But keep in mind that the physical stress you want is dressed by the atominc internal relaxations; those would not be included in your bare CP stress, or in your bare CP stress plus contraints, and need in addition the inverse of the dynamical matrix. Any comments, anyone ? nicola Konstantin Kudin wrote: > Dear Nicola and Paolo, > > Thanks for the comments! > > However, I do not think that one needs the response functions from > DFPT to remove constraints from the stress. > > What happens is that for homogeneous strain it is probably assumed > that the fractional coordinates in the cell remain the same, however, > the lattice vectors change, and so all the atomic Cartesian coordinates > are updated. Is that how the stress actually defined? > > With constraints, the change in the lattice vectors should also update > the fractional coordinates of some atoms, leading to extra derivatives > which include the usual atomic forces of these atoms times the change > in the fractional coordinates. > > Kostya > > > > > --- Nicola Marzari wrote: > > >> >>Hi Kostya, >> >> >>a quick comment - Don Hamann has written a fairly extensive PRB this >>year discussing many of these issues: Vol 72, 350105 (2005). >> >>CP stresses are (I hope) a derivative with respect to the strain >>tensor - ie.e they do not take into account that >>atoms can relax in response to the stress (the paradigmatic case is >>the >>response to a strain in the 111 direction in silicon - the internal >>strain parameter measures how the distance between the two atoms >>in the unit cell changes in response to the symmetry-breaking >>stress). >>So, you have the bare stress, calculated by CP and/or PWSCF, but you >>want the renormalized one "dressed" by the relaxations (mediated by >>the >>inverse of the dynamical matrix, and by the coupling between >>displacements and strains). The constraint will allow you to >>renormalize appropriately the bare stress, if you have all the >>response >>functions from DFPT (and their correct long-wavelength limit) >>but it might be easier to do it by finite differeces of the energy >>along >>the strain direction, while constraining the atoms. >> >>By the way - the dressing of a perturbation by the ionic relaxations >>is very relevant for piezoelectricity (e.g. Stefano de Gironcoli >>1989 PRL) or for the interactions in substitutional alloy >>(PRL 72 4001 (1994) - some of the issues with the long wavelength >>limit >>are discussed there). >> >>Best, >> >> nicola >> >> >> >> >>Konstantin Kudin wrote: >> >> >>> Hi there, >>> >>> I have a basic question on the stress tensor. >>> >>> Presumably, the stress tensor computed in the CP implies >> >>continuous >> >>>stretching of the underlying system. Is this correct? Is the >> >>derivative >> >>>with respect to % of stretch, or is it to extra length in Bohrs, >> >>for >> >>>example? >>>What are the units of the stress in CP ? >>> >>> If one projected out certain forces (for constrained coordinates), >>>then the stress needs to be corrected for the fact that certain >> >>forces >> >>>are no longer there. Basically, the original stress was computed as >> >>if >> >>>all the forces were present, now, however, some of them are >> >>missing. >> >>>This means that if the system were to be stretched, the constrained >>>coordinates would still be constrained. >>> >>> My question is how I can correct the stress in a system with >>>constraints if this needs to be done. >>> >>> Thanks! >>> Kostya >>> >>> >>> >>>__________________________________ >>>Start your day with Yahoo! - Make it your home page! >>>http://www.yahoo.com/r/hs >>>_______________________________________________ >>>Pw_forum mailing list >>>Pw_forum at pwscf.org >>>http://www.democritos.it/mailman/listinfo/pw_forum >> >>-- >>--------------------------------------------------------------------- >>Prof Nicola Marzari Department of Materials Science and Engineering >>13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA >>tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu >>_______________________________________________ >>Pw_forum mailing list >>Pw_forum at pwscf.org >>http://www.democritos.it/mailman/listinfo/pw_forum >> > > > > > > __________________________________ > Yahoo! Music Unlimited > Access over 1 million songs. Try it free. > http://music.yahoo.com/unlimited/ > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From jasonsun98 at hotmail.com Thu Oct 13 09:05:44 2005 From: jasonsun98 at hotmail.com (sun jason) Date: Thu, 13 Oct 2005 07:05:44 +0000 Subject: [Pw_forum] RE: Pw_forum digest, Vol 1 #775 - 2 msgs In-Reply-To: <20051013053508.1606.5817.Mailman@democritos.sissa.it> Message-ID: An HTML attachment was scrubbed... URL: /pipermail/attachments/20051013/55542bda/attachment.htm From degironc at sissa.it Thu Oct 13 09:53:31 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Thu, 13 Oct 2005 09:53:31 +0200 (CEST) Subject: [Pw_forum] Question on stress in a system with constraints In-Reply-To: <434D2D48.10207@mit.edu> References: <20051012152804.8020.qmail@web52013.mail.yahoo.com> <434D2D48.10207@mit.edu> Message-ID: Dear Nicola and Kostya, I did not think deeply about the question, so I may be wrong, but I agree with Kostya. Stress is a first order derivative and as such only GS properties should be needed. Without constraint it is related to homogeneous deformation, in presence of constraints it is related to a locally inhomogeneous deformation but Hellman-Feynman theorem should still apply. I guess the result will be something like the unconstraint stress plus a term \sum_i,gamma F_i,gamma {\partial tau_i,gamma \over \partial \epsilon_{alpha,beta} } / omega where F_i,gamma is the (unconstrained) gamma component of the force on atom i and {\partial tau_i,gamma \over \partial \epsilon_{alpha,beta} } is the first order variation of the position of atom i with distortion needed to fulfill the constraint. In order to compute the stress induced by a displacement one indeed need a linear response calculation and one could think that also a term involving the internal strain parameter (Delta) is needed. something like \sigma^constr_{alpha,beta} = \sigma^unconstr_{alpha,beta} + \sum_i,gamma F_i,gamma * {\partial tau_i,gamma \over \partial \epsilon_{alpha,beta} } + \sum_i,gamma Delta_{alpha,beta;i,gamma} * delta tau_{i,gamma} but stress is computed at zero deformation therefore delta tau_{i,gamma} is actually zero stefano On Wed, 12 Oct 2005, Nicola Marzari wrote: > > > > Hi Kostya, > > I think that you do not need the response functions if you want to > calculate the bare stress (i.e. derivative with respect to strain, > in which fractional coordinates remain the same). If you want to have the > bare stress plus constraints, you need to know at least the tensor that > couples strain and forces. > > But keep in mind that the physical stress you want is dressed by the > atominc internal relaxations; those would not be included in your > bare CP stress, or in your bare CP stress plus contraints, and need in > addition the inverse of the dynamical matrix. > > Any comments, anyone ? > > nicola > > > Konstantin Kudin wrote: >> Dear Nicola and Paolo, >> >> Thanks for the comments! >> >> However, I do not think that one needs the response functions from >> DFPT to remove constraints from the stress. >> >> What happens is that for homogeneous strain it is probably assumed >> that the fractional coordinates in the cell remain the same, however, >> the lattice vectors change, and so all the atomic Cartesian coordinates >> are updated. Is that how the stress actually defined? >> >> With constraints, the change in the lattice vectors should also update >> the fractional coordinates of some atoms, leading to extra derivatives >> which include the usual atomic forces of these atoms times the change >> in the fractional coordinates. >> >> Kostya >> >> >> >> --- Nicola Marzari wrote: >> >> >>> >>> Hi Kostya, >>> >>> >>> a quick comment - Don Hamann has written a fairly extensive PRB this >>> year discussing many of these issues: Vol 72, 350105 (2005). >>> >>> CP stresses are (I hope) a derivative with respect to the strain >>> tensor - ie.e they do not take into account that >>> atoms can relax in response to the stress (the paradigmatic case is >>> the >>> response to a strain in the 111 direction in silicon - the internal >>> strain parameter measures how the distance between the two atoms >>> in the unit cell changes in response to the symmetry-breaking >>> stress). >>> So, you have the bare stress, calculated by CP and/or PWSCF, but you >>> want the renormalized one "dressed" by the relaxations (mediated by >>> the >>> inverse of the dynamical matrix, and by the coupling between >>> displacements and strains). The constraint will allow you to >>> renormalize appropriately the bare stress, if you have all the >>> response >>> functions from DFPT (and their correct long-wavelength limit) >>> but it might be easier to do it by finite differeces of the energy >>> along the strain direction, while constraining the atoms. >>> >>> By the way - the dressing of a perturbation by the ionic relaxations >>> is very relevant for piezoelectricity (e.g. Stefano de Gironcoli >>> 1989 PRL) or for the interactions in substitutional alloy >>> (PRL 72 4001 (1994) - some of the issues with the long wavelength >>> limit are discussed there). >>> >>> Best, >>> >>> nicola >>> >>> >>> >>> >>> Konstantin Kudin wrote: >>> >>> >>>> Hi there, >>>> >>>> I have a basic question on the stress tensor. >>>> >>>> Presumably, the stress tensor computed in the CP implies >>> >>> continuous >>> >>>> stretching of the underlying system. Is this correct? Is the >>> >>> derivative >>> >>>> with respect to % of stretch, or is it to extra length in Bohrs, >>> >>> for >>> >>>> example? >>>> What are the units of the stress in CP ? >>>> >>>> If one projected out certain forces (for constrained coordinates), >>>> then the stress needs to be corrected for the fact that certain >>> >>> forces >>> >>>> are no longer there. Basically, the original stress was computed as >>> >>> if >>> >>>> all the forces were present, now, however, some of them are >>> >>> missing. >>> >>>> This means that if the system were to be stretched, the constrained >>>> coordinates would still be constrained. >>>> >>>> My question is how I can correct the stress in a system with >>>> constraints if this needs to be done. >>>> >>>> Thanks! >>>> Kostya >>>> >>>> >>>> __________________________________ Start your day with Yahoo! >>>> - Make it your home page! http://www.yahoo.com/r/hs >>>> _______________________________________________ >>>> Pw_forum mailing list >>>> Pw_forum at pwscf.org >>>> http://www.democritos.it/mailman/listinfo/pw_forum >>> >>> -- >>> --------------------------------------------------------------------- >>> Prof Nicola Marzari Department of Materials Science and Engineering >>> 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA >>> tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu >>> _______________________________________________ >>> Pw_forum mailing list >>> Pw_forum at pwscf.org >>> http://www.democritos.it/mailman/listinfo/pw_forum >>> >> >> >> >> >> __________________________________ Yahoo! Music Unlimited >> Access over 1 million songs. Try it free. >> http://music.yahoo.com/unlimited/ >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum > > From degironc at sissa.it Thu Oct 13 10:21:29 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Thu, 13 Oct 2005 10:21:29 +0200 (CEST) Subject: [Pw_forum] RE: Pw_forum digest, Vol 1 #775 - 2 msgs In-Reply-To: References: Message-ID: is the value of c/a (celldm(3)) used in the scf calculation the same as the one used to generate the q-points ? Is the crystal symmetry actually lower than tetragonal in the case ? If so you need to compute additional q-points that would be equivalent in the Bravais lattice but are not in your crystal. Run your scf calculation with automatic grid 4 4 4 0 0 0 and stop it as soon as k-points are gerenated and see how many are there. These are the q-points to be used in the phonon calculation. Maybe you got a message like (i dont remember it exactly) " symmetry not allowed due to fractional translation " this means your FFT grid is incompatible with the fractional translation of your crystal and the accepted symmetry is lower than it could. Adjust nr1,nr2,nr3 in the input so that fractional translations connect FFT grid points and all symmetry is restored. stefano On Thu, 13 Oct 2005, sun jason wrote: > > Dear all, > > when I doing a phonon spectra test of a?primitive tetragonal lattice with > 4*4*4 uniform q grid, the nscf and the phonon calculation is?right. > but?when I perform?q2r.x, the program stopped, the error message said: > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > ???? task #********** > ???? from init : error #?????? 197 > ????? missing q-point(s)! > ?%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > my input file is like this > > ?&input > ?? nr1=4, nr2=4, nr3=4, zasr=.true., fild='test444.fc' > ?/ > ?18 > ?test.dynG > ?dyn.0.2500000,0.0000000,0.0000000 > ?dyn.0.5000000,0.0000000,0.0000000 > ?dyn.0.2500000,0.2500000,0.0000000 > ?dyn.0.5000000,0.2500000,0.0000000 > ?dyn.0.5000000,0.5000000,0.0000000 > ?dyn.0.0000000,0.0000000,0.1745709 > ?dyn.0.2500000,0.0000000,0.1745709 > ?dyn.0.5000000,0.0000000,0.1745709 > ?dyn.0.2500000,0.2500000,0.1745709 > ?dyn.0.5000000,0.2500000,0.1745709 > ?dyn.0.5000000,0.5000000,0.1745709 > ?dyn.0.0000000,0.0000000,0.3491419 > ?dyn.0.2500000,0.0000000,0.3491419 > ?dyn.0.5000000,0.0000000,0.3491419 > ?dyn.0.2500000,0.2500000,0.3491419 > ?dyn.0.5000000,0.2500000,0.3491419 > ?dyn.0.5000000,0.5000000,0.3491419 > > ----------------------------------- > > I calculate the phonon spectra of another lattice, just change the > lattice constants, atomic species and q vector, the spectra is right. so > what's wrong with it? > > thank you in advance. > > Best, > > > ============================================== > Jian SUN > Physics Dept. of Nanjing University > National Lab. of Solid State Microstructures > 22 Hankou Road, Gulou District > Nanjing, Jiangsu Province > 210093 > China > ============================================== > _______________________________________________ Pw_forum mailing list > Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum > From yay451 at mail.usask.ca Thu Oct 13 16:41:22 2005 From: yay451 at mail.usask.ca (Yansun Yao) Date: Thu, 13 Oct 2005 08:41:22 -0600 Subject: [Pw_forum] Question on based-centered monoclinic cell input In-Reply-To: <200510121719.33992.giannozz@nest.sns.it> References: <1128971625.434abd69aba9e@webmail.usask.ca> <200510121719.33992.giannozz@nest.sns.it> Message-ID: <1129214482.434e7212dd2a8@webmail.usask.ca> Dear Paolo Thank you very much for your help! Yansun Yao From liyanpcl at yahoo.com.cn Fri Oct 14 09:03:32 2005 From: liyanpcl at yahoo.com.cn (li yan) Date: Fri, 14 Oct 2005 15:03:32 +0800 (CST) Subject: [Pw_forum] a question about the structure optim In-Reply-To: <1129214482.434e7212dd2a8@webmail.usask.ca> Message-ID: <20051014070333.28403.qmail@web15610.mail.cnb.yahoo.com> dear Polo, I met a problem when I was doing structure optim of beta-phase AgI. The space group is P63MC which has 12 symmitry operations. The atoms positions are Ag 0.3333, 0.6667, 0.000 Ag 0.6667, 0.3333, 0.500 I 0.3333, 0.6667, 0.3345 I 0.6667, 0.3333, 0.8345 I first do the scf. In agi.scf.out , there is only 2 symmitry operations, and the total force is 514.926. And P=********* 26318.28 -19.20 -0.01 -19.20 26296.11 0.00 -0.01 0.00 556494 My questions are that why the total force and P is so large and why the pwscf can't find 12 symmitry operations. --------------------------------- DO YOU YAHOO!? ???????????2G?????pop3?????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051014/6307494a/attachment.htm From giannozz at nest.sns.it Fri Oct 14 11:07:37 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 14 Oct 2005 11:07:37 +0200 Subject: [Pw_forum] a question about the structure optim In-Reply-To: <20051014070333.28403.qmail@web15610.mail.cnb.yahoo.com> References: <20051014070333.28403.qmail@web15610.mail.cnb.yahoo.com> Message-ID: <200510141107.37426.giannozz@nest.sns.it> On Friday 14 October 2005 09:03, li yan wrote: > dear Polo, messages to pw_forum are addressed to everybody, not to a single person. The next messages addressed to me are guaranteed not to get any answer from me. > My questions are that why the total force and P is so large most likely, because your structure is far from equilibrium. DId you check your structure with xcrysden? Is the cutoff appropriate (for P in particular)? > and why the pwscf can't find 12 symmetry operations. most likely, because you used too few significant figures in atomic positions. It is explained in detail in the manual Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From afloris at physik.fu-berlin.de Fri Oct 14 12:53:11 2005 From: afloris at physik.fu-berlin.de (Andrea Floris) Date: Fri, 14 Oct 2005 12:53:11 +0200 (CEST) Subject: [Pw_forum] Re: projwfc.f90 In-Reply-To: <20051011231901.22029.12292.Mailman@democritos.sissa.it> References: <20051011231901.22029.12292.Mailman@democritos.sissa.it> Message-ID: Many thanks Paolo for your suggestions, very useful. cheers Andrea > Message: 7 > From: Paolo Giannozzi > Organization: Scuola Normale Superiore di Pisa > To: pw_forum at pwscf.org > Subject: Re: [Pw_forum] projwfc.f90 > Date: Tue, 11 Oct 2005 17:33:46 +0200 > Reply-To: pw_forum at pwscf.org > > On Tuesday 11 October 2005 16:52, Andrea Floris wrote: > > > does it exists some documentation (papers, some notes or whatever) > > on which the program projwfc.f90 (and the routines that it calls) is > > based? > > some time ago I wrote some notes (latex file below) about Loewdin > charges. Hope this helps > > Paolo > -- > Paolo Giannozzi e-mail: giannozz at nest.sns.it > Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 > Piazza dei Cavalieri 7 I-56126 Pisa, Italy > -- > \documentclass[a4paper,12pt]{article} > > \begin{document} > > \section{L\"owdin and Mulliken population analysis} > > We have $| \phi_\mu\rangle$ atomic (or atomic-like) > orbitals, not necessarily orthonormal, and we want > to write KS orbitals for our system: > $|\psi_\alpha\rangle$, as sums over said > atomic orbitals: > $$ > |\psi_\alpha\rangle = \sum_\mu c_\mu^{(\alpha)} | \phi_\mu \rangle. > $$ > The generalized orthonormality relations for KS orbitals > is written as: > $$ > \langle\psi_\alpha| \hat S |\psi_\beta\rangle = \delta_{\alpha\beta}. > $$ > where $\hat S$ is an operator defined in the US PP framework. > The charge density $\rho({\bf r})$ is given by > $$ > \rho({\bf r}) = \sum_{\mu,\nu} P_{\mu\nu} > \left (\phi^*_\nu \hat S \phi_\mu\right) ({\bf r}) > $$ > where > $$ > P_{\mu\nu} = \sum_\alpha c_\mu^{(\alpha)} c_\nu^{(\alpha)*} > $$ > define an operator $\hat P$, and > $$ > \left (\phi^*_\nu \hat S \phi_\mu\right) ({\bf r}) = > \phi^*_\nu({\bf r}) \phi_\mu({\bf r}) + > \sum_{lm} \langle\phi_\nu|\beta_l\rangle q_{lm}({\bf r}) > \langle\beta_m|\phi_\mu\rangle > $$ > where the $\beta$'s and $q$'s are components of the US PP. > > \subsection{Mulliken population analysis} > > We write the total number of electrons $N$ as: > $$ > N = \int \rho({\bf r}) d{\bf r} > = \mbox{Tr} \hat P\hat S = \sum_\mu (\sum_\nu P_{\mu\nu} S_{\nu\mu}) > \equiv \sum_\mu q_\mu > $$ > where $q_\mu$ is the {\em Mulliken charge} associated to state $\mu$ : > $$ > q_\mu = \sum_\nu \sum_\alpha c_\mu^{(\alpha)} c_\nu^{(\alpha)*} S_{\nu\mu} > $$ > and > $$ > S_{\nu\mu} = \langle\phi_\nu|\hat S|\phi_\mu\rangle > = \int \phi^*_\nu({\bf r}) \phi_\mu({\bf r}) d{\bf r} + > \sum_{lm} \langle\phi_\nu|\beta_l\rangle Q_{lm} > \langle\beta_m|\phi_\mu\rangle > $$ > with > $$ > Q_{lm}=\int q_{lm}({\bf r}) d{\bf r}. > $$ > In general, this matrix is not diagonal, even with NC PP. > The coefficients $c_\mu^{(\alpha)}$ are obtained by inverting the linear > system: > $$ > \langle \phi_\nu | \hat S | \psi_\alpha\rangle = > \sum_\mu c_\mu^{(\alpha)} S_{\nu\mu} > $$ > that is > $$ > c_\mu^{(\alpha)} = (\hat S^{-1})_{\mu\nu} > \langle \phi_\nu | \hat S | \psi_\alpha\rangle > $$ > and finally > $$ > q_\mu = \sum_{\alpha\nu} > \langle \psi_\alpha | \hat S | \phi_\mu\rangle > (\hat S^{-1})_{\mu\nu} > \langle \phi_\nu | \hat S | \psi_\alpha\rangle. > $$ > > \subsection{L\"owdin population analysis} > > The total number of electrons $N$ can be alternatively written as > $$ > N = \mbox{Tr} \left[\hat S^{1/2}\hat P \hat S^{1/2}\right] > = \sum_\mu \tilde q_\mu. > $$ > where $\tilde q_\mu$ is called {\em L\"owdin charge} associated to state > $\mu$. Let us introduce an auxiliary set of atomic orbitals > $\tilde\phi$ via : > $$ > |\tilde\phi_\mu\rangle = \sum_\nu (\hat S^{-1/2})_{\nu\mu}| \phi_\nu\rangle, > \qquad > | \phi_\mu\rangle = \sum_\nu (\hat S^{1/2})_{\nu\mu}|\tilde\phi_\nu\rangle, > $$ > for which the generalized orthonormality relation > $$ > \langle\tilde\phi_\mu|\hat S| \tilde\phi_\nu\rangle = \delta_{\mu\nu} > $$ > holds. The KS orbitals for our systems can be rewritten as > $$ > |\psi_\alpha\rangle = \sum_\mu c_\mu^{(\alpha)} | \phi_\mu \rangle > = \sum_\mu \tilde c_\mu^{(\alpha)} | \tilde\phi_\mu > \rangle > $$ > where > $$ > \tilde c_\mu^{(\alpha)} = \sum_\nu (\hat S^{1/2})_{\nu\mu} c_\nu^{(\alpha)}. > $$ > By comparison with the above expression of $\tilde q_\mu$ we get > $$ > \tilde q_\mu = \sum_\alpha |\tilde c_\mu^{(\alpha)}|^2 > = |\langle \tilde\phi_\mu | \hat S | \psi_\alpha\rangle|^2 > $$ > > Reference: {\em Modern Quantum Chemistry}, A. Szabo and N. Ostlund (Dover, NY > 1996), > p. 153 > \end{document} > > From yay451 at mail.usask.ca Fri Oct 14 17:11:29 2005 From: yay451 at mail.usask.ca (Yansun Yao) Date: Fri, 14 Oct 2005 09:11:29 -0600 Subject: [Pw_forum] vc-relaxation question Message-ID: <1129302689.434fcaa11d2e4@webmail.usask.ca> Dear pwscf users, I met a problem when I did vc-relaxtion. First of all vc-relaxtion ran perfect for me- I got good enough forces, stress, and total energy. But when I use the optimized parameters as input to do a scf run, forces,stress and total energy changed. Stress changed a lot espercially. I am doing a orthorhombic stucture with the Gamma only pw.x. To double check the results, I used both ibrav=0 and ibrav=8 describtion and those two yield the same results. I attached some information below. Could anybody help me to figure out the problem? Thank you so much! Yansun Yao vc-relax input: &control calculation='vc-relax', restart_mode='from_scratch' prefix='AlOH3' pseudo_dir = '/home/espresso-2.1.4/pseudo/', outdir='/home/tmp/' tstress=.ture., tprnfor=.true., nstep=200, dt = 50.0, / &system ibrav = 0, celldm(1)=8.3161, nat= 16, ntyp= 3, ecutwfc = 70.0, nosym=.t., / &electrons mixing_beta = 0.4 mixing_mode = 'local-TF' conv_thr = 1.0d-8 / &ions upscale=20 / &CELL cell_dynamics='damp-pr', press=0.0010, / ATOMIC_SPECIES Al 26.98 Al.cpi.UPF O 15.9994 O.cpi.UPF H 1.008 H.cpi.UPF ATOMIC_POSITIONS {crystal} Al 0.0447200 0.8555000 0.7500000 1 1 0 Al 0.5447200 0.6445000 0.2500000 1 1 0 Al 0.9552800 0.1445000 0.2500000 1 1 0 **some more here** CELL_PARAMETERS {cubic} 1.000000000 0.000000000 0.000000000 0.000000000 2.141773000 0.000000000 0.000000000 0.000000000 0.646533000 K_POINTS {gamma} ************************************************ The final estimate: input alat = 8.3161 (a.u.) CELL_PARAMETERS (alat) 0.976352770 0.000021968 0.000000000 0.000047050 2.048611386 0.000000000 0.000000000 0.000000000 0.729286228 ATOMIC_POSITIONS (crystal) Al 0.079415970 0.852920557 0.750000000 Al 0.579460122 0.647129732 0.250000000 Al 0.920584030 0.147079443 0.250000000 ****some more here**** ************************************************* The scf input after vc-relaxation: &control calculation='scf' pseudo_dir = '/home/espresso-2.1.4/pseudo/', outdir='/home/tmp/' title='Test aloh3 gamma only' prefix='aloh3_gamma' tprnfor=.true., tstress=.true., / &system ibrav = 0, celldm(1) = 8.3161, nat= 16, ntyp= 3, ecutwfc = 70.0, / &electrons mixing_beta = 0.4 mixing_mode = 'local-TF' conv_thr = 1.0d-8 / ATOMIC_SPECIES Al 26.98 Al.cpi.UPF O 15.9994 O.cpi.UPF H 1.008 H.cpi.UPF ATOMIC_POSITIONS (crystal) Al 0.079415970 0.852920557 0.750000000 Al 0.579460122 0.647129732 0.250000000 Al 0.920584030 0.147079443 0.250000000 ****some more here****** CELL_PARAMETERS {cubic} 0.976352770 0.000021968 0.000000000 0.000047050 2.048611386 0.000000000 0.000000000 0.000000000 0.729286228 K_POINTS {gamma} ************************************************ And the result shows: Total force = 0.002185 Total SCF correction = 0.000079 entering subroutine stress ... total stress (ryd/bohr**3) (kbar) P= 47.14 0.00009568 0.00000209 0.00000000 14.08 0.31 0.00 0.00000209 0.00003429 0.00000000 0.31 5.04 0.00 0.00000000 0.00000000 0.00083142 0.00 0.00 122.31 From degironc at sissa.it Fri Oct 14 17:24:42 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Fri, 14 Oct 2005 17:24:42 +0200 Subject: [Pw_forum] vc-relaxation question References: <1129302689.434fcaa11d2e4@webmail.usask.ca> Message-ID: <434FCDBA.6010408@sissa.it> Stress converge slowly with cutoff, try to increase the cutoff and check again. Notice that during vc relaxation the basis set, originally defined as the plane waves contained inside the cutoff sphere, deforms with the recirocal lattice and the sphere becomes an ellipsoid with higher effective cutoff in the directions where the cell shinked and lower effective cutoff where the cell expanded. If the basis is complete enough this does not create problems but if it is not you can have large differences when after relaxation you recalculate (with an again spherical cutoff) the stress in the supposedly relaxed structure. In your calculation the z direction expanded substantially (12 %) and your effecctive cutoff in that direction is only some 3/4 of the original one which might be not accurate enough for Oxygen. stefano Yansun Yao wrote: >Dear pwscf users, > I met a problem when I did vc-relaxtion. First of all vc-relaxtion ran >perfect for me- I got good enough forces, stress, and total energy. But when I >use the optimized parameters as input to do a scf run, forces,stress and total >energy changed. Stress changed a lot espercially. > I am doing a orthorhombic stucture with the Gamma only pw.x. To double check >the results, I used both ibrav=0 and ibrav=8 describtion and those two yield >the same results. > I attached some information below. Could anybody help me to figure out the >problem? Thank you so much! > Yansun Yao >vc-relax input: >&control > calculation='vc-relax', > restart_mode='from_scratch' > prefix='AlOH3' > pseudo_dir = '/home/espresso-2.1.4/pseudo/', > outdir='/home/tmp/' > tstress=.ture., > tprnfor=.true., > nstep=200, > dt = 50.0, > / > &system > ibrav = 0, celldm(1)=8.3161, > nat= 16, ntyp= 3, > ecutwfc = 70.0, nosym=.t., > / > &electrons > mixing_beta = 0.4 > mixing_mode = 'local-TF' > conv_thr = 1.0d-8 > / > &ions > upscale=20 > / > &CELL > cell_dynamics='damp-pr', press=0.0010, > / >ATOMIC_SPECIES > Al 26.98 Al.cpi.UPF > O 15.9994 O.cpi.UPF > H 1.008 H.cpi.UPF >ATOMIC_POSITIONS {crystal} >Al 0.0447200 0.8555000 0.7500000 1 1 0 >Al 0.5447200 0.6445000 0.2500000 1 1 0 >Al 0.9552800 0.1445000 0.2500000 1 1 0 >**some more here** >CELL_PARAMETERS {cubic} > 1.000000000 0.000000000 0.000000000 > 0.000000000 2.141773000 0.000000000 > 0.000000000 0.000000000 0.646533000 >K_POINTS {gamma} >************************************************ >The final estimate: > input alat = 8.3161 (a.u.) >CELL_PARAMETERS (alat) > 0.976352770 0.000021968 0.000000000 > 0.000047050 2.048611386 0.000000000 > 0.000000000 0.000000000 0.729286228 > >ATOMIC_POSITIONS (crystal) >Al 0.079415970 0.852920557 0.750000000 >Al 0.579460122 0.647129732 0.250000000 >Al 0.920584030 0.147079443 0.250000000 >****some more here**** >************************************************* >The scf input after vc-relaxation: >&control > calculation='scf' > pseudo_dir = '/home/espresso-2.1.4/pseudo/', > outdir='/home/tmp/' > title='Test aloh3 gamma only' > prefix='aloh3_gamma' > tprnfor=.true., > tstress=.true., > / > &system > ibrav = 0, celldm(1) = 8.3161, > nat= 16, ntyp= 3, > ecutwfc = 70.0, > / > &electrons > mixing_beta = 0.4 > mixing_mode = 'local-TF' > conv_thr = 1.0d-8 > / >ATOMIC_SPECIES > Al 26.98 Al.cpi.UPF > O 15.9994 O.cpi.UPF > H 1.008 H.cpi.UPF >ATOMIC_POSITIONS (crystal) >Al 0.079415970 0.852920557 0.750000000 >Al 0.579460122 0.647129732 0.250000000 >Al 0.920584030 0.147079443 0.250000000 >****some more here****** >CELL_PARAMETERS {cubic} > 0.976352770 0.000021968 0.000000000 > 0.000047050 2.048611386 0.000000000 > 0.000000000 0.000000000 0.729286228 >K_POINTS {gamma} >************************************************ >And the result shows: > Total force = 0.002185 Total SCF correction = 0.000079 > entering subroutine stress ... > > total stress (ryd/bohr**3) (kbar) P= 47.14 > 0.00009568 0.00000209 0.00000000 14.08 0.31 0.00 > 0.00000209 0.00003429 0.00000000 0.31 5.04 0.00 > 0.00000000 0.00000000 0.00083142 0.00 0.00 122.31 > > > > > >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From L.Boeri at fkf.mpg.de Fri Oct 14 19:46:33 2005 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Fri, 14 Oct 2005 19:46:33 +0200 (CEST) Subject: [Pw_forum] charge density and pot file In-Reply-To: <200510111900.10851.giannozz@nest.sns.it> References: <200510111707.16019.giannozz@nest.sns.it> <200510111900.10851.giannozz@nest.sns.it> Message-ID: dear Paolo, thanks a lot for your answer. Actually, in the chdens.f90 file there was all that was needed: the rho (or potential) is stored in G space in the rhog array, and the g vectors in a vector which is called g. So I just had to print them out! thanks a lot for your response, it was very useful, best, Lilia On Tue, 11 Oct 2005, Paolo Giannozzi wrote: > On Tuesday 11 October 2005 17:41, Lilia Boeri wrote: > > > In particular, I have problems in locating which part of the code > > attaches an index "ig" to a particular plane wave: i.e. how is the > > index i related to ix, iy,iz? > > it depends on what ix, iy, iz are ... There are two different ways > of storing G-components. One is as list of components for > wavevector g_i (i=1,ngm): g_i = ( G(1,i), G(2,i), G(3,i) ) * 2pi/a, > where variable G is contained in module gvect. The other way > is the "FFT grid" that is used to perform FFT. The latter is > related to the former by a set of indices NL: rho(g_i) = rho_fft(nl(i)) > Unless you need to do something special, this should be more > or less all you need > > Paolo > > -- > Paolo Giannozzi e-mail: giannozz at nest.sns.it > Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 > Piazza dei Cavalieri 7 I-56126 Pisa, Italy > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From yang_jun-jun at hotmail.com Sun Oct 16 16:23:23 2005 From: yang_jun-jun at hotmail.com (Jun-jun Yang) Date: Sun, 16 Oct 2005 07:23:23 -0700 Subject: [Pw_forum] Very long time of computation Message-ID: Hi, I'm a newbie user of pwscf, and working on total energy calculation of chromite mineral as my first step for getting familiar with the program. But when I run the program, it looks like stpping at wavefunction calculation and seems endless. starting charge 543.97604, renormalized to 544.00000 Starting wfc are atomic The longest time it stopped (maybe still running) here was over 5 hours on one processor. The unit cell of the mineral has 56 atoms (Z=8) and crystal system is cubic. I'm using USPP potential. Please refer to snippet of my input file below. Do I make anything wrong? Is such a long computation time reasonable? &control calculation = 'scf' restart_mode='from_scratch', prefix='fecr2o4', tstress = .true. tprnfor = .true. pseudo_dir = '../../pseudo/', outdir = $HOME/temp/ / &system ibrav = 1, celldm(1) = 15.8321, nat = 56 , ntyp = 3, ecutwfc = 40.0 / &electrons diagonalization = cg, mixing_mode = 'plain', mixing_beta = 0.7, conv_thr = 1.0d-8 / ATOMIC_SPECIES Fe 55.847 Fe.pbe-sp-van.UPF Cr 51.996 Cr.pbe-sp-van.UPF O 15.999 O.pbe-van_ak.UPF ATOMIC_POSITIONS angstrom Fe 0.00000 0.00000 0.00000 Cr 5.23625 5.23625 5.23625 O 3.23391 3.23391 3.23391 ...... ...... K_POINTS automatic 8 8 8 0 0 0 Please help and thanks in advance. Jun-jun Yang _________________________________________________________________ Don?t just search. Find. Check out the new MSN Search! http://search.msn.click-url.com/go/onm00200636ave/direct/01/ From eyvaz_isaev at yahoo.com Sun Oct 16 17:31:26 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Sun, 16 Oct 2005 08:31:26 -0700 (PDT) Subject: [Pw_forum] Very long time of computation In-Reply-To: Message-ID: <20051016153126.39289.qmail@web60312.mail.yahoo.com> Hi, For 56 atoms and 8x8x8 q-points it is OK. You can decrease ecutwfc down to 30Ry. Much more you can gain using just the Gamma point (1 1 1 0 0 0). Your calculations will be more time demanding if you turn to spin-polarized case. Even for parallel calculations it will take time. Best regards, Eyvaz. P.S. Besides, you forgot ecutrho parameter which is important for US pseudopotentials. --- Jun-jun Yang wrote: > Hi, > > I'm a newbie user of pwscf, and working on total > energy calculation of > chromite mineral as > my first step for getting familiar with the program. > But when I run the > program, it looks like > stpping at wavefunction calculation and seems > endless. > > starting charge 543.97604, renormalized to 544.00000 > Starting wfc are atomic > > The longest time it stopped (maybe still running) > here was over 5 hours on > one processor. > The unit cell of the mineral has 56 atoms (Z=8) and > crystal system is cubic. > I'm using USPP > potential. Please refer to snippet of my input file > below. Do I make > anything wrong? Is such > a long computation time reasonable? > > &control > calculation = 'scf' > restart_mode='from_scratch', > prefix='fecr2o4', > tstress = .true. > tprnfor = .true. > pseudo_dir = '../../pseudo/', > outdir = $HOME/temp/ > / > &system > ibrav = 1, celldm(1) = 15.8321, nat = 56 , ntyp > = 3, > ecutwfc = 40.0 > / > &electrons > diagonalization = cg, > mixing_mode = 'plain', > mixing_beta = 0.7, > conv_thr = 1.0d-8 > / > ATOMIC_SPECIES > Fe 55.847 Fe.pbe-sp-van.UPF > Cr 51.996 Cr.pbe-sp-van.UPF > O 15.999 O.pbe-van_ak.UPF > ATOMIC_POSITIONS angstrom > Fe 0.00000 0.00000 0.00000 > Cr 5.23625 5.23625 5.23625 > O 3.23391 3.23391 3.23391 > ...... > ...... > > K_POINTS automatic > 8 8 8 0 0 0 > > Please help and thanks in advance. > > > Jun-jun Yang > > _________________________________________________________________ > Don?t just search. Find. Check out the new MSN > Search! > http://search.msn.click-url.com/go/onm00200636ave/direct/01/ > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From giannozz at nest.sns.it Mon Oct 17 09:30:26 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 17 Oct 2005 09:30:26 +0200 Subject: [Pw_forum] Very long time of computation In-Reply-To: References: Message-ID: <200510170930.26900.giannozz@nest.sns.it> On Sunday 16 October 2005 16:23, Jun-jun Yang wrote: > I'm a newbie user of pwscf, and working on total energy calculation > of chromite mineral as my first step for getting familiar with the > program. [...] The unit cell of the mineral has 56 atoms (Z=8) you should star with simpler materials if you want to get familiar with the code > tstress = .true. don't calculate stress if you do not need to > ecutwfc = 40.0 start with a smaller cutoff, increase if needed > diagonalization = cg, davidson is usually faster (but takes more memory) > K_POINTS automatic > 8 8 8 0 0 0 start with Gamma only, using all tricks to enhance the speed: K_POINTS gamma and increase if needed. The CPU time needed is proportional to the number of k-points and to the number of plane waves. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From yang_jun-jun at hotmail.com Mon Oct 17 09:54:43 2005 From: yang_jun-jun at hotmail.com (Jun-jun Yang) Date: Mon, 17 Oct 2005 00:54:43 -0700 Subject: [Pw_forum] Very long time of computation In-Reply-To: <20051016153126.39289.qmail@web60312.mail.yahoo.com> Message-ID: Thank you, Eyvaz. I tried several different settings in accordence of your recommendation. Now ecutwfc is decreased to 30 Ryd and ecutrho is set to 140 Ryd. But if I set KPOINTS gamma, the program will run into an error: Starting wfc are atomic %%%%%%%%%%%%%%%%%%%%%%% drom cdiaghg : error # 497 info =/= 0 %%%%%%%%%%%%%%%%%%%%%%% Stopping 2 If I set KPOINTS automatic 1 1 1 0 0 0 the program can continue to run (start SCF calculation) although I don't know if it can get correct result. Each iteration costs 3-4 minutes and the estimated scf accuracy seems not so good (from <160 Ryd to < 2400 Ryd, then back to 320Ryd). Why the two KPOINTS settings result in different consequences? What does the error mean? Regards, Jun-jun >From: Eyvaz Isaev >Reply-To: pw_forum at pwscf.org >To: pw_forum at pwscf.org >Subject: Re: [Pw_forum] Very long time of computation >Date: Sun, 16 Oct 2005 08:31:26 -0700 (PDT) > >Hi, > >For 56 atoms and 8x8x8 q-points it is OK. You can >decrease ecutwfc down to 30Ry. Much more you can gain >using just the Gamma point (1 1 1 0 0 0). >Your calculations will be more time demanding if you >turn to spin-polarized case. >Even for parallel calculations it will take time. > >Best regards, >Eyvaz. >P.S. Besides, you forgot ecutrho parameter which is >important for US pseudopotentials. > > >--- Jun-jun Yang wrote: > > > Hi, > > > > I'm a newbie user of pwscf, and working on total > > energy calculation of > > chromite mineral as > > my first step for getting familiar with the program. > > But when I run the > > program, it looks like > > stpping at wavefunction calculation and seems > > endless. > > > > starting charge 543.97604, renormalized to 544.00000 > > Starting wfc are atomic > > > > The longest time it stopped (maybe still running) > > here was over 5 hours on > > one processor. > > The unit cell of the mineral has 56 atoms (Z=8) and > > crystal system is cubic. > > I'm using USPP > > potential. Please refer to snippet of my input file > > below. Do I make > > anything wrong? Is such > > a long computation time reasonable? > > > > &control > > calculation = 'scf' > > restart_mode='from_scratch', > > prefix='fecr2o4', > > tstress = .true. > > tprnfor = .true. > > pseudo_dir = '../../pseudo/', > > outdir = $HOME/temp/ > > / > > &system > > ibrav = 1, celldm(1) = 15.8321, nat = 56 , ntyp > > = 3, > > ecutwfc = 40.0 > > / > > &electrons > > diagonalization = cg, > > mixing_mode = 'plain', > > mixing_beta = 0.7, > > conv_thr = 1.0d-8 > > / > > ATOMIC_SPECIES > > Fe 55.847 Fe.pbe-sp-van.UPF > > Cr 51.996 Cr.pbe-sp-van.UPF > > O 15.999 O.pbe-van_ak.UPF > > ATOMIC_POSITIONS angstrom > > Fe 0.00000 0.00000 0.00000 > > Cr 5.23625 5.23625 5.23625 > > O 3.23391 3.23391 3.23391 > > ...... > > ...... > > > > K_POINTS automatic > > 8 8 8 0 0 0 > > > > Please help and thanks in advance. > > > > > > Jun-jun Yang > > > > >_________________________________________________________________ > > Don?t just search. Find. Check out the new MSN > > Search! > > >http://search.msn.click-url.com/go/onm00200636ave/direct/01/ > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > > >__________________________________ >Yahoo! Mail - PC Magazine Editors' Choice 2005 >http://mail.yahoo.com >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum _________________________________________________________________ Is your PC infected? Get a FREE online computer virus scan from McAfee? Security. http://clinic.mcafee.com/clinic/ibuy/campaign.asp?cid=3963 From chaohao2002 at 163.com Mon Oct 17 10:10:01 2005 From: chaohao2002 at 163.com (Chaohao Hu) Date: Mon, 17 Oct 2005 16:10:01 +0800 Subject: [Pw_forum] a question about parameters for ecutwfc and ecutrho Message-ID: <20051017080504.9A8CD112FAD@democritos.sissa.it> Dear all users, Now I am confused to the difference between ecutwfc and ecutrho. In general plane wave code, only the energy cutoff for wavefunctions is adequate. Why do we need ecutrho parameter in PWSCF? In the help file, The explaination for ecutrho is too simple. Who can tell more details and orginal reference about it? Thanks in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051017/90f4d051/attachment.htm From giannozz at nest.sns.it Mon Oct 17 10:39:19 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 17 Oct 2005 10:39:19 +0200 Subject: [Pw_forum] Very long time of computation In-Reply-To: References: Message-ID: <200510171039.19937.giannozz@nest.sns.it> On Monday 17 October 2005 09:54, Jun-jun Yang wrote: > Why the two KPOINTS settings result in different consequences? they don't. If you are ABSOLUTELY sure that the same calculation, done with K_POINTS Gamma and done with K_POINTS 1 0.0 0.0 0.0 0 0 0 or K_POINTS (auto) 1 1 1 0 0 0 yields different results (one set of results meaningful and one set not: both sets meaningless is another story), please provide a test job Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Mon Oct 17 18:16:34 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 17 Oct 2005 18:16:34 +0200 Subject: [Pw_forum] a question about parameters for ecutwfc and ecutrho In-Reply-To: <20051017080504.9A8CD112FAD@democritos.sissa.it> References: <20051017080504.9A8CD112FAD@democritos.sissa.it> Message-ID: <200510171816.34864.giannozz@nest.sns.it> On Monday 17 October 2005 10:10, Chaohao Hu wrote: > I am confused to the difference between ecutwfc and ecutrho. > In general plane wave code, using norm-conserving pseudopotentials, > only the energy cutoff for wavefunctions is adequate. > Why do we need ecutrho parameter in PWSCF? because it may be useful with ultrasoft pseudopotentials. See for instance Phys. Rev. B 47, 10142 (1993); J. Chem. Phys. 120, 5903-5915 (2004). The latter is available here: http://web1.sns.it/~giannozz/Papers/JCP_57.pdf Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From proffess at yandex.ru Mon Oct 17 21:21:06 2005 From: proffess at yandex.ru (Sergey Lisenkov) Date: Mon, 17 Oct 2005 23:21:06 +0400 (MSD) Subject: [Pw_forum] Convergence with vc-relax Message-ID: <4353F9A2.000002.23046@mfront7.yandex.ru> Dear PWscf users, I have a question regarding convergence criteria during vc-relaxation. At my last iteration (not finished yet) I have: Total force = 0.000324 Total SCF correction = 0.000008 entering subroutine stress ... total stress (ryd/bohr**3) (kbar) P= -0.15 0.00000113 -0.00000182 -0.00000043 0.17 -0.27 -0.06 -0.00000182 -0.00000232 0.00000001 -0.27 -0.34 0.00 -0.00000043 0.00000001 -0.00000192 -0.06 0.00 -0.28 Entering Dynamics; it = 514 time = 1.86219 pico-seconds AFAIK, convergence criteria for stresses should be 0.5, which was already reached. Am I wrong? What is then criteria for VC-relax? PW's version is 2.1.4 Thanks a lot, Best wishes, Sergey From naromero at gmail.com Tue Oct 18 22:54:30 2005 From: naromero at gmail.com (Nichols A. Romero) Date: Tue, 18 Oct 2005 16:54:30 -0400 Subject: [Pw_forum] structural relaxation with constraints Message-ID: <6ac064b60510181354v5a19253et8d7eea05eddb6d38@mail.gmail.com> Hi, I looked in Doc/INPUT_PW file to see what type of optimizations with constraints are available in PWSCF. It looks like the only option is to the keep distances between atoms fixed. Has anyone coded up other constraints, e.g. keeping the cell shape fixed (i.e. fcc, bcc, etc.) but varying the volume? Thanks, -- Nichols A. Romero, PhD. 1613 Denise Dr. Apt. D Forest Hill, MD 21050 443-567-8328 (C) 410-306-0709 (O) From zucco at dipteris.unige.it Wed Oct 19 14:08:14 2005 From: zucco at dipteris.unige.it (Marino Vetuschi Zuccolini) Date: Wed, 19 Oct 2005 14:08:14 +0200 Subject: [Pw_forum] Unit cell vibrations Message-ID: Hello to all, I've two questions for the audience: - can you suggest me how can I obtain the full set of vibrations related to a unit cell consisting of Z formulas ? - I'd like to depict the potential energy surface of a tetragonal crystal, what is the better way to do so? (I admit my pseudopotential are good..) Thanks in advance to all for your helps Marino ******************************************************* Marino Vetuschi Zuccolini zucco at dipteris.unige.it PhD / Geochemist Laboratory of Geochemistry http://www.dipteris.unige.it/geochimica Visit our BAXEICO computing cluster homepage http://qed.dipteris.unige.it/ganglia DIPartimento per lo studio della TErra e delle sue RISorse - Universit? di Genova Tel. ++39 010 3538136 Fax. ++39 010 352169 Corso Europa 26, 16132 - Genova - Italy -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 908 bytes Desc: not available Url : /pipermail/attachments/20051019/db43c484/attachment.bin From zjspam at gmail.com Wed Oct 19 16:46:22 2005 From: zjspam at gmail.com (Jian ZHOU) Date: Wed, 19 Oct 2005 22:46:22 +0800 Subject: [Pw_forum] GGA version of pseudo potential of Pb Message-ID: <8d38992a0510190746n6a6be1c8j4bef110bc975fd70@mail.gmail.com> Dear all, It seems that there does not have a GGA version of pseudo potential of element Pb (Atomic number: 82). Does any one have a pw91 GGA version? Thank you! regards, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051019/c923a2bf/attachment.htm From naromero at gmail.com Wed Oct 19 20:43:50 2005 From: naromero at gmail.com (Nichols A. Romero) Date: Wed, 19 Oct 2005 14:43:50 -0400 Subject: [Pw_forum] Nose Thermostat Message-ID: <6ac064b60510191143u62ea15em6b0ca2f34d3de535@mail.gmail.com> Hi, I searched the archive for Nose thermostat in PWSCF and found only a few references to it from back in November. Has anyone implemented the Nose thermostat in PWSCF? Bests. -- Nichols A. Romero, PhD. 1613 Denise Dr. Apt. D Forest Hill, MD 21050 443-567-8328 (C) 410-306-0709 (O) From konstantin_kudin at yahoo.com Wed Oct 19 21:20:34 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Wed, 19 Oct 2005 12:20:34 -0700 (PDT) Subject: [Pw_forum] Nose Thermostat In-Reply-To: <6ac064b60510191143u62ea15em6b0ca2f34d3de535@mail.gmail.com> Message-ID: <20051019192034.83960.qmail@web52002.mail.yahoo.com> It is in the CVS version of CP (not PWSCF). Kostya --- "Nichols A. Romero" wrote: > Hi, > > I searched the archive for Nose thermostat in PWSCF and found only a > few references to it from back in November. > > Has anyone implemented the Nose thermostat in PWSCF? > > Bests. > -- > Nichols A. Romero, PhD. > 1613 Denise Dr. Apt. D > Forest Hill, MD 21050 > 443-567-8328 (C) > 410-306-0709 (O) > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From stargmoon at yahoo.com Wed Oct 19 22:18:45 2005 From: stargmoon at yahoo.com (stargmoon) Date: Wed, 19 Oct 2005 13:18:45 -0700 (PDT) Subject: [Pw_forum] installation of version2.1.5 Message-ID: <20051019201845.71275.qmail@web33203.mail.mud.yahoo.com> Dear PWscf community, I am new in PWscf, this is the first time I install PWscf. My machine type is i686-pc-linux-gnu. There are more than one fortran compilers on it. If I simple run ./configure, I got the following messages: ------------------------------------------------- WARNING: the following problems were detected: unsupported C/Fortran compilers combination: CC=pgcc, F77=mpif77, F90=mpif90 you may not be able to compile this program ------------------------------------------------- I tried to specify the compiler by running ./compiler F90=g77 CC=gcc then, I got: ------------------------------------------------- WARNING: the following problems were detected: unsupported C/Fortran compilers combination: CC=gcc, F77=g77, F90=g77 you may not be able to compile this program -------------------------------------------------- Could you please let me know what is wrong and what I should do? Thanks in advance! Stargmoon __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From eyvaz_isaev at yahoo.com Wed Oct 19 23:40:12 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Wed, 19 Oct 2005 14:40:12 -0700 (PDT) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <20051019201845.71275.qmail@web33203.mail.mud.yahoo.com> Message-ID: <20051019214013.40193.qmail@web60315.mail.yahoo.com> Hi, > ------------------------------------------------- > WARNING: the following problems were detected: > > unsupported C/Fortran compilers combination: > CC=pgcc, F77=mpif77, F90=mpif90 > > you may not be able to compile this program This combination is unsuitable because parallel version of Fortran compilers are assumed to be used, but CC compiler is serial. It seems strange, may be your system is misconfigured. > ------------------------------------------------- > > I tried to specify the compiler by running > ./compiler > F90=g77 CC=gcc then, I got: > ------------------------------------------------- > WARNING: the following problems were detected: > > unsupported C/Fortran compilers combination: > CC=gcc, F77=g77, F90=g77 > g77 compiler is actually can not compile the code. You need F90 compiler, indeed. You can try ./configure CC=mpicc F77=mpif77 F90=mpif90 for parallel code. Otherwise, use pgcc, pgf77, and pgf90. Good luck, Eyvaz. __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ From giannozz at nest.sns.it Thu Oct 20 09:39:30 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 20 Oct 2005 09:39:30 +0200 Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <20051019201845.71275.qmail@web33203.mail.mud.yahoo.com> References: <20051019201845.71275.qmail@web33203.mail.mud.yahoo.com> Message-ID: <200510200939.30786.giannozz@nest.sns.it> On Wednesday 19 October 2005 22:18, stargmoon wrote: > If I simple run ./configure, I got the following messages: > ------------------------------------------------- > WARNING: the following problems were detected: > > unsupported C/Fortran compilers combination: > CC=pgcc, F77=mpif77, F90=mpif90 > > you may not be able to compile this program > ------------------------------------------------- did you try? the worst thing that can happen to you if you try and it doesn't work, is that you wasted up to 5' or so. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From alberto.milani at chem.polimi.it Thu Oct 20 11:47:19 2005 From: alberto.milani at chem.polimi.it (Alberto Milani) Date: Thu, 20 Oct 2005 11:47:19 +0200 Subject: [Pw_forum] About pseudopotential conversion Message-ID: <200510201147.19621.alberto.milani@chem.polimi.it> Hello, I have generated a NCPP for C using FHI98PP code and I used fhi2upf.x to convert it. I chose l_max=2 as maximum angular momentum to generate pseudopotential components. In FHI98PP tutorial and in CPC 119 (1999) 67-98 it is said that for the local potential "l_loc=l_max=2 is the common practice for sp-type materials..." So I set "l local (max=2) > 2" while running fhi2upf.x . In the .UPF pseudopotential my "Max angular momentum component" in <> is 1 instead of 2. What is happening? I took a look to fhi2upf.f90 and I noticed that the lmax_ variable is defined in some confusing ways (at least for me). I think that another source of confusion is due to the first line of FHI98PP files PP.cpi where the second element is not lmax but lmax+1... Anyway, what I have to do to avoid errors in conversion? Thanks in advance A. M. From giannozz at nest.sns.it Thu Oct 20 12:09:16 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 20 Oct 2005 12:09:16 +0200 Subject: [Pw_forum] About pseudopotential conversion In-Reply-To: <200510201147.19621.alberto.milani@chem.polimi.it> References: <200510201147.19621.alberto.milani@chem.polimi.it> Message-ID: <200510201209.16817.giannozz@nest.sns.it> On Thursday 20 October 2005 11:47, Alberto Milani wrote: > In the .UPF pseudopotential my "Max angular momentum component" > in <> is 1 instead of 2. correct: the UPF file contains PPs in separable (Kleinman-Bylander) form, so you have l=0 and l=1 projectors, l=2 as the local potential. P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From g.ballabio at cineca.it Thu Oct 20 14:44:02 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Thu, 20 Oct 2005 14:44:02 +0200 (MEST) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <20051019201845.71275.qmail@web33203.mail.mud.yahoo.com> (from stargmoon@yahoo.com on Wed Oct 19 22:18:45 2005) References: <20051019201845.71275.qmail@web33203.mail.mud.yahoo.com> Message-ID: <1129812240l.7418l.2l@nb-ballabio.cineca.it> On 10/19/2005 10:18:45 PM, stargmoon wrote: > ------------------------------------------------- > WARNING: the following problems were detected: > > unsupported C/Fortran compilers combination: > CC=pgcc, F77=mpif77, F90=mpif90 > > you may not be able to compile this program > ------------------------------------------------- "Unsupported" doesn't mean "doesn't work". It means "not known to work", but that could be just because nobody has tried yet (or someone has tried, but hasn't told us). Actually if your "mpif90" wraps pgf90, it will most likely work. Please try. > I tried to specify the compiler by running ./compiler > F90=g77 CC=gcc then, I got: g77 by contrast is known NOT to be able to compile Espresso. You need a real Fortran 90 compiler. Gerardo From stargmoon at yahoo.com Thu Oct 20 15:20:02 2005 From: stargmoon at yahoo.com (stargmoon) Date: Thu, 20 Oct 2005 06:20:02 -0700 (PDT) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <1129812240l.7418l.2l@nb-ballabio.cineca.it> Message-ID: <20051020132002.80609.qmail@web33211.mail.mud.yahoo.com> Thanks a lot, Gerardo, Paolo and Eyvaz for your reply. I tried to do "make all" by using combination "CC=pgcc, F77=mpif77, F90=mpif90", and "CC=mpicc, F77=mpif77, F90=mpif90", neither of them works. Could you please suggest me what should I do next? Another question, has anybody successfully complied PWscf with pathscale compiler? We have another cluster which has pathscale compiler. Best regards, Stargmoon --- Gerardo Ballabio wrote: > On 10/19/2005 10:18:45 PM, stargmoon wrote: > > ------------------------------------------------- > > WARNING: the following problems were detected: > > > > unsupported C/Fortran compilers combination: > > CC=pgcc, F77=mpif77, F90=mpif90 > > > > you may not be able to compile this program > > ------------------------------------------------- > > "Unsupported" doesn't mean "doesn't work". It means > "not known to > work", but that could be just because nobody has > tried yet (or > someone has tried, but hasn't told us). Actually if > your "mpif90" > wraps pgf90, it will most likely work. Please try. > > > I tried to specify the compiler by running > ./compiler > > F90=g77 CC=gcc then, I got: > > g77 by contrast is known NOT to be able to compile > Espresso. You need > a real Fortran 90 compiler. > > Gerardo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Music Unlimited Access over 1 million songs. Try it free. http://music.yahoo.com/unlimited/ From proffess at yandex.ru Thu Oct 20 16:47:26 2005 From: proffess at yandex.ru (Sergey Lisenkov) Date: Thu, 20 Oct 2005 18:47:26 +0400 (MSD) Subject: [Pw_forum] Compiling espresso-2.1.5 on SGI ALTIX 350 Message-ID: <4357ADFE.000006.22044@ariel.yandex.ru> Dear PWscf users and authors, I have troubles with compiling pwscf code on SGI altix 350 code. I took the make.sys file from install directory. We have scsl library and I choose shmem option for compiling. But during linking stage I got: ecl ../flib/ptools.a ../flib/flib.a ../clib/clib.a -L/sw/sgi/mpt/mpt-1.11/usr/lib/ -lsma -lscs /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_printf' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_global_rembias' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_globally_accessible' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_progress' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_comm_copy' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_fast_fence' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_symmetric_addr' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `_fastbcopy' /sw/sgi/mpt/mpt-1.11/usr/lib//libsma.so: undefined reference to `MPI_SGI_spmd_model' make[1]: *** [memory.x] Error 1 make[1]: Leaving directory `/home/inna/espresso-2.1.5/PW' make: *** [pw] Error 2 I used nm command and see that inside libsma.so there are all undefined routines without underscore: nm /sw/sgi/mpt/mpt-1.11/usr/lib/libsma.so | grep -i MPI_SGI_printf U MPI_SGI_printf So, how to solve this problem? Thanks a lot, Best wishes, Sergey From stargmoon at yahoo.com Thu Oct 20 16:49:12 2005 From: stargmoon at yahoo.com (stargmoon) Date: Thu, 20 Oct 2005 07:49:12 -0700 (PDT) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <1129812240l.7418l.2l@nb-ballabio.cineca.it> Message-ID: <20051020144912.19419.qmail@web33207.mail.mud.yahoo.com> Hello Gerardo, When I run "make all", I got the following error messages: ar: avrec.o: No such file or directory make[1]: *** [flib.a] Error 1 make[1]: Leaving directory `/home/gao/pwscf/espresso-2.1.5/flib' make: *** [libs] Error 2 However, I couldn't find what's wrong for the compilation of avrec.f90. Best, Stargmoon --- Gerardo Ballabio wrote: > On 10/19/2005 10:18:45 PM, stargmoon wrote: > > ------------------------------------------------- > > WARNING: the following problems were detected: > > > > unsupported C/Fortran compilers combination: > > CC=pgcc, F77=mpif77, F90=mpif90 > > > > you may not be able to compile this program > > ------------------------------------------------- > > "Unsupported" doesn't mean "doesn't work". It means > "not known to > work", but that could be just because nobody has > tried yet (or > someone has tried, but hasn't told us). Actually if > your "mpif90" > wraps pgf90, it will most likely work. Please try. > > > I tried to specify the compiler by running > ./compiler > > F90=g77 CC=gcc then, I got: > > g77 by contrast is known NOT to be able to compile > Espresso. You need > a real Fortran 90 compiler. > > Gerardo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From giannozz at nest.sns.it Thu Oct 20 16:54:28 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 20 Oct 2005 16:54:28 +0200 Subject: [Pw_forum] Compiling espresso-2.1.5 on SGI ALTIX 350 In-Reply-To: <4357ADFE.000006.22044@ariel.yandex.ru> References: <4357ADFE.000006.22044@ariel.yandex.ru> Message-ID: <200510201654.28445.giannozz@nest.sns.it> On Thursday 20 October 2005 16:47, Sergey Lisenkov wrote: > I have troubles with compiling pwscf code on SGI altix 350 code not really: you have troubles LINKING the pwscf code to SGI mathematical libraries P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From jmorrone at Princeton.EDU Thu Oct 20 16:54:39 2005 From: jmorrone at Princeton.EDU (Joseph Morrone) Date: Thu, 20 Oct 2005 10:54:39 -0400 Subject: [Pw_forum] units of electron-phonon matrix elements Message-ID: <83DC4262-DD3A-4296-B4D3-1C40F48724CC@Princeton.EDU> Hi, Can anyone tell me what the units are of the electron-phonon matrix elements (those that are calculated in subroutine "elphel")? Are they energy / distance (rydberg / bohr), as they are usually reported in the literature, or are they something else? your help is very much appreciated! Joe Morrone From giannozz at nest.sns.it Thu Oct 20 17:04:06 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 20 Oct 2005 17:04:06 +0200 Subject: [Pw_forum] units of electron-phonon matrix elements In-Reply-To: <83DC4262-DD3A-4296-B4D3-1C40F48724CC@Princeton.EDU> References: <83DC4262-DD3A-4296-B4D3-1C40F48724CC@Princeton.EDU> Message-ID: <200510201704.06296.giannozz@nest.sns.it> On Thursday 20 October 2005 16:54, Joseph Morrone wrote: > Can anyone tell me what the units are of the electron-phonon > matrix elements (those that are calculated in subroutine "elphel")? as a rule everything is in atomic (Ry) units unless otherwise stated P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From proffess at yandex.ru Thu Oct 20 18:51:59 2005 From: proffess at yandex.ru (Sergey Lisenkov) Date: Thu, 20 Oct 2005 20:51:59 +0400 (MSD) Subject: [Pw_forum] Compiling espresso-2.1.5 on SGI ALTIX 350 In-Reply-To: <200510201654.28445.giannozz@nest.sns.it> References: <4357ADFE.000006.22044@ariel.yandex.ru> <200510201654.28445.giannozz@nest.sns.it> Message-ID: <4357CB2F.000008.03877@pantene.yandex.ru> Dear Paolo, I am a bit confused, because when I linked with Intel MKL libraries, the errors were the same. I think these errors come from using SHMEM cpp flag. What is exatly the problem? in libsma.so library? or in SCSL library? How to solve it? Thanks a lot, Sergey From eyvaz_isaev at yahoo.com Thu Oct 20 22:01:56 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Thu, 20 Oct 2005 13:01:56 -0700 (PDT) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <20051020132002.80609.qmail@web33211.mail.mud.yahoo.com> Message-ID: <20051020200156.23332.qmail@web60313.mail.yahoo.com> Hi, So strange. How about "./configure.old pc_pgi" if you prefer a serial, or "./configure.old beowulf" for parallel (PGI compiler, Linux PC cluster) executable. Bests, Eyvaz. --- stargmoon wrote: > Thanks a lot, Gerardo, Paolo and Eyvaz for your > reply. > I tried to do "make all" by using combination > "CC=pgcc, F77=mpif77, F90=mpif90", and "CC=mpicc, > F77=mpif77, F90=mpif90", neither of them works. > Could > you please suggest me what should I do next? > > Another question, has anybody successfully complied > PWscf with pathscale compiler? We have another > cluster > which has pathscale compiler. > > Best regards, > > Stargmoon > > --- Gerardo Ballabio wrote: > > > On 10/19/2005 10:18:45 PM, stargmoon wrote: > > > > ------------------------------------------------- > > > WARNING: the following problems were detected: > > > > > > unsupported C/Fortran compilers combination: > > > CC=pgcc, F77=mpif77, F90=mpif90 > > > > > > you may not be able to compile this program > > > > ------------------------------------------------- > > > > "Unsupported" doesn't mean "doesn't work". It > means > > "not known to > > work", but that could be just because nobody has > > tried yet (or > > someone has tried, but hasn't told us). Actually > if > > your "mpif90" > > wraps pgf90, it will most likely work. Please try. > > > > > I tried to specify the compiler by running > > ./compiler > > > F90=g77 CC=gcc then, I got: > > > > g77 by contrast is known NOT to be able to compile > > Espresso. You need > > a real Fortran 90 compiler. > > > > Gerardo > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > > __________________________________ > Yahoo! Music Unlimited > Access over 1 million songs. Try it free. > http://music.yahoo.com/unlimited/ > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From stargmoon at yahoo.com Thu Oct 20 23:54:24 2005 From: stargmoon at yahoo.com (stargmoon) Date: Thu, 20 Oct 2005 14:54:24 -0700 (PDT) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <20051020200156.23332.qmail@web60313.mail.yahoo.com> Message-ID: <20051020215424.47902.qmail@web33208.mail.mud.yahoo.com> Hi Eyvaz, Thanks a lot. By using "./configure.old beowulf", I successfully compiled version2.1.5. But I don't know whether this will affect the performance or there will be some bugs when I run the code. Thanks again! Stargmoon --- Eyvaz Isaev wrote: > Hi, > > So strange. How about "./configure.old pc_pgi" if > you > prefer a serial, or "./configure.old beowulf" for > parallel (PGI compiler, Linux PC cluster) > executable. > > Bests, > Eyvaz. > > --- stargmoon wrote: > > > Thanks a lot, Gerardo, Paolo and Eyvaz for your > > reply. > > I tried to do "make all" by using combination > > "CC=pgcc, F77=mpif77, F90=mpif90", and "CC=mpicc, > > F77=mpif77, F90=mpif90", neither of them works. > > Could > > you please suggest me what should I do next? > > > > Another question, has anybody successfully > complied > > PWscf with pathscale compiler? We have another > > cluster > > which has pathscale compiler. > > > > Best regards, > > > > Stargmoon > > > > --- Gerardo Ballabio wrote: > > > > > On 10/19/2005 10:18:45 PM, stargmoon wrote: > > > > > > ------------------------------------------------- > > > > WARNING: the following problems were detected: > > > > > > > > unsupported C/Fortran compilers combination: > > > > CC=pgcc, F77=mpif77, F90=mpif90 > > > > > > > > you may not be able to compile this program > > > > > > ------------------------------------------------- > > > > > > "Unsupported" doesn't mean "doesn't work". It > > means > > > "not known to > > > work", but that could be just because nobody has > > > tried yet (or > > > someone has tried, but hasn't told us). Actually > > if > > > your "mpif90" > > > wraps pgf90, it will most likely work. Please > try. > > > > > > > I tried to specify the compiler by running > > > ./compiler > > > > F90=g77 CC=gcc then, I got: > > > > > > g77 by contrast is known NOT to be able to > compile > > > Espresso. You need > > > a real Fortran 90 compiler. > > > > > > Gerardo > > > > > > _______________________________________________ > > > Pw_forum mailing list > > > Pw_forum at pwscf.org > > > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > > > > > > > > __________________________________ > > Yahoo! Music Unlimited > > Access over 1 million songs. Try it free. > > http://music.yahoo.com/unlimited/ > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > > > __________________________________ > Yahoo! Mail - PC Magazine Editors' Choice 2005 > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com From g.ballabio at cineca.it Fri Oct 21 11:51:17 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Fri, 21 Oct 2005 11:51:17 +0200 (MEST) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <20051020215424.47902.qmail@web33208.mail.mud.yahoo.com> (from stargmoon@yahoo.com on Thu Oct 20 23:54:24 2005) References: <20051020215424.47902.qmail@web33208.mail.mud.yahoo.com> Message-ID: <1129888276l.9272l.0l@nb-ballabio.cineca.it> On 10/20/2005 11:54:24 PM, stargmoon wrote: > Hi Eyvaz, > > Thanks a lot. By using "./configure.old beowulf", I > successfully compiled version2.1.5. I'd really like to see the differences between the working "make.sys" and the configure-generated one. Gerardo From stargmoon at yahoo.com Fri Oct 21 17:01:52 2005 From: stargmoon at yahoo.com (stargmoon) Date: Fri, 21 Oct 2005 08:01:52 -0700 (PDT) Subject: [Pw_forum] installation of version2.1.5 In-Reply-To: <1129888276l.9272l.0l@nb-ballabio.cineca.it> Message-ID: <20051021150152.46124.qmail@web33203.mail.mud.yahoo.com> Hi Gerardo, I attached the two make.sys files here, make.sys_1 is the configure-generated "make.sys" and make.sys_2 is the working one. Best, Stargmoon --- Gerardo Ballabio wrote: > On 10/20/2005 11:54:24 PM, stargmoon wrote: > > Hi Eyvaz, > > > > Thanks a lot. By using "./configure.old beowulf", > I > > successfully compiled version2.1.5. > > I'd really like to see the differences between the > working "make.sys" > and the configure-generated one. > > Gerardo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com -------------- next part -------------- A non-text attachment was scrubbed... Name: make.sys_1 Type: application/octet-stream Size: 1590 bytes Desc: 1074575357-make.sys_1 Url : /pipermail/attachments/20051021/0455d550/attachment.obj -------------- next part -------------- A non-text attachment was scrubbed... Name: make.sys_2 Type: application/octet-stream Size: 1794 bytes Desc: 2901908261-make.sys_2 Url : /pipermail/attachments/20051021/0455d550/attachment-0001.obj From giannozz at nest.sns.it Fri Oct 21 17:15:29 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 21 Oct 2005 17:15:29 +0200 Subject: [Pw_forum] Compiling espresso-2.1.5 on SGI ALTIX 350 In-Reply-To: <4357CB2F.000008.03877@pantene.yandex.ru> References: <4357ADFE.000006.22044@ariel.yandex.ru> <200510201654.28445.giannozz@nest.sns.it> <4357CB2F.000008.03877@pantene.yandex.ru> Message-ID: <200510211715.29684.giannozz@nest.sns.it> On Thursday 20 October 2005 18:51, Sergey Lisenkov wrote: > I am a bit confused, because when I linked with Intel MKL > libraries, the errors were the same. I think these errors > come from using SHMEM cpp flag. the messages you showed say that something (apparently related to MPI) is missing to libsma.so, required by -lsma > What is exatly the problem? in libsma.so library? or in SCSL > library? How to solve it? inquire with your local SGI wizard (or become one :-) ) P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From cbarreteau at cea.fr Fri Oct 21 17:49:08 2005 From: cbarreteau at cea.fr (Cyrille Barreteau) Date: Fri, 21 Oct 2005 17:49:08 +0200 (CEST) Subject: [Pw_forum] mixing in non-collinear case In-Reply-To: <200510201209.16817.giannozz@nest.sns.it> References: <200510201147.19621.alberto.milani@chem.polimi.it> <200510201209.16817.giannozz@nest.sns.it> Message-ID: <1169.81.56.70.96.1129909748.squirrel@dsm-mail> Dear pwscf users and developpers, I was wondering what kind of mixing is used in the non collinear version of pwscf (pwnc) And more importantly which quantity is mixed? i) the density matrix rho=(rho up up, rho up dn rho dn up, rho dn dn) ii) the magnetic vector in xyz coordinate n,mx,my,mz or iii) the magnetic vector in spherical coordinate n,|m|,theta,phi The case iii) seems rather attractive but I guess one should not treat n,|m| on the same footing as theta and phi. cyrille -- ================================================================== Cyrille Barreteau | phone : +33 (0)1 69 08 29 51 CEA Saclay | fax : +33 (0)1 69 08 84 46 DSM/DRECAM/SPCSI | email : cyrille.barreteau at cea.fr Batiment 462 | 91191 Gif sur Yvette Cedex FRANCE ~~~~~~~~~~~~~~~~~~~~~~~~ http://www-drecam.cea.fr/spcsi/index.php http://www-drecam.cea.fr/Images/Pisp/cbarreteau/cbarreteau_fr.html http://www-drecam.cea.fr/Phocea/Membres/Cours/index.php ================================================================== From jmorrone at Princeton.EDU Fri Oct 21 21:02:33 2005 From: jmorrone at Princeton.EDU (Joseph Morrone) Date: Fri, 21 Oct 2005 15:02:33 -0400 Subject: [Pw_forum] units of electron-phonon matrix elements Message-ID: <4E9A6CB2-CDA3-43C4-9E55-3E5002D987EC@princeton.edu> Thank you very much for your reply. What I was also curious about was whether or not this matrix element that is calculated by the subroutine "elphel" is in units of energy / distance or if it includes other units such as the mass, which are present in the eigenmodes during some phases of the code's operation. It seems that upon an inspection of the subroutine "elphsum," which uses these matrix elements, that this may be the case. your help is very much appreciated! Joe Morrone From leoant21 at hotmail.com Mon Oct 24 08:53:04 2005 From: leoant21 at hotmail.com (=?gb2312?B?1cUguumx8g==?=) Date: Mon, 24 Oct 2005 06:53:04 +0000 Subject: [Pw_forum] Convergence problems on supercell calculation of MgB_2 Message-ID: Dear all: I am now doing some calculation of 2?2?1 supercell of MgB_2 without any atom moved, however, after 100 iterations the package tell me that the convergence has NOT been achieved, while with 3?3?1 supercell I had successfully got scf caculation done. The E_thr is 1.0E-8, while the smallest estimated energy accuracy is 2.0E-8. The package I used for 2?2 caculation is the latest version, while 3?3?s is not. Is there any comment on this and advice on how to improve the calculation? Thank you. Regards. _________________________________________________________________ ?????????????? MSN Messenger: http://messenger.msn.com/cn From liyanpcl at yahoo.com.cn Tue Oct 25 10:43:09 2005 From: liyanpcl at yahoo.com.cn (li yan) Date: Tue, 25 Oct 2005 16:43:09 +0800 (CST) Subject: [Pw_forum] questions about Raman tensor Message-ID: <20051025084309.71028.qmail@web15608.mail.cnb.yahoo.com> dear all, i want to follow the example15 to study how to calculate the Raman intensities. i have three questions about that. thank you for your answers to any of the three questions. (1) i cannot understand the selection "b_length", could you please give me some papers about that. (2) whether the nonlinear susceptibility is defined as the first order derivative of the polarization with respect to the displacements of the atoms. (3) in the file alas.ram, to every atom, there are three 3*3 matrix. whether these matrixes are defined as the Taylor expansion of the polarization with respect to the displacements of the atoms. --------------------------------- ?????????????????????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051025/4416415d/attachment.htm From zjspam at gmail.com Tue Oct 25 14:28:33 2005 From: zjspam at gmail.com (Jian ZHOU) Date: Tue, 25 Oct 2005 20:28:33 +0800 Subject: [Pw_forum] k point parallel Message-ID: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> Dear all It is said in the manual that pwscf can be paralleled by "k points parallelization" or by " PW parallelization". Since in our Gigabit ethernet, the parallelization performances is not so good, so, is it possible for pwscf to parallel only on the k points, not on the PW? Thank you. Best wishes, Jian -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051025/1eb0b5a1/attachment.htm From akohlmey at vitae.cmm.upenn.edu Tue Oct 25 14:55:20 2005 From: akohlmey at vitae.cmm.upenn.edu (Axel Kohlmeyer) Date: Tue, 25 Oct 2005 08:55:20 -0400 Subject: [Pw_forum] k point parallel In-Reply-To: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> Message-ID: <435E2B38.4020507@cmm.upenn.edu> Jian ZHOU wrote: > Dear all dear jian, > > It is said in the manual that pwscf can be paralleled by "k points > parallelization" or by " PW parallelization". Since in our Gigabit > ethernet, the parallelization performances is not so good, so, is it > possible for pwscf to parallel only on the k points, not on the PW? yes. please have a look at the manual. you need to give the additional flag -npool to the executables. this decides, how many k-point pools will be used. note that this should be a divisor of the number of cpus requested and the number of k-points. for optimal performance, you also need a _local_ scratch area on each node (and specify that in your input) and set the keyword wf_collect=.true. so that the final restart is collected to a shared directory, or else you would most likely not be able to restart or continue with the next job, as you would need to have the exact same nodes in the exact same order assigned for that job. there are more details about this in the manual and Doc/INPUT_PW. this also has been discussed before, so have a look at the mailing list archives as well (you can even (ab-)use google for that, by adding site:www.democritos.it to your search). regards, axel. > > Thank you. > > Best wishes, > > Jian > > -- ======================================================================= Axel Kohlmeyer e-mail: akohlmey at cmm.upenn.edu, tel: ++1-215-898-1582 Center for Molecular Modeling -- University of Pennsylvania Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323 ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From Giovanni.Cantele at na.infn.it Tue Oct 25 14:57:09 2005 From: Giovanni.Cantele at na.infn.it (Giovanni Cantele) Date: Tue, 25 Oct 2005 14:57:09 +0200 Subject: [Pw_forum] k point parallel In-Reply-To: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> Message-ID: <435E2BA5.9020001@na.infn.it> Jian ZHOU wrote: > Dear all > > It is said in the manual that pwscf can be paralleled by "k points > parallelization" or by " PW parallelization". Since in our Gigabit > ethernet, the parallelization performances is not so good, so, is it > possible for pwscf to parallel only on the k points, not on the PW? > > Thank you. > > Best wishes, > > Jian > > Try to use pools (for k-point) parallelization (see http://www.pwscf.org/guide/2.1.4/html-single/manual.html#SECTION00040000000000000000 for running on parallel machines ) For example, running on 16 cpus with no pools the beginning of the output looks as follows: Program PWSCF v.3.0 starts ... Today is 25Oct2005 at 14:49:47 Parallel version (MPI) Number of processors in use: 16 R & G space division: proc/pool = 16 Ultrasoft (Vanderbilt) Pseudopotentials Current dimensions of program pwscf are: ntypx = 10 npk = 40000 lmax = 3 nchix = 6 ndmx = 2000 nbrx = 14 nqfx = 8 Planes per process (thick) : nr3 = 96 npp = 6 ncplane = 9216 Proc/ planes cols G planes cols G columns G Pool (dense grid) (smooth grid) (wavefct grid) 1 6 447 28493 6 447 28493 119 3935 2 6 448 28494 6 448 28494 119 3935 3 6 447 28491 6 447 28491 119 3935 4 6 447 28491 6 447 28491 119 3935 5 6 447 28491 6 447 28491 119 3935 6 6 447 28491 6 447 28491 119 3933 7 6 447 28491 6 447 28491 119 3933 8 6 447 28491 6 447 28491 119 3933 9 6 447 28493 6 447 28493 119 3935 10 6 447 28491 6 447 28491 119 3935 11 6 447 28491 6 447 28491 118 3930 12 6 447 28491 6 447 28491 118 3930 13 6 447 28491 6 447 28491 118 3930 14 6 447 28491 6 447 28491 119 3933 15 6 447 28491 6 447 28491 119 3933 16 6 447 28491 6 447 28491 119 3933 0 96 7153 455863 96 7153 455863 1901 62933 .... in which only PW parallelization occurred. If you run with pools, within each pool the R & G space division occurs on a number of processors which is the total number divided by the number of pools. So, if you use as many pools as the total number of cpus non PW parallelization should occur. For example running on 16 cpus with 16 pools you find Program PWSCF v.3.0 starts ... Today is 25Oct2005 at 14:47:28 Parallel version (MPI) Number of processors in use: 16 K-points division: npool = 16 Ultrasoft (Vanderbilt) Pseudopotentials Current dimensions of program pwscf are: ntypx = 10 npk = 40000 lmax = 3 nchix = 6 ndmx = 2000 nbrx = 14 nqfx = 8 Planes per process (thick) : nr3 = 96 npp = 96 ncplane = 9216 Proc/ planes cols G planes cols G columns G Pool (dense grid) (smooth grid) (wavefct grid) 1 96 7153 455863 96 7153 455863 1901 62933 0 96 7153 455863 96 7153 455863 1901 62933 ... and you see that there is one cpu for each pool and so no parallelization on PWs can be done. Giovanni -- Dr. Giovanni Cantele Coherentia CNR-INFM and Dipartimento di Scienze Fisiche Universita' di Napoli "Federico II" Complesso Universitario di Monte S. Angelo - Ed. G Via Cintia, I-80126, Napoli, Italy Phone: +39 081 676910 Fax: +39 081 676346 E-mail: Giovanni.Cantele at na.infn.it Web: http://people.na.infn.it/~cantele **************** RICERCA E FUTURO CAMMINANO INSIEME: DA 41 MESI I LAVORATORI DELLA RICERCA HANNO IL CONTRATTO SCADUTO, IL FUTURO NON DEVE SCADERE! per maggiori info: http://www.ge.cnr.it/rsu/rsu.htm **************** From mverissi at ictp.it Tue Oct 25 14:57:36 2005 From: mverissi at ictp.it (Marcos Verissimo Alves) Date: Tue, 25 Oct 2005 14:57:36 +0200 (CEST) Subject: [Pw_forum] Magnetism in carbon structures - technical issues In-Reply-To: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> Message-ID: <36534.10.50.40.120.1130245056.squirrel@webmail1.ictp.trieste.it> Hi all, Maybe this is off-topic, but... I want to perform some studies on magnetism on carbon structures using pwscf. In the pdf file included with the espresso package, regarding the generation of pseudopotentials (PSP for short), it is mentioned that PSP for magnetic system must include nonlinear core corrections (NLCC for laziness). My questions are: 1) Is this really mandatory for all magnetic systems? And in the particular case of Carbon and Oxygen, would NLCC be mandatory, too? 2) Instead of nonlinear core corrections, could the 1s electrons be included in Vanderbilt or RKKJ ultrasoft PSPs? Would the computational cost be much larger in this case, as compared to NLCC? This study would require up to about ~128 atoms, Oxygen included. 3) Would anyone have good USPPs for Carbon and Oxygen, using NLCC and GGA-PBE - or be so kind as to send me an input file to generate them? Troullier-Martins PSPs I can generate, but I don't know much (actually, almost nothing) about USPP. Thanks in advance, Marcos -- Dr. Marcos Verissimo Alves Post-Doctoral Fellow - Condensed Matter and Statistical Physics Group International Centre for Theoretical Physics Strada Costiera, 11 Room 230, Main Building Trieste, Italy - 34014 Telephone: +39 040 2240 357 -------------------- "So long and thanks for all the fish." - Quoted from the dolphins of "The Hitchhiker's Guide to the Galaxy" From zjspam at gmail.com Tue Oct 25 16:18:26 2005 From: zjspam at gmail.com (Jian ZHOU) Date: Tue, 25 Oct 2005 22:18:26 +0800 Subject: [Pw_forum] Re: k point parallel In-Reply-To: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> Message-ID: <8d38992a0510250718p66eea6f3t27cf064ce53ab1f9@mail.gmail.com> Thank you axel and Giovanni. On 10/25/05, Jian ZHOU wrote: > > Dear all > > It is said in the manual that pwscf can be paralleled by "k points > parallelization" or by " PW parallelization". Since in our Gigabit ethernet, > the parallelization performances is not so good, so, is it possible for > pwscf to parallel only on the k points, not on the PW? > > Thank you. > > Best wishes, > > Jian > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051025/f806a321/attachment.htm From konstantin_kudin at yahoo.com Tue Oct 25 20:05:37 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Tue, 25 Oct 2005 11:05:37 -0700 (PDT) Subject: [Pw_forum] k point parallel In-Reply-To: <435E2B38.4020507@cmm.upenn.edu> Message-ID: <20051025180537.81976.qmail@web52008.mail.yahoo.com> > yes. please have a look at the manual. you need to give the > additional > flag -npool to the executables. this decides, how many k-point pools > will be used. note that this should be a divisor of the number of > cpus > requested and the number of k-points. for optimal performance, you > also need a _local_ scratch area on each node (and specify that > in your input) and set the keyword wf_collect=.true. so that the > final restart is collected to a shared directory, or else you would > most likely not be able to restart or continue with the next job, as > you would need to have the exact same nodes in the exact same order > assigned for that job. Axel, Good tips! Also note that most PWSCF jobs can be restarted from 'rho' files alone without 'wfc' files (with very minor extra cpu time), thus saving the effort to collect and manage the wavefunction file. I would also like to add that the upcoming "official" version of Espresso will allow to specify separate directories for 'wfc' files. This way they can go to local nodes, while the other restart info goes to some other directory perhaps even using NFS. Kostya __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From TAOJ0003 at ntu.edu.sg Wed Oct 26 14:02:03 2005 From: TAOJ0003 at ntu.edu.sg (#TAO JUNGUANG#) Date: Wed, 26 Oct 2005 20:02:03 +0800 Subject: [Pw_forum] why I can not get the correct band gap of TiO2? Message-ID: <34C4FA35021357469685D8102806A1C002905B96@mail01.student.main.ntu.edu.sg> My input file is as following, &CONTROL calculation='nscf', restart_mode='from_scratch', pseudo_dir='/home/tao/pseudo/', tstress=.true. tprnfor=.true. / &SYSTEM ibrav=6, a=4.594,c=2.958, nat=6,ntyp=2, ecutwfc=16, / &ELECTRONS conv_thr=1.0d-8, mixing_beta=0.7, mixing_mode='plain', / &IONS ion_dynamics='bfgs', upscale=10 / ATOMIC_SPECIES Ti 47.867 Ti.vdb.UPF O 15.9994 O.vdb.UPF ATOMIC_POSITIONS crystal Ti 0.00 0.00 0.00 Ti 0.5 0.5 0.5 O 0.3053 0.3053 0.00 O 0.6947 0.6947 0.00 O 0.1947 0.8053 0.50 O 0.8053 0.1947 0.50 K_POINTS 36 0.5 0.5 0.5 1 0.4 0.4 0.4 1 0.3 0.3 0.3 1 0.2 0.2 0.2 1 0.1 0.1 0.1 1 0.0 0.0 0.0 1 0.0 0.0 0.1 1 0.0 0.0 0.2 1 0.0 0.0 0.3 1 0.0 0.0 0.4 1 0.0 0.0 0.5 1 0.0 0.0 0.6 1 0.0 0.0 0.7 1 0.0 0.0 0.8 1 0.0 0.0 0.9 1 0.0 0.0 1.0 1 0.0 0.1 1.0 1 0.0 0.2 1.0 1 0.0 0.3 1.0 1 0.0 0.4 1.0 1 0.0 0.5 1.0 1 0.0 0.6 1.0 1 0.0 0.7 1.0 1 0.0 0.8 1.0 1 0.0 0.9 1.0 1 0.0 1.0 1.0 1 0.0 0.9 0.9 1 0.0 0.8 0.8 1 0.0 0.7 0.7 1 0.0 0.6 0.6 1 0.0 0.5 0.5 1 0.0 0.4 0.4 1 0.0 0.3 0.3 1 0.0 0.2 0.2 1 0.0 0.1 0.1 1 0.0 0.0 0.0 1 And the output file is: ....... .... End of band structure calculation k = 0.5000 0.5000 0.5000 band energies (ev): -48.0011 -47.9986 -24.3665 -24.3599 -23.9018 -23.9017 -23.6774 -23.6773 -9.6133 -9.6128 -9.5064 -9.5063 3.0737 3.0742 3.4060 3.4062 4.6156 4.6161 5.5717 5.5719 6.6187 6.6188 6.7179 6.7182 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051026/bea2ebd0/attachment.htm From eyvaz_isaev at yahoo.com Wed Oct 26 14:38:29 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Wed, 26 Oct 2005 05:38:29 -0700 (PDT) Subject: [Pw_forum] why I can not get the correct band gap of TiO2? In-Reply-To: <34C4FA35021357469685D8102806A1C002905B96@mail01.student.main.ntu.edu.sg> Message-ID: <20051026123829.93707.qmail@web60323.mail.yahoo.com> Hi, First, some technical comments. > > ecutwfc=16, It looks very small, use 30Ry. Besides add ecutrho, say, 300Ry, for the augmented charge. > > &IONS > > ion_dynamics='bfgs', > > upscale=10 > > / For the band gap problem you do not need ionic dynamics. > > K_POINTS > > 36 K-points you used looks strange. How did you generate these points? They have equal weight and seem to be k-points along high symmetry directions. Use K_POINTS {automatic}. Now about your question. DFT is known to give underestimated and sometime wrong band gap for semiconductors and insulators. You can improve it using the LDA+U approximation. As far as I am aware Exact Exchange is still in development in Espresso. Bests, Eyvaz. __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com From TAOJ0003 at ntu.edu.sg Wed Oct 26 16:19:07 2005 From: TAOJ0003 at ntu.edu.sg (#TAO JUNGUANG#) Date: Wed, 26 Oct 2005 22:19:07 +0800 Subject: [Pw_forum] why I can not get the correct band gap of TiO2? References: <20051026123829.93707.qmail@web60323.mail.yahoo.com> Message-ID: <34C4FA35021357469685D8102806A1C00203DB24@mail01.student.main.ntu.edu.sg> Very thanks for your help. I have revised my Input file as you said, but from the output file, ........ End of band structure calculation k = 0.0000 0.0000 0.0000 band energies (ev): -48.0496 -48.0314 -24.6221 -24.6221 -24.5729 -24.5342 -24.4355 -24.4355 -10.8942 -9.9348 -9.1108 -9.1108 1.4602 1.7866 2.9518 3.3254 3.3254 4.5197 4.5197 4.9432 5.9659 6.2973 6.2973 6.9316 from the output file, I think the band gap is (1.4602-(-9.1108)=10.571eV),first, am I right getting band gap like this? Then, if that is the case, the band gap is very large comparing with the experiment value 3.2eV. (Additionally, other people hav givin a value of 2.0eV using pwscf, which is more reasonable than mine). Yes, maybe pwscf is not the best way to get the band gap of TiO2, however my results is very strange, so, I just want to figure out what wrong . Thanks! ________________________________ From: pw_forum-admin at pwscf.org on behalf of Eyvaz Isaev Sent: Wed 10/26/2005 8:38 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] why I can not get the correct band gap of TiO2? Hi, First, some technical comments. > > ecutwfc=16, It looks very small, use 30Ry. Besides add ecutrho, say, 300Ry, for the augmented charge. > > &IONS > > ion_dynamics='bfgs', > > upscale=10 > > / For the band gap problem you do not need ionic dynamics. > > K_POINTS > > 36 K-points you used looks strange. How did you generate these points? They have equal weight and seem to be k-points along high symmetry directions. Use K_POINTS {automatic}. Now about your question. DFT is known to give underestimated and sometime wrong band gap for semiconductors and insulators. You can improve it using the LDA+U approximation. As far as I am aware Exact Exchange is still in development in Espresso. Bests, Eyvaz. __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/ms-tnef Size: 6163 bytes Desc: not available Url : /pipermail/attachments/20051026/8102b106/attachment.bin From g.ballabio at cineca.it Wed Oct 26 17:20:13 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Wed, 26 Oct 2005 17:20:13 +0200 (MEST) Subject: [Pw_forum] why I can not get the correct band gap of TiO2? In-Reply-To: <34C4FA35021357469685D8102806A1C00203DB24@mail01.student.main.ntu.edu.sg> (from TAOJ0003@ntu.edu.sg on Wed Oct 26 16:19:07 2005) References: <20051026123829.93707.qmail@web60323.mail.yahoo.com> <34C4FA35021357469685D8102806A1C00203DB24@mail01.student.main.ntu.edu.sg> Message-ID: <1130340012l.4450l.0l@nb-ballabio.cineca.it> On 10/26/2005 04:19:07 PM, #TAO JUNGUANG# wrote: > End of band structure calculation > k = 0.0000 0.0000 0.0000 band energies (ev): > -48.0496 -48.0314 -24.6221 -24.6221 -24.5729 -24.5342 -24.4355 > -24.4355 > -10.8942 -9.9348 -9.1108 -9.1108 1.4602 1.7866 2.9518 > 3.3254 > 3.3254 4.5197 4.5197 4.9432 5.9659 6.2973 6.2973 > 6.9316 > from the output file, I think the band gap is > (1.4602-(-9.1108)=10.571eV),first, am I right getting band gap like > this? No. You aren't getting any band gap unless you calculate at least one empty band. In your input "nbnd" isn't specified, thus the default number of bands is computed -- that is, only occupied ones. Indeed, O has 6 valence electrons, Ti has 12 (assuming that semicore 3s and 3p are included). You have 4 O + 2 Ti = 48 electrons in your cell, at 2 electrons per band, this makes exactly 24 bands. You must specify an "nbnd" of at least 25. Gerardo From wmbmacam at lg.ehu.es Wed Oct 26 17:25:19 2005 From: wmbmacam at lg.ehu.es (Miguel Martinez) Date: Wed, 26 Oct 2005 17:25:19 +0200 Subject: [Pw_forum] why I can not get the correct band gap of TiO2? In-Reply-To: <34C4FA35021357469685D8102806A1C00203DB24@mail01.student.main.ntu.edu.sg> References: <20051026123829.93707.qmail@web60323.mail.yahoo.com> <34C4FA35021357469685D8102806A1C00203DB24@mail01.student.main.ntu.edu.sg> Message-ID: <435F9FDF.40408@lg.ehu.es> #TAO JUNGUANG# wrote: > End of band structure calculation > k = 0.0000 0.0000 0.0000 band energies (ev): [...] > I think the band gap is (1.4602-(-9.1108)=10.571eV) Hmmm... maybe I am wrong, but I would think this is the gap... AT (0,0,0)!!! You should plot the bands in several directions to see what the gap is. It might well be an indirect gap (the valence band maxima and the conduction band minima occur at different q_points), and, as such, only observable in a plot or in a very careful analysis. Hope it helps, Miguel -- ---------------------------------------- Miguel Mart?nez Canales Dto. F?sica de la Materia Condensada UPV/EHU Facultad de Ciencia y Tecnolog?a Apdo. 644 48080 Bilbao (Spain) Fax: +34 94 601 3500 Tlf: +34 94 601 5437 ---------------------------------------- 2005 International Year of Physics "But you don't ask with respect. You don't offer friendship. You don't even think to call me Godfather." Don Vito Corleone From TAOJ0003 at ntu.edu.sg Wed Oct 26 18:03:22 2005 From: TAOJ0003 at ntu.edu.sg (#TAO JUNGUANG#) Date: Thu, 27 Oct 2005 00:03:22 +0800 Subject: [Pw_forum] =?gb2312?B?tPC4tDogW1B3X2ZvcnVtXSB3aHkgSSBjYW4gbm90IGdldCB0aGUgYw==?= =?gb2312?B?b3JyZWN0IGJhbmQgZ2FwIG9mIFRpTzI/?= Message-ID: <34C4FA35021357469685D8102806A1C002905B97@mail01.student.main.ntu.edu.sg> Dear Ballabio, Thanks a lot!! Based on your suggestion, I make an another run and get a reasonable results. However, why in the valance band , there has a huge gap?( see the values -48.0314 and -24.6221) ,and can you tell how to use bands.x and forthmore band_plot.x to plot the band structure? Thanks Regards Tao Junguang -----????----- ???: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] ?? Gerardo Ballabio ????: 2005?10?26? 23:20 ???: pw_forum at pwscf.org ??: Re: [Pw_forum] why I can not get the correct band gap of TiO2? On 10/26/2005 04:19:07 PM, #TAO JUNGUANG# wrote: > End of band structure calculation > k = 0.0000 0.0000 0.0000 band energies (ev): > -48.0496 -48.0314 -24.6221 -24.6221 -24.5729 -24.5342 -24.4355 > -24.4355 > -10.8942 -9.9348 -9.1108 -9.1108 1.4602 1.7866 2.9518 > 3.3254 > 3.3254 4.5197 4.5197 4.9432 5.9659 6.2973 6.2973 > 6.9316 > from the output file, I think the band gap is > (1.4602-(-9.1108)=10.571eV),first, am I right getting band gap like > this? No. You aren't getting any band gap unless you calculate at least one empty band. In your input "nbnd" isn't specified, thus the default number of bands is computed -- that is, only occupied ones. Indeed, O has 6 valence electrons, Ti has 12 (assuming that semicore 3s and 3p are included). You have 4 O + 2 Ti = 48 electrons in your cell, at 2 electrons per band, this makes exactly 24 bands. You must specify an "nbnd" of at least 25. Gerardo _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From TAOJ0003 at ntu.edu.sg Wed Oct 26 18:10:30 2005 From: TAOJ0003 at ntu.edu.sg (#TAO JUNGUANG#) Date: Thu, 27 Oct 2005 00:10:30 +0800 Subject: [Pw_forum] =?gb2312?B?tPC4tDogW1B3X2ZvcnVtXSB3aHkgSSBjYW4gbm90IGdldCB0aGUgYw==?= =?gb2312?B?b3JyZWN0IGJhbmQgZ2FwIG9mIFRpTzI/?= Message-ID: <34C4FA35021357469685D8102806A1C002905B98@mail01.student.main.ntu.edu.sg> Dear Martinez, Thanks for suggestion. The gap of TiO2 is occurred at (0 0 0) , so I just take a rough analysis at this point. By the way, do you know how to plot out the band structure based on the output file of pwscf? Thx in advance. Regards! Tao Junguang -----????----- ???: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] ?? Miguel Martinez ????: 2005?10?26? 23:25 ???: pw_forum at pwscf.org ??: Re: [Pw_forum] why I can not get the correct band gap of TiO2? #TAO JUNGUANG# wrote: > End of band structure calculation > k = 0.0000 0.0000 0.0000 band energies (ev): [...] > I think the band gap is (1.4602-(-9.1108)=10.571eV) Hmmm... maybe I am wrong, but I would think this is the gap... AT (0,0,0)!!! You should plot the bands in several directions to see what the gap is. It might well be an indirect gap (the valence band maxima and the conduction band minima occur at different q_points), and, as such, only observable in a plot or in a very careful analysis. Hope it helps, Miguel -- ---------------------------------------- Miguel Mart?nez Canales Dto. F?sica de la Materia Condensada UPV/EHU Facultad de Ciencia y Tecnolog?a Apdo. 644 48080 Bilbao (Spain) Fax: +34 94 601 3500 Tlf: +34 94 601 5437 ---------------------------------------- 2005 International Year of Physics "But you don't ask with respect. You don't offer friendship. You don't even think to call me Godfather." Don Vito Corleone _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From eyvaz_isaev at yahoo.com Wed Oct 26 18:25:56 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Wed, 26 Oct 2005 09:25:56 -0700 (PDT) Subject: [Pw_forum] ΄πΈ΄: [Pw_forum] why I can not get the correct band gap of TiO2? In-Reply-To: <34C4FA35021357469685D8102806A1C002905B98@mail01.student.main.ntu.edu.sg> Message-ID: <20051026162556.84593.qmail@web60320.mail.yahoo.com> Hi, Thanks Gerardo explaining the band gap issue. > By the way, do > you know how to plot out the band structure based on > the output file of pwscf? Thx in advance. > To me the best way is the use of a program I sent to forum some time ago. I also promised attache it to the Fermi Surface program, but, sorry, I did not it yet. There are a lot useless things I have to do. But I will do that definitely. In any case besides SCF calculations you have to do NSCF ones for a number of k-points along high symmetry directions. Bests, Eyvaz. __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com From eyvaz_isaev at yahoo.com Wed Oct 26 18:29:32 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Wed, 26 Oct 2005 09:29:32 -0700 (PDT) Subject: [Pw_forum] ΄πΈ΄: [Pw_forum] why I can not get the correct band gap of TiO2? In-Reply-To: <34C4FA35021357469685D8102806A1C002905B97@mail01.student.main.ntu.edu.sg> Message-ID: <20051026162932.85113.qmail@web60320.mail.yahoo.com> > However, why in the > valance band , there has a huge gap?( see the values > -48.0314 and -24.6221) , These are Ti semicore states. Bests, Eyvaz. __________________________________ Start your day with Yahoo! - Make it your home page! http://www.yahoo.com/r/hs From giannozz at nest.sns.it Wed Oct 26 18:36:30 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 26 Oct 2005 18:36:30 +0200 Subject: [Pw_forum] =?gb2312?q?=B4=F0=B8=B4=3A_=5BPw=5Fforum=5D_why_I_can_not_get_the?= =?gb2312?q?_correct_band_gap_of?= TiO2? In-Reply-To: <34C4FA35021357469685D8102806A1C002905B97@mail01.student.main.ntu.edu.sg> References: <34C4FA35021357469685D8102806A1C002905B97@mail01.student.main.ntu.edu.sg> Message-ID: <200510261836.30794.giannozz@nest.sns.it> On Wednesday 26 October 2005 18:03, #TAO JUNGUANG# wrote: > [...] can you tell how to use bands.x and forthmore band_plot.x > to plot the band structure? Thanks there is an example of band structure plotting in examples/example05 P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From quevedin at gmail.com Thu Oct 27 10:53:15 2005 From: quevedin at gmail.com (Lucas Fernandez Seivane) Date: Thu, 27 Oct 2005 10:53:15 +0200 Subject: [Pw_forum] Building a chain Message-ID: <2ebe300a0510270153i7090beb8l7cb69a3dd807afdc@mail.gmail.com> Hi friends of the list. I amt trying to simulate a linear chain (single atom for the moment, and I put the following input. I put the following input: &system ibrav= 14, a=2.80, b=12, c=12, cosab=0.0, cosbc=0.0,cosac=0.0 Is this correct? Since PWSCF takes advantages of yhe symmetries, I don't know if I am doing 'the best' election for the ibrav. Best wishes -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20051027/5e33d648/attachment.htm From eyvaz_isaev at yahoo.com Thu Oct 27 13:09:20 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Thu, 27 Oct 2005 04:09:20 -0700 (PDT) Subject: [Pw_forum] Building a chain In-Reply-To: <2ebe300a0510270153i7090beb8l7cb69a3dd807afdc@mail.gmail.com> Message-ID: <20051027110921.24601.qmail@web60325.mail.yahoo.com> Hi, Presumably, you should use the simple tetragonal lattice (ibrav=6), but not triclinic. Bests, Eyvaz. --- Lucas Fernandez Seivane wrote: > Hi friends of the list. > > I amt trying to simulate a linear chain (single atom > for the moment, and I > put the following input. I put the following input: > > &system > ibrav= 14, > a=2.80, b=12, c=12, > cosab=0.0, cosbc=0.0,cosac=0.0 > > Is this correct? Since PWSCF takes advantages of yhe > symmetries, I don't > know if I am doing 'the best' election for the > ibrav. > > > Best wishes > __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com From L.Boeri at fkf.mpg.de Thu Oct 27 14:27:08 2005 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Thu, 27 Oct 2005 14:27:08 +0200 (CEST) Subject: [Pw_forum] Pseudopotential Format In-Reply-To: <8d38992a0510250718p66eea6f3t27cf064ce53ab1f9@mail.gmail.com> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <8d38992a0510250718p66eea6f3t27cf064ce53ab1f9@mail.gmail.com> Message-ID: Dear PWSCF users, I have been trying to convert a Li pseudopotential downloaded from the ABINIT database (LDA TM pseudopotential generated with the FHI98 code) to PWSCF UPF format; I thought I had succeeded in the conversion because the total energies and other properties (bulk modulus, phonon frequencies etc) were very similar in the two programs (the difference in total energies is of the order of 1mRy). The problem is that when I looked at the scf potential in real space given by the two programs the shape was not the same... The difference is really in the ionic part, which means that the pseudopotential is not the same in the two calculations. - Has any of you had any experience in converting the pseudopotentials from one code to the other? - Where can I find documentation about the UPF format ? I have a doubt about the meaning of the "maximum angular component"? What is the difference between this and the "number of wavefunctions and number of projectors"? Where should I type the l local? Sorry for the confusion, thank you in advance for your response, Lilia p.s. I am attaching the headers of the two files, in case any of you notice what is wrong in my conversion ----------------------------------------------------- ABINIT ----------------------------------------------------- lithium, fhi98PP : Trouiller-Martins-type, LDA Ceperley/Alder Perdew/Wang (1992), l= 2 local 3.000 1.000 021003 zatom,zion,pspdat 6 7 3 2 433 0 pspcod,pspxc,lmax,lloc,mmax,r2well 4.763 1.000 0.000 rchrg, fchrg, qchrg 5--- These two lines are available for giving more information, later 6 7-Here follows the cpi file from the fhi98pp code- 1.00000E+00 4 <---- This should be the valence charge and the number of wavefunctions 0.0000 0.0000 0.0000 0.0000 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 0.0000 .00e+00 .00e+00 433 0.10247000000000E+01 ---------------------------------------------------------- UPF ---------------------------------------------------------- 0 Version Number Li Element NC Norm - Conserving pseudopotential T Nonlinear Core Correction SLA PW NOGX NOGC Exchange-Correlation functional 1.00000000000 Z valence 0.00000000000 Total energy 0.0000000 0.0000000 Suggested cutoff for wfc and rho 3 Max angular momentum component 433 Number of points in mesh 4 3 Number of Wavefunctions, Number of Projectors Wavefunctions nl l occ 2s 0 1.00 2p 1 0.00 3d 2 0.00 4f 3 0.00 ----------------------------------------------------------- From degironc at sissa.it Thu Oct 27 16:07:00 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Thu, 27 Oct 2005 16:07:00 +0200 Subject: [Pw_forum] Pseudopotential Format References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <8d38992a0510250718p66eea6f3t27cf064ce53ab1f9@mail.gmail.com> Message-ID: <4360DF04.3040602@sissa.it> In the upftools directory of the distribution there is the file UPF that contains the definition of UPF format. UPF stores pseudopotentials in the KB form. There is a local potential and a series of projectors. Potentials for individual angular momenta ARE NOT stored. The choice of lloc (to be provided when converting the pseudopotential to UPF format) determines which semilocal potential is used as the local one. Stefano de Gironcoli Lilia Boeri wrote: >Dear PWSCF users, > > >I have been trying to convert a Li pseudopotential downloaded from the >ABINIT database (LDA TM pseudopotential generated with the FHI98 code) to PWSCF >UPF format; I thought I had succeeded in the conversion >because the total energies and other properties (bulk modulus, phonon >frequencies etc) were very similar in the two programs (the difference in >total energies is of the order of 1mRy). The problem is that when I looked >at the scf potential in real space given by the two programs the shape was >not the same... The difference is really in the ionic part, which means >that the pseudopotential is not the same in the two calculations. > >- Has any of you had any experience in converting the pseudopotentials >from one code to the other? >- Where can I find documentation about the UPF format ? I have a doubt >about the meaning of the "maximum angular component"? What is the >difference between this and the "number of wavefunctions and number of >projectors"? Where should I type the l local? > >Sorry for the confusion, thank you in advance for your response, >Lilia > >p.s. I am attaching the headers of the two files, in case any of you >notice what is wrong in my conversion > >----------------------------------------------------- >ABINIT >----------------------------------------------------- >lithium, fhi98PP : Trouiller-Martins-type, LDA Ceperley/Alder Perdew/Wang >(1992), l= 2 local > 3.000 1.000 021003 zatom,zion,pspdat > 6 7 3 2 433 0 >pspcod,pspxc,lmax,lloc,mmax,r2well > 4.763 1.000 0.000 rchrg, fchrg, qchrg >5--- These two lines are available for giving more information, later >6 >7-Here follows the cpi file from the fhi98pp code- >1.00000E+00 4 <---- This should be the valence charge and the number >of wavefunctions > 0.0000 0.0000 0.0000 0.0000 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 > 0.0000 .00e+00 .00e+00 >433 0.10247000000000E+01 > >---------------------------------------------------------- >UPF >---------------------------------------------------------- > > > 0 Version Number > Li Element > NC Norm - Conserving pseudopotential > T Nonlinear Core Correction > SLA PW NOGX NOGC Exchange-Correlation functional > 1.00000000000 Z valence > 0.00000000000 Total energy > 0.0000000 0.0000000 Suggested cutoff for wfc and rho > 3 Max angular momentum component > 433 Number of points in mesh > 4 3 Number of Wavefunctions, Number of Projectors > Wavefunctions nl l occ > 2s 0 1.00 > 2p 1 0.00 > 3d 2 0.00 > 4f 3 0.00 > >----------------------------------------------------------- > >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From giannozz at nest.sns.it Thu Oct 27 17:13:46 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 27 Oct 2005 17:13:46 +0200 Subject: [Pw_forum] Pseudopotential Format In-Reply-To: References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <8d38992a0510250718p66eea6f3t27cf064ce53ab1f9@mail.gmail.com> Message-ID: <200510271713.46529.giannozz@nest.sns.it> On Thursday 27 October 2005 14:27, Lilia Boeri wrote: > I have been trying to convert a Li pseudopotential downloaded from > the ABINIT database (LDA TM pseudopotential generated with the > FHI98 code) to PWSCF UPF format; I thought I had succeeded in the > conversion because the total energies and other properties (bulk > modulus, phonon frequencies etc) were very similar in the two > programs (the difference in total energies is of the order of 1mRy). > The problem is that when I looked at the scf potential in real space > given by the two programs the shape was not the same... The > difference is really in the ionic part, which means that the > pseudopotential is not the same in the two calculations. I already heard complaints that the conversion to UPF of FHI98 pseudopotentials occasionally doesn't work, in particular when the "nonlinear core correction" is present. In order to solve this problem, one needs to know what is inside the FHI98 files (what is inside the UPF format is documented in file upftools/UPF) and to modify the converter accordingly. Since the reward-to-effort ratio of such operation is unfavourable, I am afraid that this is not going to happen anytime soon. Before claiming that the PP is not correctly converted, however, you should try to verify what you get in an atomic calculation: the atomic code should work with UPF files, if the radial grid does not start from 0. If you can well reproduce one-electron levels and total energy differences on different atomic configurations, your PP does the job. The difference you observe may be due to something else (for instance, you are not really looking at the same quantity) Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From L.Boeri at fkf.mpg.de Thu Oct 27 20:21:55 2005 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Thu, 27 Oct 2005 20:21:55 +0200 (CEST) Subject: [Pw_forum] Pseudopotential Format In-Reply-To: <200510271713.46529.giannozz@nest.sns.it> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <8d38992a0510250718p66eea6f3t27cf064ce53ab1f9@mail.gmail.com> <200510271713.46529.giannozz@nest.sns.it> Message-ID: dear Paolo, thanks a lot for answering. I actually tried all the possible tests with atomic, and they work nicely (total energies of different atomic configuratons differ by less than 1d-4 Ry, wavefunctions look all right, as well as logarithmic derivatives). As I told you, also in the bulk properties the difference between the two codes are tiny, but there is some difference in the 1d projection of the density and potential. Is there a way to plot the potential with atomic ? Best, Lilia On Thu, 27 Oct 2005, Paolo Giannozzi wrote: > On Thursday 27 October 2005 14:27, Lilia Boeri wrote: > > > I have been trying to convert a Li pseudopotential downloaded from > > the ABINIT database (LDA TM pseudopotential generated with the > > FHI98 code) to PWSCF UPF format; I thought I had succeeded in the > > conversion because the total energies and other properties (bulk > > modulus, phonon frequencies etc) were very similar in the two > > programs (the difference in total energies is of the order of 1mRy). > > The problem is that when I looked at the scf potential in real space > > given by the two programs the shape was not the same... The > > difference is really in the ionic part, which means that the > > pseudopotential is not the same in the two calculations. > > I already heard complaints that the conversion to UPF of FHI98 > pseudopotentials occasionally doesn't work, in particular when the > "nonlinear core correction" is present. In order to solve this problem, > one needs to know what is inside the FHI98 files (what is inside the > UPF format is documented in file upftools/UPF) and to modify the > converter accordingly. Since the reward-to-effort ratio of such > operation is unfavourable, I am afraid that this is not going > to happen anytime soon. > > Before claiming that the PP is not correctly converted, however, > you should try to verify what you get in an atomic calculation: > the atomic code should work with UPF files, if the radial grid > does not start from 0. If you can well reproduce one-electron levels > and total energy differences on different atomic configurations, > your PP does the job. The difference you observe may be due > to something else (for instance, you are not really looking at the > same quantity) > > Paolo > -- > Paolo Giannozzi e-mail: giannozz at nest.sns.it > Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 > Piazza dei Cavalieri 7 I-56126 Pisa, Italy > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From giannozz at nest.sns.it Thu Oct 27 22:08:37 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 27 Oct 2005 22:08:37 +0200 Subject: [Pw_forum] Magnetism in carbon structures - technical issues In-Reply-To: <36534.10.50.40.120.1130245056.squirrel@webmail1.ictp.trieste.it> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <36534.10.50.40.120.1130245056.squirrel@webmail1.ictp.trieste.it> Message-ID: <200510272208.37932.giannozz@nest.sns.it> On Tuesday 25 October 2005 14:57, Marcos Verissimo Alves wrote: > [...] it is mentioned that PSP for magnetic system must include > nonlinear core corrections (NLCC for laziness). My questions are: > > 1) Is this really mandatory for all magnetic systems? And in the > particular case of Carbon and Oxygen, would NLCC be mandatory, too? mandatory maybe not, but it may be very useful > 2) Instead of nonlinear core corrections, could the 1s electrons be > included in Vanderbilt or RKKJ ultrasoft PSPs? it doesn't look like a good idea to me Paolo From giannozz at nest.sns.it Fri Oct 28 18:12:31 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 28 Oct 2005 18:12:31 +0200 Subject: [Pw_forum] Pseudopotential Format In-Reply-To: References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <200510271713.46529.giannozz@nest.sns.it> Message-ID: <200510281812.31091.giannozz@nest.sns.it> On Thursday 27 October 2005 20:21, Lilia Boeri wrote: > Is there a way to plot the potential with atomic ? presently no, although it should be quite easy to write the desired quantity in a format that can be read by gnuplot, xmgrace or whatever plotting package one likes P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From stewart at cnf.cornell.edu Fri Oct 28 18:30:52 2005 From: stewart at cnf.cornell.edu (stewart at cnf.cornell.edu) Date: Fri, 28 Oct 2005 12:30:52 -0400 Subject: [Pw_forum] Restarting calculations In-Reply-To: <200510281812.31091.giannozz@nest.sns.it> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <200510271713.46529.giannozz@nest.sns.it> <200510281812.31091.giannozz@nest.sns.it> Message-ID: <20051028163052.13103.qmail@xuxa.iecc.com> Hi all, I would like to restart pw calculations that haven't converged. Unfortunately, I run into efermit errors when I try to restart the calculations. I have encountered this before when I am taking a converged calculations and doing some processing on it to get band, dos, etc with non-scf runs. In this case, I believe I removed the information on smearing and tetrahedra integration (found in a suggestion in a previous post) and everything ran. Does anyone have suggestions for getting restart calculations to work when you are still seeking convergence. I am looking at spin polarized calculations. Could the spin polarization lead to problems? In past posts, there seem to be some suggestions that information is lost that is required to restart spin calculations. Thanks, Derek ##################################### Derek Stewart, Ph. D. Scientific Computation Associate 250 Duffield Hall Cornell Nanoscale Facility (CNF) Ithaca, NY 14853 stewart (at) cnf.cornell.edu (607) 255-2856 From giannozz at nest.sns.it Fri Oct 28 18:52:34 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 28 Oct 2005 18:52:34 +0200 Subject: [Pw_forum] Restarting calculations In-Reply-To: <20051028163052.13103.qmail@xuxa.iecc.com> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <200510281812.31091.giannozz@nest.sns.it> <20051028163052.13103.qmail@xuxa.iecc.com> Message-ID: <200510281852.34617.giannozz@nest.sns.it> On Friday 28 October 2005 18:30, stewart at cnf.cornell.edu wrote: > I would like to restart pw calculations that haven't converged. > Unfortunately, I run into efermit errors when I try to restart the > calculations. I already heard about this. I think it is fixed now. Which version are you using? P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From stewart at cnf.cornell.edu Fri Oct 28 19:45:08 2005 From: stewart at cnf.cornell.edu (stewart at cnf.cornell.edu) Date: Fri, 28 Oct 2005 13:45:08 -0400 Subject: [Pw_forum] Re: Restarting calculations In-Reply-To: <200510281852.34617.giannozz@nest.sns.it> References: <8d38992a0510250528o29afe214rd83442916c9bd0c1@mail.gmail.com> <200510281812.31091.giannozz@nest.sns.it> <20051028163052.13103.qmail@xuxa.iecc.com> <200510281852.34617.giannozz@nest.sns.it> Message-ID: <20051028174508.13286.qmail@xuxa.iecc.com> Hi Paolo, I am currently using espresso version 2.1.5 running on redhat-linux enterprise with the intel v8 compilers. I downloaded version 2.1.5 back in August. Have there been any small changes to the distribution since then? Thanks, Derek ################################ Derek Stewart, Ph. D. Scientific Computation Associate 250 Duffield Hall Cornell Nanoscale Facility (CNF) Ithaca, NY 14853 stewart (at) cnf.cornell.edu Paolo Giannozzi writes: > On Friday 28 October 2005 18:30, stewart at cnf.cornell.edu wrote: > >> I would like to restart pw calculations that haven't converged. >> Unfortunately, I run into efermit errors when I try to restart the >> calculations. > > I already heard about this. I think it is fixed now. Which version > are you using? P. > > -- > Paolo Giannozzi e-mail: giannozz at nest.sns.it > Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 > Piazza dei Cavalieri 7 I-56126 Pisa, Italy > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum From giannozz at nest.sns.it Fri Oct 28 20:21:48 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 28 Oct 2005 20:21:48 +0200 Subject: [Pw_forum] Fwd: BOMD / FPMD / CP Message-ID: <200510282021.49004.giannozz@nest.sns.it> Paul Tangney asked me to post the following message: ---------- Forwarded Message ---------- Subject: BOMD / FPMD / CP Date: Friday 28 October 2005 19:06 From: Paul Tangney To: giannozz at nest.sns.it Hello, I am using Espresso on an Opteron cluster with LAM/MPI and Portland compilers. My system is an Indole molecule surrounded by 63 waters. I have a few questions: (1) Is FPMD still being maintained ? Some of the features that used to be in FPMD a few years ago are no longer working and I had to delve into the code to 'reactivate' them. e.g. velocity rescaling. Also, some of the default parameters are different from what the documentation says they are. (2) I'm doing Born-Oppenheimer dynamics with DIIS. It works great for a while (although I'm sure the parameters I'm using are far from optimal) and then, after a few hundred steps, the performance drops by a factor of about 100000. The simulation basically stops. Has anybody else experienced this ? Is there a memory leak ? (3) Does anybody know which code (CP/FPMD) is faster for Born-Oppenheimer MD using norm-conserving PPs and which minimization method works best for water ? Is any effort being put into making BOMD fast in these codes or is the focus only on CPMD ? Thanks for your help. Paul ------------------------------------------------------- From tangney at civet.berkeley.edu Sat Oct 29 02:25:46 2005 From: tangney at civet.berkeley.edu (Paul Tangney) Date: Fri, 28 Oct 2005 17:25:46 -0700 Subject: [Pw_forum] Mermin free energy & Fermi-Dirac smearing Message-ID: <4362C18A.5060405@civet.berkeley.edu> Hello, I am trying to calculate the total (free) energy of a semiconductor with electrons at a high temperature. I am doing this by tinkering with the Fermi-Dirac smearing scheme in PWSCF. Now, as far as I know, if F-D smearing is used purely as a computational device for metals, the ground state energy should be calculated (estimated) by E_0 = E - TS/2 , where E is the standard Kohn-Sham energy and the second term is half the electronic entropy energy. Is this correct ? What I want is to calculate F = E - TS, the free energy. However, when I looked in the code (subroutines "gweights" and "w1gauss" ) it appears to me that the ground state energy for metals is being calculated as E_0 = E - TS. In other words, the correction term is twice what it should be. Am I right about this, or am I reading the code wrong ? Factors of two can have many origins. Is the variable "demet" TS or TS/2 ? If I can figure this out, I can figure out how to get the quantity I want. Thanks for any help, Paul -- ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo Dr. Paul Tangney Theory of Nanostructured Materials Facility The Molecular Foundry Lawrence Berkeley National Lab. E-mail: PTTangney at lbl.gov 1 Cyclotron Road, Bldg 66 Phone: (510) 642-2635 Berkeley, CA 94720 Fax : (510) 643-9345 ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo From marzari at MIT.EDU Sat Oct 29 12:46:46 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Sat, 29 Oct 2005 06:46:46 -0400 Subject: [Pw_forum] Mermin free energy & Fermi-Dirac smearing In-Reply-To: <4362C18A.5060405@civet.berkeley.edu> References: <4362C18A.5060405@civet.berkeley.edu> Message-ID: <43635316.7040101@mit.edu> Hi Paul, have a look at this: http://www.democritos.it/pipermail/pw_forum/2004-January/000785.html Basically, you should use E-TS/2 to extrapolate a finite-temperature calculation done with fermi-dirac or gaussian smearing to its zero temperature limit. Note that there is never a good reason to do this - you should use instead methfessel-paxton or marzari-vanderbilt. If you really want the "physical" E-TS at finite T, then you should use Fermi-Dirac, and use E-TS in your work - no corrections to it to extrapolate to zero-temperature (note my comments in the previous link about the Vertstraete-Gonze approach to use cold smearing to extrapolate to a finite Fermi-Dirac distribution) What PWSCF calls "total energy" in the presence of a temperature is actually E-TS (and I really, really think we should call it "total free energy", to decrease this permanent state of confusion), and this E-TS is what you should use. Note that very little has been done to figure out what the real exchange-correlation functional at finite-temperature is (in fact, we always use the T=0 one...), and that it's even more difficult than usual to express from a charge density the correct free energy of a system, since that contains an ensemble-average at a given temperature. The Sprik PRL from ~2003 on ions in solution can be very illuminating on the problem. Not sure what "demet" is. nicola Paul Tangney wrote: > Hello, > > I am trying to calculate the total (free) energy of > a semiconductor with electrons at a high temperature. > I am doing this by tinkering with the Fermi-Dirac smearing > scheme in PWSCF. Now, as far as I know, if F-D smearing is > used purely as a computational device for metals, the > ground state energy should be calculated (estimated) by > E_0 = E - TS/2 , where E is the standard Kohn-Sham energy > and the second term is half the electronic entropy energy. > Is this correct ? > > What I want is to calculate F = E - TS, the free energy. > However, when I looked in the code (subroutines "gweights" and > "w1gauss" ) it appears to me that the ground state energy > for metals is being calculated as E_0 = E - TS. > In other words, the correction term is twice what it should > be. Am I right about this, or am I reading the code wrong ? > Factors of two can have many origins. > Is the variable "demet" TS or TS/2 ? > > If I can figure this out, I can figure out how to get > the quantity I want. > > Thanks for any help, > > > Paul > > -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From marzari at MIT.EDU Sat Oct 29 12:48:39 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Sat, 29 Oct 2005 06:48:39 -0400 Subject: [Pw_forum] Question on stress in a system with constraints In-Reply-To: References: <20051012152804.8020.qmail@web52013.mail.yahoo.com> <434D2D48.10207@mit.edu> Message-ID: <43635387.2060503@mit.edu> Stefano de Gironcoli wrote: > > Dear Nicola and Kostya, > I did not think deeply about the question, so I may be wrong, but I agree > with Kostya. > > Stress is a first order derivative and as such only GS properties should > be needed. > Yes - apologies I forgot to correct this. For some reason I was thinking at elastic constants, thus at second derivatives of the energy. Bottom line - all my previous emails would be correct if you were looking at an elastic constant. For stress, what Stefano and Kostya said was right. nicola --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From tangney at civet.berkeley.edu Sat Oct 29 19:05:58 2005 From: tangney at civet.berkeley.edu (Paul Tangney) Date: Sat, 29 Oct 2005 10:05:58 -0700 (PDT) Subject: [Pw_forum] More on F-D smearing Message-ID: Thanks Nicola, I think that answered my question. If you're right and what PWSCF calls "total energy" is E-TS, then "demet" is -TS and is incorrectly called the "correction for metals" in the code. As you say, the "correction" should be -TS/2. If F-D smearing is not a good computational device then definitely, I agree, the code should call this "Total Energy" "Free Energy" instead. Otherwise, if F-D is to be used purely as a smearing technique, the correct correction should be used...i.e. "demet" should be changed in the code from -TS to -TS/2. I'm interested in using F-D in a *physical* way. I want F = E - TS. So I guess I don't need to change the code at all. By the way, on the system that I'm studying (crystalline Tellurium), the application of a finite temperature actually *lowers* the free energy. Obviously this is unphysical....although perhaps not totally surprising, given that I'm using a finite (but dense : 10x10x10) k-point grid. Its still a little worrying though. Regards, Paul -- ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo Dr. Paul Tangney Theory of Nanostructured Materials Facility The Molecular Foundry Lawrence Berkeley National Lab. E-mail: PTTangney at lbl.gov 1 Cyclotron Road, Bldg 66 Phone: (510) 642-2635 Berkeley, CA 94720 Fax : (510) 643-9345 ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo From umari at MIT.EDU Sat Oct 29 21:54:43 2005 From: umari at MIT.EDU (Paolo Umari) Date: Sat, 29 Oct 2005 15:54:43 -0400 Subject: [Pw_forum] Re: Fwd: BOMD / FPMD / CP (Paolo Giannozzi) Message-ID: <20051029155443.gx9vez9bms1wc88s@webmail.mit.edu> Dear Paul, With the new quantum-espresso release, the code cp.x will be able to perform BO molecular dynamics, through conjugate gradient optimization of the electronic states (instead of Verlet). It works also for metals through ensemble DFT. Paolo Umari. From marzari at MIT.EDU Sun Oct 30 00:59:42 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Sat, 29 Oct 2005 18:59:42 -0400 Subject: [Pw_forum] More on F-D smearing In-Reply-To: References: Message-ID: <4363FEDE.5030202@mit.edu> Hi Paul, I think there are different world-views behind the nomenclature - if you think at the smearing orginating in a convolution of the density of states, you tend to think at the variational functional as an energy. I like to think at smearing as a ficititious temperature, hence the free energy. Maybe we should sort this once for all. Stay tuned. > By the way, on the system that I'm studying (crystalline > Tellurium), the application of a finite temperature actually > *lowers* the free energy. Obviously this is unphysical....although > perhaps not totally surprising, given that I'm using a finite > (but dense : 10x10x10) k-point grid. Its still a little worrying though. No, no, it is ok, for a regular smearing - the free energy is quadratic around T=0, and the second-order coefficient is negative. Have a look at paragraphs 4.2 and 4.3 of http://nnn.mit.edu/phd , in which the Taylor expansion of the free energy around T=0 is discussed (there is some subtlety having to do with the fact that the partial and the total derivative of the free energy with respect to T are equal, since the free energy is a minimum with respect to psi_i and f_ij). So, nothing to worry. Not yet. nicola --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From tangney at civet.berkeley.edu Sun Oct 30 06:28:57 2005 From: tangney at civet.berkeley.edu (Paul Tangney) Date: Sat, 29 Oct 2005 22:28:57 -0700 (PDT) Subject: [Pw_forum] F-D Smearing etc.. Message-ID: Hi Nicola, I obviously wasn't clear. I agree that E-TS should be called the free energy. E-TS/2, on the other hand, is the estimate of the ground state energy for T=0....when T is used as a computational device. I don't agree that F = E-TS should be lower than the ground state energy (i.e. F, when T=0) which is what I'm seeing in my simulations. Of course -TS is always negative because S and T are positive, but the magnitude of TS should be smaller than the difference between E(T>0) and E(T=0). I can't download your thesis from where I am right now, but I'm pretty sure that the Kohn-Sham ground state is always lower than the ground state of the Mermin functional for finite T, and equal to it at T=0. I think the code is not actually calculating what I want. What I want is F=E-TS, minimized with respect to the occupation numbers and the orbitals. I don't think the code calculates this...it just goes to self consistency with occupations determined by a F-D distribution. I'm not sure about this, but I think that what the code calculates can, in principle, depend on when (i.e. at which iteration) the excited states are occupied. If at the first iteration we have a really crap band structure, and then we occupy it with our F-D dist, this occupation can determine what happens in the rest of the self consistent calculation. We might get something different if we began by only occupying valence bands, and then at iteration 5 (say) started imposing the F-D distribution. I don't know if this is why F decreases as T increases in my calculations. Anyway, I'm still worried....but its almost Sunday, so I'll have a day of rest. Regards, Paul -- ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo Dr. Paul Tangney Theory of Nanostructured Materials Facility The Molecular Foundry Lawrence Berkeley National Lab. E-mail: PTTangney at lbl.gov 1 Cyclotron Road, Bldg 66 Phone: (510) 642-2635 Berkeley, CA 94720 Fax : (510) 643-9345 ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo From marzari at MIT.EDU Sun Oct 30 13:31:09 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Sun, 30 Oct 2005 07:31:09 -0500 Subject: [Pw_forum] F-D Smearing etc.. In-Reply-To: References: Message-ID: <4364BD0D.6000203@mit.edu> Hi Paul, I've uploaded in http://nnn.mit.edu/paul a plot of E-TS (free), E (total), and E-TS/2 (corrected) for bulk Al, and Gauassian smearing. This should show you what to expect for the change in temperature of the different quantities. The analytical proofs are in the thesis. Note that if you were using Methfessel-Paxton or Marzari-Vanderbilt, the free energy wouldn't have a quadratic term, and would look almost identical to the corrected energy of the plot attached. > What I want is F=E-TS, minimized with respect > to the occupation numbers and the orbitals. I don't think > the code calculates this...it just goes to self consistency > with occupations determined by a F-D distribution. All codes calculate exactly what you want, i.e. F=E-TS minimized with respecte to occupation numbers and orbitals. They do this either using a direct minimization recipe on all those degrees of freedom (this is the "ensemble-DFT" philosophy, now available in the CVS of Quantum-Espresso, thanks to Paolo Umari and Ismaila Dabo), or they reach iteratively selfconsistency (PWSCF and all other codes). Once you are *fully* selfconsistent (i.e. the eigenvectors of your Hamiltonian, populated with occupations given by the thermal-distribution of their eigenvalues, produce a charge density that is identical to the previous one), you have the same ground state of the direct minimization. I think we should continue this is in private - but again, have a careful look at the reference first. nicola --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From yang_jun-jun at hotmail.com Sun Oct 30 14:55:55 2005 From: yang_jun-jun at hotmail.com (Jun-jun Yang) Date: Sun, 30 Oct 2005 05:55:55 -0800 Subject: [Pw_forum] SCF convergency Message-ID: Dear list users, I'm working on scf calculation of mineral chromite. I think the calculation may be on the right track but I still have problem on its convergency. The accuracy alway osillates between about 1.5e-5 and 1 (Ry) after 10th iteration (100 in total) although the lower value had nearly reached the convergence threshold conv_thr = 1.0d-5. I have tried other two minimization algorithms (david and diis), and it seems that the result with cg is the best. My co-worker is trying to do the same calculation with CPMD (v3.9.2). He did electron annealing first to bring the total energy to a would-be global(?) minima, then switch to scf calculation using diis or cg with preconditioning. Unfortunately, he got very similar system behaviour with mine. As my next step, I prepare to set conv_thr = 5.0e-5 and continue my calculation with different a0, then use the Etot - volume data to fit an EOS and compare the result with other experimental data. This could tell me if the calculations are reasonable. As both my co-worker and I are beginers of the relevant software, and the calculations are very time-consumming, before I proceed to my next step, I ask your help to see whether I made anything wrong in the input, or let me know if I need to do this or that for correct practice. Best regards and thanks in advance. Yang Jun-jun Below is the snippet of my input file: =================================== &control calculation = 'scf' restart_mode='from_scratch' prefix='fecr2o4' tstress = .false. tprnfor = .false. pseudo_dir = '$HOME/espresso-2.1.5/pseudo' outdir = $HOME/tmp/ wf_collect = .true. / &system ibrav = 1, celldm(1) = 15.8321, nat = 56, ntyp = 3 nbnd = 330 nspin = 2, degauss = 0.5 occupations = 'smearing' smearing = 'gaussian' starting_magnetization(1) = 0.0 ecutwfc = 25.0 ecutrho = 140.0 / &electrons diagonalization = 'cg' mixing_mode = 'plain' mixing_beta = 0.40 conv_thr = 1.0d-5 / ATOMIC_SPECIES Fe 55.847 Fe.pbe-sp-van.UPF Cr 51.996 Cr.pbe-sp-van.UPF O 15.999 O.pbe-van_ak.UPF ATOMIC_POSITIONS angstrom Fe 0.00000 0.00000 0.00000 Cr 5.23625 5.23625 5.23625 O 3.23391 3.23391 3.23391 Fe 0.00000 4.18900 4.18900 Fe 4.18900 0.00000 4.18900 [...] K_POINTS automatic 1 1 1 0 0 0 I have sent this mail three times with a graph attached. The first two posts had been rejected because size of the mail was a little larger than the limit. The third one should be below 40KB, but I haven't seen its appearing on the list yet. So, I have to post it again without a graph. I apologize for that you may receive two mails with similar content. _________________________________________________________________ On the road to retirement? Check out MSN Life Events for advice on how to get there! http://lifeevents.msn.com/category.aspx?cid=Retirement From eyvaz_isaev at yahoo.com Sun Oct 30 17:35:32 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Sun, 30 Oct 2005 08:35:32 -0800 (PST) Subject: [Pw_forum] SCF convergency In-Reply-To: Message-ID: <20051030163532.64336.qmail@web60322.mail.yahoo.com> Hi, Hopefully you know that celldm(1)=15.8321 is in atomic units, but not Angstrom. You asked to do spin-polarized calculations (nspin=2), but put initial magnetic splitting to zero. So, you have to change starting_magnetization(1) = 0.9 and add more 2 same lines for Cr and O atoms (your first atomic type is Fe). Presumably, you should not specify the bands number, so, remove nbnd=330. Be careful with this number, it means that you have around 600-660 electrons in your crystal. If you are sure that your nbnd covers bands above the Fermi level,then it is OK. Ecutrho seems to be low, increase up to 200Ry. Try mixing_beta=0.1. degauss=0.5? It is very large. Try degauss=0.03, it is in Ry. To me conv_thr=1.d-5 seems too large. Default conv_thr (1d-6) looks more acceptable. Finally, if something went OK, be ready for LDA+U calculations. Fe-Cr-O looks like a strongly correlated system. Best regards, Eyvaz. > &system > ibrav = 1, celldm(1) = 15.8321, nat = 56, ntyp = > 3 > nbnd = 330 > nspin = 2, degauss = 0.5 > occupations = 'smearing' > smearing = 'gaussian' > starting_magnetization(1) = 0.0 > ecutwfc = 25.0 > ecutrho = 140.0 > / > &electrons > diagonalization = 'cg' > mixing_mode = 'plain' > mixing_beta = 0.40 > conv_thr = 1.0d-5 > / > > ATOMIC_SPECIES > Fe 55.847 Fe.pbe-sp-van.UPF > Cr 51.996 Cr.pbe-sp-van.UPF > O 15.999 O.pbe-van_ak.UPF > > ATOMIC_POSITIONS angstrom > Fe 0.00000 0.00000 0.00000 > Cr 5.23625 5.23625 5.23625 > O 3.23391 3.23391 3.23391 > Fe 0.00000 4.18900 4.18900 > Fe 4.18900 0.00000 4.18900 > [...] > > K_POINTS automatic > 1 1 1 0 0 0 > > > I have sent this mail three times with a graph > attached. The first two posts > had been rejected > because size of the mail was a little larger than > the limit. The third one > should be below 40KB, > but I haven't seen its appearing on the list yet. > So, I have to post it > again without a graph. I > apologize for that you may receive two mails with > similar content. > > _________________________________________________________________ > On the road to retirement? Check out MSN Life Events > for advice on how to > get there! > http://lifeevents.msn.com/category.aspx?cid=Retirement > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From silviu at Princeton.EDU Sun Oct 30 22:58:15 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Sun, 30 Oct 2005 23:58:15 +0200 Subject: [Pw_forum] magnetization and spin contamination Message-ID: <436541F7.7080809@Princeton.EDU> Hi, I have a few questions related to spin-polarized calculations in PWSCF: 1. Is there a way to get the spin contamination information? 2. One can constrain the total magnetization. The INPUT_PW file provides some information on how it is done in the non-collinear case, but the collinear case (nspin=2) is more obscured. Could someone explain it some more or point me to a reference? 3. Can one remove the spin-contamination by constraining to total magnetization to the correct value s(s+1)? Thanks, Silviu. From degironc at sissa.it Mon Oct 31 10:43:01 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Mon, 31 Oct 2005 10:43:01 +0100 Subject: [Pw_forum] More on F-D smearing References: Message-ID: <4365E725.3030609@sissa.it> Dear Paul, I just wanted to confirm all what Nicola was writing. The quantity called total energy in PWSCF is the variational quantity in presence of finite smearing, whose functional derivatives are the K-S equations, as solved by the code, and whose derivatives (forces and stress) can be computed exploiting Hellmann-Feynman theorem . Demet is the quantity that needs to be added to the band energy \sum f_i \epsilon_i in order to make it equal to the integral computed from the smeared density of state \int \epsilon n(\epsilon) d\epsilon. It is essential to recover the variational character of the functional. In this sense it is a correction to the usual definition of eband needed in the metallic case. Since it generates confusion "correction for metals" is probably an unlucky name. If the smearing function is chosen to be Fermi-Dirac and degauss is interpreted as a temperature it happens that the variational quantity at finite temperature is the free energy and this is the quantity computed and printed as total energy by the code. I agree total free energy would be more appropriate in this case. Demet in this case equals -TS. However, if the smearing is Gaussian, cold-smearing or Methfessel-Paxton there is no real temperature around and no need to talk about free energy and my feeling is that calling it total free energy would also generate some confusion. The termodynamic analogue is sometimes useful to understand, for instance, why the "total (free) energy" decreases quadratically with increasing "temperature" with gaussian and F-D smearing while is much flatted with M-V or M-P smearing. Note that in the case of Gaussian or F-D smearing, although the quantity (F+E)/2 = E - TS/2 is closer to the T=0 limit, it SHOULD NOT be used as a the estimate for the energy since it IS NOT variational and therefore not consistent with computed forces . One should use instead M-V or M-P smearing that are both more insensitive to the value of smearing AND variational. Finally, in order to perform accurate finite temperature calcualtion with reasonable number of k-points one should combine M-V or M-P smearing (in order to speed up k-point convergence) with F-D (in order to have finite temperature). This is probably what is done in the paper by Gonze and coworkers mentioned by Nicola (I did not read it yet so I may be wrong). This is presently not implemented in PWSCF but it should be rather straightforward to use as smearing width the convolution of a M-V/M-P function (with a given degauss) and a F-D function (with the desired physical temperature). In this case the "total energy" printed by PWSCF would be the approximate "total free energy" but the "correction for metals" will not be immediately related to -TS (one should deconvolute it somehow) . Since this issues keep generating confusion among users, I agree we should make an effort to clarify them. stefano Paul Tangney wrote: >Thanks Nicola, > >I think that answered my question. >If you're right and what PWSCF calls "total energy" >is E-TS, then "demet" is -TS and is incorrectly >called the "correction for metals" in the code. >As you say, the "correction" should be -TS/2. > >If F-D smearing is not a good computational device >then definitely, I agree, the code should call this >"Total Energy" "Free Energy" instead. Otherwise, if F-D >is to be used purely as a smearing technique, the >correct correction should be used...i.e. "demet" >should be changed in the code from -TS to -TS/2. > >I'm interested in using F-D in a *physical* way. I want F = E - TS. >So I guess I don't need to change the code at all. > >By the way, on the system that I'm studying (crystalline >Tellurium), the application of a finite temperature actually >*lowers* the free energy. Obviously this is unphysical....although >perhaps not totally surprising, given that I'm using a finite >(but dense : 10x10x10) k-point grid. Its still a little worrying though. > >Regards, > >Paul > > > > From shaposh at isp.nsc.ru Mon Oct 31 12:30:23 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Mon, 31 Oct 2005 17:30:23 +0600 Subject: [Pw_forum] CVS cp.x Message-ID: <1130758223.3422.22.camel@m00> Hellow to all. CVS (3.0) version of cp.x does not seem to be functional, am i right? It always terminates with "Interpolation tables recalculation not implemented" error. Best Regards, Alexander Shaposhnikov From yang_jun-jun at hotmail.com Mon Oct 31 14:38:56 2005 From: yang_jun-jun at hotmail.com (Jun-jun Yang) Date: Mon, 31 Oct 2005 05:38:56 -0800 Subject: [Pw_forum] SCF convergency In-Reply-To: <20051030163532.64336.qmail@web60322.mail.yahoo.com> Message-ID: Eyvaz and list users, Thank you for your timely response. I changed some of the parameters, and now the calculation converges within 50 iterations (conv_thr=1.0e-6). But I noticed that the total energy is a little higher than before (-5807 Ry vs. -5865 Ry). It seems that it's most likely a local minimum. Should I decrease the convergence threshold to 1e-7 or 1e-8? Probably, I need to start the calculation from another point to avoid the local minimum. How can we do this in PWScf? In my calculation, as you suggested, I set start_magnetization to 0.9 for Fe, but I really don't know what number I should give for Cr. I just set 0.9 for Cr, 0.0 for O. Are there any guideline to set these values? Regards, Yang, Jun-jun >From: Eyvaz Isaev >Reply-To: pw_forum at pwscf.org >To: pw_forum at pwscf.org >Subject: Re: [Pw_forum] SCF convergency >Date: Sun, 30 Oct 2005 08:35:32 -0800 (PST) > >Hi, > >Hopefully you know that celldm(1)=15.8321 is in atomic >units, but not Angstrom. > >You asked to do spin-polarized calculations (nspin=2), >but put initial magnetic splitting to zero. >So, you have to change >starting_magnetization(1) = 0.9 and add more 2 same >lines for Cr and O atoms (your first atomic type is >Fe). Presumably, you should not specify the bands >number, so, remove nbnd=330. Be careful with this >number, it means that you have around 600-660 >electrons in your crystal. If you are sure that your >nbnd covers bands above the Fermi level,then it is OK. > > >Ecutrho seems to be low, increase up to 200Ry. >Try mixing_beta=0.1. >degauss=0.5? It is very large. Try degauss=0.03, it is >in Ry. >To me conv_thr=1.d-5 seems too large. Default conv_thr >(1d-6) looks more acceptable. > >Finally, if something went OK, be ready for LDA+U >calculations. Fe-Cr-O looks like a strongly correlated >system. > >Best regards, >Eyvaz. > > > &system > > ibrav = 1, celldm(1) = 15.8321, nat = 56, ntyp = > > 3 > > nbnd = 330 > > nspin = 2, degauss = 0.5 > > occupations = 'smearing' > > smearing = 'gaussian' > > starting_magnetization(1) = 0.0 > > ecutwfc = 25.0 > > ecutrho = 140.0 > > / > > &electrons > > diagonalization = 'cg' > > mixing_mode = 'plain' > > mixing_beta = 0.40 > > conv_thr = 1.0d-5 > > / > > > > ATOMIC_SPECIES > > Fe 55.847 Fe.pbe-sp-van.UPF > > Cr 51.996 Cr.pbe-sp-van.UPF > > O 15.999 O.pbe-van_ak.UPF > > > > ATOMIC_POSITIONS angstrom > > Fe 0.00000 0.00000 0.00000 > > Cr 5.23625 5.23625 5.23625 > > O 3.23391 3.23391 3.23391 > > Fe 0.00000 4.18900 4.18900 > > Fe 4.18900 0.00000 4.18900 > > [...] > > > > K_POINTS automatic > > 1 1 1 0 0 0 > > > > > > I have sent this mail three times with a graph > > attached. The first two posts > > had been rejected > > because size of the mail was a little larger than > > the limit. The third one > > should be below 40KB, > > but I haven't seen its appearing on the list yet. > > So, I have to post it > > again without a graph. I > > apologize for that you may receive two mails with > > similar content. > > > > >_________________________________________________________________ > > On the road to retirement? Check out MSN Life Events > > for advice on how to > > get there! > > >http://lifeevents.msn.com/category.aspx?cid=Retirement > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > > >__________________________________ >Yahoo! Mail - PC Magazine Editors' Choice 2005 >http://mail.yahoo.com >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum _________________________________________________________________ Don?t just search. Find. Check out the new MSN Search! http://search.msn.click-url.com/go/onm00200636ave/direct/01/ From eyvaz_isaev at yahoo.com Mon Oct 31 18:13:45 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Mon, 31 Oct 2005 09:13:45 -0800 (PST) Subject: [Pw_forum] SCF convergency In-Reply-To: Message-ID: <20051031171345.16335.qmail@web60316.mail.yahoo.com> Hi, --- Jun-jun Yang wrote: > But I noticed that the total energy is a > little higher than before (-5807 Ry > vs. -5865 Ry). It seems that it's most likely a > local minimum. Why you think about a local minima? > Should I decrease the convergence > threshold to 1e-7 or 1e-8? For test purpose, I assume, you can check lower conv_thr, too. > Probably, I need to start the calculation from > another point to avoid > the local minimum. How can we do this in PWScf? > Not so clear what what do you want. Please clarify. > In my calculation, as you suggested, I set > start_magnetization to 0.9 for > Fe, but I really don't > know what number I should give for Cr. I just set > 0.9 for Cr, 0.0 for O. Are > there any guideline > to set these values? starting_magnetization (and not start_magnetization) is not a magnetic moment. It is (rho_up - rho_dn)/(rho_up + rho_dn), obviously. See "starting_magnetization" section in /Doc/INPUT_PW. For O you can also specify non-zero starting_magnetization. Bests, Eyvaz. __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From konstantin_kudin at yahoo.com Mon Oct 31 21:19:50 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Mon, 31 Oct 2005 12:19:50 -0800 (PST) Subject: [Pw_forum] CVS cp.x In-Reply-To: <1130758223.3422.22.camel@m00> Message-ID: <20051031201951.86821.qmail@web52006.mail.yahoo.com> This code does work for me ... Could you run the examples that come with the code and see if they work? What is your architecture/libraries/compilers ? Kostya --- Alexander Shaposhnikov wrote: > Hellow to all. > > CVS (3.0) version of cp.x does not seem to be functional, > am i right? It always terminates with "Interpolation tables > recalculation not implemented" error. > > Best Regards, > Alexander Shaposhnikov > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com From konstantin_kudin at yahoo.com Mon Oct 31 22:33:15 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Mon, 31 Oct 2005 13:33:15 -0800 (PST) Subject: [Pw_forum] mixed 2nd derivatives - wrt atomic positions and strain In-Reply-To: <20051031201951.86821.qmail@web52006.mail.yahoo.com> Message-ID: <20051031213315.59595.qmail@web52008.mail.yahoo.com> Hi there, 2nd derivatives wrt (with respect to) atomic positions twice are force constants, and can be used to compute phonons. 2nd derivatives wrt strain twice give stress tensor. Does anyone know a good reference where I could read about mixed 2nd derivatives of these two? Thanks! Kostya __________________________________ Yahoo! FareChase: Search multiple travel sites in one click. http://farechase.yahoo.com