From luch001 at 126.com Sun Oct 1 03:52:46 2006 From: luch001 at 126.com (luch001) Date: Sun, 1 Oct 2006 09:52:46 +0800 (CST) Subject: [Pw_forum] LATTICE VECTOR DEFINATION Message-ID: <451F1F6E.000125.14752@bj126app2.126.com> Dear all, I am so puzzled by the definition of the lattice vector for monoclinic lattice in PWSCF package. Would you kindly tell me the definition or where the definition is given in the PWSCF package. Any reply will be appreciated. Best wishes. Yours sincerely Jingyun Zhang National Lab of Superhard Materials, Jilin University, Changchun 130012, P. R. China. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061001/dfc3090c/attachment.htm From luch001 at 126.com Sun Oct 1 03:22:26 2006 From: luch001 at 126.com (luch001) Date: Sun, 1 Oct 2006 09:22:26 +0800 (CST) Subject: [Pw_forum] LATTICE VECTOR DEFINATION Message-ID: <451F1852.000112.13326@bj126app2.126.com> Dear all users, I am so puzzled by the definition of the lattice vector for monoclinic lattice in PWSCF package. Would you kindly tell me the definition or where the definition is given in the package. Any reply will be appreciated. Best wishes. Yours sincerely Jingyun Zhang National Lab of Superhard Materials, Jilin University, Changchun 130012, P. R. China. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061001/4047518f/attachment.htm From luch001 at 126.com Sun Oct 1 04:10:34 2006 From: luch001 at 126.com (luch001) Date: Sun, 1 Oct 2006 10:10:34 +0800 (CST) Subject: [Pw_forum] lattice vector defination Message-ID: <451F239A.00008C.19659@bj126app8.126.com> Dear all , Sorry for the repeating of my question. I am so puzzled by the definition of the lattice vector for monoclinic lattice (especially the base-centered monoclinic ) in PWSCF package. Would you kindly tell me the definition or where the definition is given in the package. Any reply will be appreciated. Best wishes. Yours sincerely Jingyun Zhang National Lab of Superhard Materials, Jilin University, Changchun 130012, P. R. China. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061001/9dd5d4e8/attachment.htm From giannozz at nest.sns.it Sun Oct 1 05:01:12 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 1 Oct 2006 05:01:12 +0200 Subject: [Pw_forum] lattice vector defination In-Reply-To: <451F239A.00008C.19659@bj126app8.126.com> References: <451F239A.00008C.19659@bj126app8.126.com> Message-ID: On Oct 1, 2006, at 4:10 , luch001 wrote: > I am so puzzled by the definition of the lattice vector for > monoclinic > lattice (especially the base-centered monoclinic ) in PWSCF package. in file Doc/INPUT_PW, towards the end, the definition of lattice vectors is explained for all lattices. Paolo From baroni at sissa.it Sun Oct 1 10:53:16 2006 From: baroni at sissa.it (Stefano Baroni) Date: Sun, 1 Oct 2006 10:53:16 +0200 Subject: [Pw_forum] lattice vector defination In-Reply-To: <451F239A.00008C.19659@bj126app8.126.com> References: <451F239A.00008C.19659@bj126app8.126.com> Message-ID: Sorry for repeating the answer: in file Doc/INPUT_PW, towards the end, the definition of lattice vectors is explained for all lattices. ;-) take care and have a nice (working) week end Stefano On Oct 1, 2006, at 4:10 AM, luch001 wrote: > > > > Dear all , > Sorry for the repeating of my question. I am so puzzled by the > definition of the lattice vector for monoclinic > > lattice (especially the base-centered monoclinic ) in PWSCF package. > > Would you kindly tell me the definition or where the definition is > given in the package. > Any reply will be appreciated. > > Best wishes. > > Yours sincerely > Jingyun Zhang > > > National Lab of Superhard Materials, Jilin University, Changchun > 130012, P. R. China. > > > > > > > > > 3G ? ? ? ? ??? ? ? ? ? ? ? ? ? ? > ? ? ? ? ? 3G ? ? ? ? ? ? ?280 ? ? ? ? ? ? ? ? > ? ? ? ? ? --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061001/15af0825/attachment.htm From giannozz at nest.sns.it Sun Oct 1 14:47:54 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 1 Oct 2006 14:47:54 +0200 Subject: [Pw_forum] lattice vector defination In-Reply-To: <451F239A.00008C.19659@bj126app8.126.com> References: <451F239A.00008C.19659@bj126app8.126.com> Message-ID: On Oct 1, 2006, at 4:10 , luch001 wrote: > Would you kindly tell me [...] where the definition is given in the > package. subroutine flib/latgen.f90 Paol From wierzbom at ts.infn.it Mon Oct 2 12:08:58 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Mon, 2 Oct 2006 12:08:58 +0200 (CEST) Subject: [Pw_forum] a question about electron-phonon coupling In-Reply-To: References: Message-ID: On Fri, 29 Sep 2006, li quan wrote: >> >> Dear Gosia, > > > What's your meaning of "calculate some "grid" of phonons and use > q2r and matdyn"? I can not understand it quite clearly. Would you explain it > more explicitly? I mean: 1) to calculate el-ph at many phonon vectors q (grid of q points) 2) to use dynamical matrix and el-ph element (g_k,k+q for i,j bands and mu,nu modes) in programs from pwtools: q2r and matdyn -- check example07 > How to apply the acoustic sum rule in the el-ph calculation? By setting "zasr" in the input to q2r and matdyn (check example07) > Is it suitable for acoustic phonons? Yes, 'zasr' is designed for accustic phonons because those are zero at q=0 and those are not accurate at small q. Gosia From lanhaiping at gmail.com Mon Oct 2 14:37:05 2006 From: lanhaiping at gmail.com (lan haiping) Date: Mon, 2 Oct 2006 20:37:05 +0800 Subject: [Pw_forum] How to consider finite system 's deformation with PW method? Message-ID: Dear All, I wanna calculate single nanotube's radial deformation with QE . But most works i knew are performed on nanotube bundle . so i wanto know whether it is possible to study single nanotube's deformation with QE code . Regards, Hai-Ping -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061002/86f3d347/attachment.htm From baroni at sissa.it Mon Oct 2 15:10:50 2006 From: baroni at sissa.it (Stefano Baroni) Date: Mon, 2 Oct 2006 15:10:50 +0200 Subject: [Pw_forum] How to consider finite system 's deformation with PW method? In-Reply-To: References: Message-ID: The answer is YES, the same way as for an isolated molecule (i.e. using supercells). S. On Oct 2, 2006, at 2:37 PM, lan haiping wrote: > Dear All, > > I wanna calculate single nanotube's radial deformation with QE . > But most works i knew are performed on nanotube bundle . > so i wanto know whether it is possible to study single nanotube's > deformation with QE code . > > Regards, > > Hai-Ping > > > > > > --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061002/8e40a9b2/attachment.htm From lanhaiping at gmail.com Mon Oct 2 18:52:30 2006 From: lanhaiping at gmail.com (lan haiping) Date: Tue, 3 Oct 2006 00:52:30 +0800 Subject: [Pw_forum] How to consider finite system 's deformation with PW method? In-Reply-To: References: Message-ID: Dear Stefano , Sorry for my misexpression of my previous question. What i really want to ask is whether QE can deal with dynamical properties of finite system under certain condition such as external pressure . For periodic crystal system, as far as i know, QE can perform calculation under external hystrostatic pressure. For single system such as nanotube with supercell, does this external hystrostatic pressure scheme still hold ? Regards, Hai-Ping On 10/2/06, Stefano Baroni wrote: > > The answer is YES, the same way as for an isolated molecule (i.e. using > supercells). S. > > On Oct 2, 2006, at 2:37 PM, lan haiping wrote: > > Dear All, > > I wanna calculate single nanotube's radial deformation with QE . > But most works i knew are performed on nanotube bundle . > so i wanto know whether it is possible to study single nanotube's > deformation with QE code . > > Regards, > > Hai-Ping > > > > > > > > > --- > Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste > [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) > > Please, if possible, don't send me MS Word or PowerPoint attachments > Why? See: http://www.gnu.org/philosophy/no-word-attachments.html > > > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061003/dfab6629/attachment.htm From marzari at MIT.EDU Mon Oct 2 20:25:31 2006 From: marzari at MIT.EDU (Nicola Marzari) Date: Mon, 02 Oct 2006 14:25:31 -0400 Subject: [Pw_forum] How to consider finite system 's deformation with PW method? In-Reply-To: References: Message-ID: <4521599B.1000102@mit.edu> lan haiping wrote: > Dear Stefano , > Sorry for my misexpression of my previous question. > What i really want to ask is whether QE can deal with dynamical > properties of finite system under certain condition such > as external pressure . > For periodic crystal system, as far as i know, QE can perform > calculation under external hystrostatic pressure. For single system such > as nanotube with supercell, does this external hystrostatic pressure > scheme still hold ? > > Regards, > > Hai-Ping Dear Hai-Ping, to apply pressure on an extended/bulk system you usually squeeze the unit cell (Parrinello-Rahman, see PRL early '80s, or recent work with Martonak and metadynamics). For a finite systems, you squeeze a unit cell filled up with a compressible medium (e.g. Martonak PRL 00), or, as recently proposed by us (Cococcioni PRL 05) you use a modified electronic enthalphy functional (The electronic enthalpy functional will soon appear in CP). In the meanwhile, you might want to look at the work of Tangney on nanotubes under pressure. All the best, 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 2586534 marzari at mit.edu http://quasiamore.mit.edu From suriano at sissa.it Mon Oct 2 21:01:17 2006 From: suriano at sissa.it (Antonio Suriano) Date: Mon, 02 Oct 2006 21:01:17 +0200 Subject: [Pw_forum] pwscf 3.* good for production? Message-ID: <452161FD.5000107@sissa.it> Hi i am moving from pwsc 2.* to 3.* to take advantage of parallel diagonalization and other stuff. Is it numerically stable? Results are consistent to previous version? I am really curious to know what are real experiences that you had about transition from the previous version to the actual. -- MAY ALL YOUR DREAMS COME TRUE! Chiacchiera con i tuoi amici in tempo reale! http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com From giannozz at nest.sns.it Mon Oct 2 21:40:48 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 2 Oct 2006 21:40:48 +0200 Subject: [Pw_forum] pwscf 3.* good for production? In-Reply-To: <452161FD.5000107@sissa.it> References: <452161FD.5000107@sissa.it> Message-ID: On Oct 2, 2006, at 21:01 , Antonio Suriano wrote: > i am moving from pwsc 2.* to 3.* to take advantage of parallel > diagonalization and other stuff. for parallel (subspace) diagonalization I would suggest to use the CVS version > Is it numerically stable? only on Monday and Wednesday. Avoid calculations on Tuesday and Friday. On Thursday, it depends on the phase of the moon. > Results are consistent to previous version? no, they are completely different, so that we can re-write old papers with new results. The ground state structure of Silicon, for istance, is a fullerene in v.3.0 and a nanotube in v.3.1. It is so much better than the diamond structure, sooo boring. Paolo Disclaimer: only the first sentence is serious From stewart at cnf.cornell.edu Mon Oct 2 23:02:10 2006 From: stewart at cnf.cornell.edu (stewart at cnf.cornell.edu) Date: Mon, 02 Oct 2006 17:02:10 -0400 Subject: [Pw_forum] 2nd Annual CNF Fall Workshop =?utf-8?B?IkJ1aWxkaW5n?= Nanostructures Bit by =?utf-8?B?Qml0Ig==?= In-Reply-To: References: <452161FD.5000107@sissa.it> Message-ID: <20061002210210.20529.qmail@xuxa.iecc.com> Dear Quantum Espresso users, I am happy to announce the 2nd Annual Cornell Nanoscale Facility Fall Workshop, ?Modeling Nanostructures Bit by Bit?. This three-day workshop will provide morning lectures on the theory behind approaches followed by hands-on sessions where participants get to work directly with the codes. This year, Dr. Nicola Marzari (MIT) and Dr. Axel Kohlmeyer (University of Pennsylvania) who contribute often to this board will be serving as invited speakers to discuss PWscf and also Carr-Parrinello molecular dynamics. In addition, we will have sessions on real space density functional approaches, classical molecular dynamics, nanophotonics, and microfluidics. To find out more, please visit: http://www.cnf.cornell.edu/cnf_fallworkshop2006.html The workshop runs from Oct. 23-25th. The registration deadline is Oct. 11th and the number of participants is limited, so please register early. Please do not hesitate to contact me if you have any questions. I hope to see many of you at Cornell this fall! Best regards, Derek **This workshop is made possible by funding through the National Nanotechnology Infrastructure Network (http://www.nnin.org) and the National Science Foundation.** ################################ 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 suriano at sissa.it Tue Oct 3 00:51:45 2006 From: suriano at sissa.it (Antonio Suriano) Date: Tue, 03 Oct 2006 00:51:45 +0200 Subject: [Pw_forum] pwscf 3.* good for production? In-Reply-To: References: <452161FD.5000107@sissa.it> Message-ID: <45219801.2080206@sissa.it> Paolo Giannozzi ha scritto: > >> Is it numerically stable? > > only on Monday and Wednesday. Avoid calculations on Tuesday > and Friday. On Thursday, it depends on the phase of the moon. > Which time zone? The one of the cluster, or the one of the user? :-) I apologize if my question was silly, but i explain it better. I just wanted to ask if the improvement in the algorithms could "translate" by some mRy results for one atom calculation. Moving this change to a cell containing 50 atoms could lead to a consistent change if the numerical gaps are systematic. The physical results (minima, adsorpion, phonon) remain unchanged even if this happens. In that case you shouldn't for example plot in the same table energies that are obtained with different versions of the program. I just wanted to know if in the wide community of pw users someone has discovered that is better not to compare absolute values that come from different branches. I suppose now that there is no risk in doing that. -- MAY ALL YOUR DREAMS COME TRUE! Chiacchiera con i tuoi amici in tempo reale! http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com From giannozz at nest.sns.it Tue Oct 3 03:38:08 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 3 Oct 2006 03:38:08 +0200 Subject: [Pw_forum] pwscf 3.* good for production? In-Reply-To: <45219801.2080206@sissa.it> References: <452161FD.5000107@sissa.it> <45219801.2080206@sissa.it> Message-ID: On Oct 3, 2006, at 24:51 , Antonio Suriano wrote: > I just wanted to ask if the improvement in the algorithms could > "translate" by some mRy results for one atom calculation. of course it can. Algorithmic changes in the calculation of matrix elements of pseudopotentials - integration, interpolation etc - can lead to minor differences in absolute energies (but not in energy differences or in physically significant quantities). The same applies to exchange-correlation functionals, where various parameters are used to decide when to cut off diverging terms. Even purely numerical differences between different machines can lead to slightly different total energies. Even rearranging do loops may lead to small differences. As a rule it is safer to use the same version of a code to perform a set of calculations. If it is needed or useful to move to another version, one has to be careful when calculating quantities (such as cohesive energies) that depend on small differences of large numbers. Paolo From stant at nc.rr.com Tue Oct 3 16:38:01 2006 From: stant at nc.rr.com (Stan Toporek) Date: Tue, 3 Oct 2006 10:38:01 -0400 Subject: [Pw_forum] Configure and Compile problems on Solaris 10 Message-ID: <20061003103801.mk894qo0gg4sw0cg@webmail.spamcop.net> All, I searched the forum for an answer to this problem but I can up with nothing. I am using Sun's compilers (SunStudio 11) on Solaris 10. When I run configure I get the following problems: checking dependencies... fft_scalar.o : @MKL_DFT_TYPE@ WARNING: dependencies not found in directory Modules directory clib : ok symrho_mag.o : @CHAR@ WARNING: dependencies not found in directory PW but configure reported success. Now I try to compile the mods and get: f90 -fast -O1 -nodepend -xvector=no -xchip=ultra3 -xarch=v8plusb -xlic_lib=sunperf -D__SUN -I../include -M. -M../Modules -M../PW -M../PH -M../iotk/src -M../CPV -c fft_base.F90 -o fft_base.o make: Fatal error: Don't know how to make target `@MKL_DFT_TYPE@' Current working directory /home/stan/download/espresso-3.1.1/Modules *** Error code 1 make: Fatal error: Command failed for target `mods' I thought the sunperf library was all I needed to build espresso. Am I missing something or is this a bug? Thanks, Stan -- Stan Toporek,MS Martin Group Graduate Student and NMR Servant North Carolina State University Department of Chemistry ----- "The visible world is the invisible organization of Energy" --Physicist Heinz Pagels There are only 10 types of people in the world - those who understand binary, and those who don't. From giannozz at nest.sns.it Tue Oct 3 16:54:28 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 3 Oct 2006 16:54:28 +0200 Subject: [Pw_forum] Configure and Compile problems on Solaris 10 In-Reply-To: <20061003103801.mk894qo0gg4sw0cg@webmail.spamcop.net> References: <20061003103801.mk894qo0gg4sw0cg@webmail.spamcop.net> Message-ID: On Oct 3, 2006, at 16:38 , Stan Toporek wrote: > I am using Sun's compilers (SunStudio 11) on Solaris 10. > When I run configure I get the following problems: > > checking dependencies... > fft_scalar.o : @MKL_DFT_TYPE@ > WARNING: dependencies not found in directory Modules > directory clib : ok > symrho_mag.o : @CHAR@ > WARNING: dependencies not found in directory PW remove the line fft_scalar.o : @MKL_DFT_TYPE@ from Modules/make.depend; remove the line symrho_mag.o : @CHAR@ from PW/make.depend; compile If you know enough about unix, please look into the script "makedeps.sh" and try to figure out why it doesn't work on Solaris 10 (it works on most machines) and how to make it work Paolo From mmiao at wsu.edu Tue Oct 3 21:09:26 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 12:09:26 -0700 Subject: [Pw_forum] Compilation on SGI Origin 300 Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034E84@EXCHANGEVS-04.ad.wsu.edu> Hi, I tried to compile Espresso-3.1.1 on a SGI origin 300 machine. The compiler is SGI MIPSpro Fortran 90 Version 7.41. I got the following error messages while it compiles the realus.f90. MODULE realus ^ f90-855 f90: ERROR REALUS, File = realus.f90, Line = 1, Column = 8 The compiler has detected errors in module "REALUS". No module information file will be created for this module. USE ions_base, ONLY : nat, ntyp => nsp, ityp ^ f90-1015 f90: ERROR NEWD_R, File = realus.f90, Line = 413, Column = 39 Local-name "NTYP" must only be referenced once in a rename-list in this scope. USE ions_base, ONLY : nat, ntyp => nsp, ityp ^ f90-1015 f90: ERROR ADDUSDENS_R, File = realus.f90, Line = 857, Column = 41 Local-name "NTYP" must only be referenced once in a rename-list in this scope. f90: SGI MIPSpro Fortran 90 Version 7.41 (f14) Tue Oct 3, 2006 11:48:25 f90: 1081 source lines f90: 3 Error(s), 0 Warning(s), 0 Other message(s), 0 ANSI(s) cf90: "explain cf90-message number" gives more information about each message *** Error code 2 (bu21) *** Error code 1 (bu21) Thnaks for your help, Maosheng From giannozz at nest.sns.it Tue Oct 3 21:45:19 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 3 Oct 2006 21:45:19 +0200 Subject: [Pw_forum] Compilation on SGI Origin 300 In-Reply-To: <13EE5E22F052EB4BBBAA85A648AEFEA8034E84@EXCHANGEVS-04.ad.wsu.edu> References: <13EE5E22F052EB4BBBAA85A648AEFEA8034E84@EXCHANGEVS-04.ad.wsu.edu> Message-ID: On Oct 3, 2006, at 21:09 , Miao, Maosheng wrote: > I tried to compile Espresso-3.1.1 on a SGI origin 300 machine. The > compiler is SGI MIPSpro Fortran 90 Version 7.41. I got the following > error messages while it compiles the realus.f90 > [...] > USE ions_base, ONLY : nat, ntyp => nsp, ityp > ^ > f90-1015 f90: ERROR NEWD_R, File = realus.f90, Line = 413, Column = 39 > Local-name "NTYP" must only be referenced once in a rename-list in > this scope. > > USE ions_base, ONLY : nat, ntyp => nsp, ityp > ^ > f90-1015 f90: ERROR ADDUSDENS_R, File = realus.f90, Line = 857, > Column = > 41 > Local-name "NTYP" must only be referenced once in a rename-list in > this scope. local name "NTYP" IS correcly referenced, whatever the compiler may say. Since however it is not actualy used in either routines, just remove it. Paolo From mmiao at wsu.edu Tue Oct 3 23:53:43 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 14:53:43 -0700 Subject: [Pw_forum] Compilation on SGI Origin 300 Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034E9E@EXCHANGEVS-04.ad.wsu.edu> Hi Paolo, Thanks. But it doesn't work. And it seems ityp is used in the subroutine. The following is the error messgae. MODULE realus ^ f90-855 f90: ERROR REALUS, File = realus.f90, Line = 1, Column = 8 The compiler has detected errors in module "REALUS". No module information file will be created for this module. nt = ityp(na) ^ f90-232 f90: ERROR NEWD_R, File = realus.f90, Line = 435, Column = 13 IMPLICIT NONE is specified in the local scope, therefore an explicit type must be specified for function "ITYP". USE ions_base, ONLY : nat => nsp ^ f90-1015 f90: ERROR ADDUSDENS_R, File = realus.f90, Line = 857, Column = 36 Local-name "NAT" must only be referenced once in a rename-list in this scope. nt = ityp(na) ^ f90-232 f90: ERROR ADDUSDENS_R, File = realus.f90, Line = 891, Column = 14 IMPLICIT NONE is specified in the local scope, therefore an explicit type must be specified for function "ITYP". f90: SGI MIPSpro Fortran 90 Version 7.41 (f14) Tue Oct 3, 2006 14:30:26 f90: 1081 source lines f90: 4 Error(s), 0 Warning(s), 0 Other message(s), 0 ANSI(s) cf90: "explain cf90-message number" gives more information about each message *** Error code 2 (bu21) *** Error code 1 (bu21) Thanks, Maosheng -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Tuesday, October 03, 2006 12:45 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] Compilation on SGI Origin 300 On Oct 3, 2006, at 21:09 , Miao, Maosheng wrote: > I tried to compile Espresso-3.1.1 on a SGI origin 300 machine. The > compiler is SGI MIPSpro Fortran 90 Version 7.41. I got the following > error messages while it compiles the realus.f90 [...] > USE ions_base, ONLY : nat, ntyp => nsp, ityp > ^ > f90-1015 f90: ERROR NEWD_R, File = realus.f90, Line = 413, Column = 39 > Local-name "NTYP" must only be referenced once in a rename-list in > this scope. > > USE ions_base, ONLY : nat, ntyp => nsp, ityp > ^ > f90-1015 f90: ERROR ADDUSDENS_R, File = realus.f90, Line = 857, Column > = > 41 > Local-name "NTYP" must only be referenced once in a rename-list in > this scope. local name "NTYP" IS correcly referenced, whatever the compiler may say. Since however it is not actualy used in either routines, just remove it. Paolo _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From giannozz at nest.sns.it Wed Oct 4 00:10:02 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 4 Oct 2006 00:10:02 +0200 Subject: [Pw_forum] Compilation on SGI Origin 300 In-Reply-To: <13EE5E22F052EB4BBBAA85A648AEFEA8034E9E@EXCHANGEVS-04.ad.wsu.edu> References: <13EE5E22F052EB4BBBAA85A648AEFEA8034E9E@EXCHANGEVS-04.ad.wsu.edu> Message-ID: <6E4AAD89-1583-4F1B-B228-D8E97143828E@nest.sns.it> On Oct 3, 2006, at 23:53 , Miao, Maosheng wrote: > Thanks. But it doesn't work. And it seems ityp is used in the > subroutine. who told you to remove the reference to ityp? you must remove "ntyp => nsp," from lines 413 and 857, as indicated in the previous message, and nothing else. P. From mmiao at wsu.edu Wed Oct 4 00:40:20 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 15:40:20 -0700 Subject: [Pw_forum] Compilation on SGI Origin 300 Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034EA9@EXCHANGEVS-04.ad.wsu.edu> Hi Paolo, Thanks. That works. But while I compile CP part, I encounter a similar problem that I can not apply the same trick. See the error message below: MODULE efield_module ^ f90-855 f90: ERROR EFIELD_MODULE, File = efield.f90, Line = 8, Column = 8 The compiler has detected errors in module "EFIELD_MODULE". No module information file will be created for this module. USE uspp, ONLY: betae => vkb, deeq ^ f90-1015 f90: ERROR DFORCE_EFIELD, File = efield.f90, Line = 153, Column = 21 Local-name "BETAE" must only be referenced once in a rename-list in this scope. USE uspp, ONLY: betae => vkb ^ f90-1015 f90: ERROR BERRY_ENERGY2, File = efield.f90, Line = 252, Column = 21 Local-name "BETAE" must only be referenced once in a rename-list in this scope. USE uspp, ONLY: betae => vkb, deeq ^ f90-1015 f90: ERROR DFORCE_EFIELD2, File = efield.f90, Line = 271, Column = 21 Local-name "BETAE" must only be referenced once in a rename-list in this scope. f90: SGI MIPSpro Fortran 90 Version 7.41 (f14) Tue Oct 3, 2006 15:20:01 f90: 290 source lines f90: 4 Error(s), 0 Warning(s), 0 Other message(s), 0 ANSI(s) cf90: "explain cf90-message number" gives more information about each message *** Error code 2 (bu21) *** Error code 1 (bu21) Thanks, Maosheng -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Tuesday, October 03, 2006 3:10 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] Compilation on SGI Origin 300 On Oct 3, 2006, at 23:53 , Miao, Maosheng wrote: > Thanks. But it doesn't work. And it seems ityp is used in the > subroutine. who told you to remove the reference to ityp? you must remove "ntyp => nsp," from lines 413 and 857, as indicated in the previous message, and nothing else. P. _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From giannozz at nest.sns.it Wed Oct 4 00:57:35 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 4 Oct 2006 00:57:35 +0200 Subject: [Pw_forum] Compilation on SGI Origin 300 In-Reply-To: <13EE5E22F052EB4BBBAA85A648AEFEA8034EA9@EXCHANGEVS-04.ad.wsu.edu> References: <13EE5E22F052EB4BBBAA85A648AEFEA8034EA9@EXCHANGEVS-04.ad.wsu.edu> Message-ID: <48EFB211-6E81-4CCF-BF86-A02EA4D19451@nest.sns.it> On Oct 4, 2006, at 24:40 , Miao, Maosheng wrote: > Thanks. That works. But while I compile CP part, I encounter a similar > problem that I can not apply the same trick. are you sure? I think you can. Maybe you haven't understood the trick: references "something => something_else" that are not used confuse the compiler (it is a bug of the SGI compiler, of course) P. From giannozz at nest.sns.it Wed Oct 4 00:58:58 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 4 Oct 2006 00:58:58 +0200 Subject: [Pw_forum] Compilation on SGI Origin 300 In-Reply-To: <13EE5E22F052EB4BBBAA85A648AEFEA8034EA9@EXCHANGEVS-04.ad.wsu.edu> References: <13EE5E22F052EB4BBBAA85A648AEFEA8034EA9@EXCHANGEVS-04.ad.wsu.edu> Message-ID: <269ABFED-4CF9-40A9-8E8D-6922D2AB484D@nest.sns.it> and please make a list of all occurrences so that they can be removed from the code - Paolo From mmiao at wsu.edu Wed Oct 4 01:00:32 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 16:00:32 -0700 Subject: [Pw_forum] Compilation on SGI Origin 300 Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034EAB@EXCHANGEVS-04.ad.wsu.edu> Hi Paolo, I read efield.f90 and betae is used in at least two subroutines. Thanks, Maosheng -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Tuesday, October 03, 2006 3:58 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] Compilation on SGI Origin 300 On Oct 4, 2006, at 24:40 , Miao, Maosheng wrote: > Thanks. That works. But while I compile CP part, I encounter a similar > problem that I can not apply the same trick. are you sure? I think you can. Maybe you haven't understood the trick: references "something => something_else" that are not used confuse the compiler (it is a bug of the SGI compiler, of course) P. _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From mmiao at wsu.edu Wed Oct 4 01:01:42 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 16:01:42 -0700 Subject: [Pw_forum] Compilation on SGI Origin 300 Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034EAC@EXCHANGEVS-04.ad.wsu.edu> After I am done, I will send you that list. Thanks. Maosheng -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Tuesday, October 03, 2006 3:59 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] Compilation on SGI Origin 300 and please make a list of all occurrences so that they can be removed from the code - Paolo _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From mmiao at wsu.edu Wed Oct 4 01:36:13 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 16:36:13 -0700 Subject: [Pw_forum] Compilation on SGI Origin 300 Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034EB3@EXCHANGEVS-04.ad.wsu.edu> Hi Paolo, Slightly different trick works for efiald.f90. There are four subroutines defined betae. Two of them are not used it and can be removed. The first one of the other two works fine, but not the second one. I just change the name of the varible from betae to betaf throughout the whole subroutine and it works. Similar thing also happens for runcp.f90. Now the CP part has been compiled. Thanks, Maosheng -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Tuesday, October 03, 2006 3:59 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] Compilation on SGI Origin 300 and please make a list of all occurrences so that they can be removed from the code - Paolo _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From mmiao at wsu.edu Wed Oct 4 01:51:27 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 16:51:27 -0700 Subject: [Pw_forum] Compilation on SGI Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034EB4@EXCHANGEVS-04.ad.wsu.edu> Hi, Thanks for helping me through the compilation. I encountered another problem. I got error message while I do "make pp". test -d bin || mkdir bin if test -d iotk ; then \ ( cd iotk ; if test "make" = "" ; then make TLDEPS= libiotk.a ; \ else make TLDEPS= libiotk.a ; fi ) ; fi cd src ; make libiotk.a `libiotk.a' is up to date. ( cd Modules ; if test "make" = "" ; then make TLDEPS= all ; \ else make TLDEPS= all ; fi ) ( cd clib ; if test "make" = "" ; then make TLDEPS= all ; \ else make TLDEPS= all ; fi ) ( cd flib ; if test "make" = "" ; then make TLDEPS= all ; \ else make TLDEPS= all ; fi ) if test -d PW ; then \ ( cd PW ; if test "make" = "" ; then make TLDEPS= all ; \ else make TLDEPS= all ; fi ) ; fi test -n "" && ( cd .. ; make || exit 1) || : if test -d PP ; then \ ( cd PP ; if test "make" = "" ; then make TLDEPS= all ; \ else make TLDEPS= all ; fi ) ; fi test -n "" && ( cd .. ; make || exit 1) || : don't know how to make @iotk_module@ (bu42). *** Error code 1 (bu21) Same thing for "make d3" and "make pwcond". Any suggestion? Thanks, Maosheng From giannozz at nest.sns.it Wed Oct 4 01:56:51 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 4 Oct 2006 01:56:51 +0200 Subject: [Pw_forum] Compilation on SGI In-Reply-To: <13EE5E22F052EB4BBBAA85A648AEFEA8034EB4@EXCHANGEVS-04.ad.wsu.edu> References: <13EE5E22F052EB4BBBAA85A648AEFEA8034EB4@EXCHANGEVS-04.ad.wsu.edu> Message-ID: <1019B2F1-3DBF-438D-BEFF-39FB4369042B@nest.sns.it> On Oct 4, 2006, at 1:51 , Miao, Maosheng wrote: > > don't know how to make @iotk_module@ (bu42). > *** Error code 1 (bu21) > > Same thing for "make d3" and "make pwcond". Any suggestion? my suggestion would be to stop using such a crappy machine. In any event, it is the same problem that was posted a few hours ago on this mailing list, and the solution is the same, i.e. remove lines with @...@ from make.depend files P. From mmiao at wsu.edu Wed Oct 4 02:11:37 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Tue, 3 Oct 2006 17:11:37 -0700 Subject: [Pw_forum] Compilation on SGI Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034EB9@EXCHANGEVS-04.ad.wsu.edu> Hi Paolo, Thanks. I don't think SGI origin is crappy. Wahtsoever, we are thinking of updating our machine. Unfortunately it will be another SGI (Altix 4700). Will it be easier to load PWscf? Thanks, Maosheng -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Tuesday, October 03, 2006 4:57 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] Compilation on SGI On Oct 4, 2006, at 1:51 , Miao, Maosheng wrote: > > don't know how to make @iotk_module@ (bu42). > *** Error code 1 (bu21) > > Same thing for "make d3" and "make pwcond". Any suggestion? my suggestion would be to stop using such a crappy machine. In any event, it is the same problem that was posted a few hours ago on this mailing list, and the solution is the same, i.e. remove lines with @...@ from make.depend files P. _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From akohlmey at cmm.chem.upenn.edu Wed Oct 4 15:36:40 2006 From: akohlmey at cmm.chem.upenn.edu (Axel Kohlmeyer) Date: Wed, 4 Oct 2006 09:36:40 -0400 Subject: [Pw_forum] Compilation on SGI In-Reply-To: <13EE5E22F052EB4BBBAA85A648AEFEA8034EB9@EXCHANGEVS-04.ad.wsu.edu> References: <13EE5E22F052EB4BBBAA85A648AEFEA8034EB9@EXCHANGEVS-04.ad.wsu.edu> Message-ID: <7b6913e90610040636q51431758t93e0b6092c4b199a@mail.gmail.com> On 10/3/06, Miao, Maosheng wrote: > Hi Paolo, > > Thanks. I don't think SGI origin is crappy. Wahtsoever, we are thinking it is not the hardware (although the SGI oxygens are a sad exception), it is the compilers and the general software design (they used to have hundreds of X11 programs with suid root). i've been forced (=paid) to support various SGI machines throughout my career so far, and the benefits were never worth the hassle (with one exception). > of updating our machine. Unfortunately it will be another SGI (Altix > 4700). Will it be easier to load PWscf? yes and no. you will get a closer look at the quirkiness of the intel compilers. ;-) since they are mostly the same on all linux machines, there is more knowledge about it 'in the wild'. ;-) the good thing that comes out of your endeavor, is the the quantum espresso code gets better regardless. thank you for that. cheers, axel. > > Thanks, Maosheng > -- ======================================================================= Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu Center for Molecular Modeling -- University of Pennsylvania Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323 tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425 ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From L.Boeri at fkf.mpg.de Wed Oct 4 15:29:32 2006 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Wed, 4 Oct 2006 15:29:32 +0200 (CEST) Subject: [Pw_forum] Question about potential (pp tool) In-Reply-To: References: Message-ID: Dear PWSCF users, I am sorry for the somehow stupid question, but the more I think about it the more I get confused... The question is: what is exactly the "total" potential which is given by pp.x when plot_num is set to 1? The manual says: 1=total potential V_bare+V_H + V_xc my question is: what is V_bare? Is it the local potential or some sort of total average potential? Thanks a lot, Lilia From niuyingli at gmail.com Wed Oct 4 15:51:11 2006 From: niuyingli at gmail.com (Yingli Niu) Date: Wed, 4 Oct 2006 21:51:11 +0800 Subject: [Pw_forum] Bug in PH/clinear.f90 Message-ID: <6398ac910610040651i14b02f91l65c2d76d7aaf7b5c@mail.gmail.com> Dear all, It looks like a bug in espresso-3.1/PH/clinear.f90 and espresso-3.1.1/PH/clinear.f90: from line 14: nkr = nk2*nk2*nk3 I think it should be: nkr = nk1*nk2*nk3 Is it right? Best Regards From wierzbom at ts.infn.it Wed Oct 4 16:02:31 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Wed, 4 Oct 2006 16:02:31 +0200 (CEST) Subject: [Pw_forum] Bug in PH/clinear.f90 In-Reply-To: <6398ac910610040651i14b02f91l65c2d76d7aaf7b5c@mail.gmail.com> References: <6398ac910610040651i14b02f91l65c2d76d7aaf7b5c@mail.gmail.com> Message-ID: On Wed, 4 Oct 2006, Yingli Niu wrote: > Dear all, > It looks like a bug in espresso-3.1/PH/clinear.f90 and > espresso-3.1.1/PH/clinear.f90: > from line 14: > > nkr = nk2*nk2*nk3 > > I think it should be: > > nkr = nk1*nk2*nk3 > > Is it right? Thank you very much Yingli! You are right :-) Gosia > > Best Regards > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From niuyingli at gmail.com Wed Oct 4 16:16:08 2006 From: niuyingli at gmail.com (Yingli Niu) Date: Wed, 4 Oct 2006 22:16:08 +0800 Subject: [Pw_forum] Bug in PH/clinear.f90 In-Reply-To: References: <6398ac910610040651i14b02f91l65c2d76d7aaf7b5c@mail.gmail.com> Message-ID: <6398ac910610040716i2c9f855bq520028b6f1263e44@mail.gmail.com> That's all right! :-) Yingli On 10/4/06, Malgorzata Wierzbowska wrote: > On Wed, 4 Oct 2006, Yingli Niu wrote: > > > Dear all, > > It looks like a bug in espresso-3.1/PH/clinear.f90 and > > espresso-3.1.1/PH/clinear.f90: > > from line 14: > > > > nkr = nk2*nk2*nk3 > > > > I think it should be: > > > > nkr = nk1*nk2*nk3 > > > > Is it right? > > Thank you very much Yingli! You are right :-) > Gosia > > > > > > > Best Regards > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From degironc at sissa.it Wed Oct 4 16:18:23 2006 From: degironc at sissa.it (degironc) Date: Wed, 04 Oct 2006 16:18:23 +0200 Subject: [Pw_forum] Question about potential (pp tool) In-Reply-To: References: Message-ID: <4523C2AF.8010609@sissa.it> It is the total local potential which inside the core region is meaningless. Only the region outside the core is "physical" . stefano Lilia Boeri wrote: >Dear PWSCF users, > >I am sorry for the somehow stupid question, but the more I think about it >the more I get confused... >The question is: what is exactly the "total" potential which is given by >pp.x when plot_num is set to 1? >The manual says: >1=total potential V_bare+V_H + V_xc > >my question is: what is V_bare? Is it the local potential or some sort of >total average potential? > >Thanks a lot, >Lilia > > >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From L.Boeri at fkf.mpg.de Wed Oct 4 16:32:05 2006 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Wed, 4 Oct 2006 16:32:05 +0200 (CEST) Subject: [Pw_forum] Question about potential (pp tool) In-Reply-To: <4523C2AF.8010609@sissa.it> References: <4523C2AF.8010609@sissa.it> Message-ID: dear Stefano, thanks a lot. Reading another message that you wrote me some time ago, I understand that when a psp is read from PWSCF one of the l components is assumed as "local" potential, while the other channels are considered non-local and obtained from the first by projection. Is this correct? If this is only a numerical trick, the results should not depend on which channel I choose as "local", is this correct? thanks, Lilia On Wed, 4 Oct 2006, degironc wrote: > It is the total local potential which inside the core region is meaningless. > Only the region outside the core is "physical" . > stefano > > Lilia Boeri wrote: > > > Dear PWSCF users, > > > > I am sorry for the somehow stupid question, but the more I think about it > > the more I get confused... The question is: what is exactly the "total" > > potential which is given by pp.x when plot_num is set to 1? > > The manual says: > > 1=total potential V_bare+V_H + V_xc > > > > my question is: what is V_bare? Is it the local potential or some sort of > > total average potential? > > > > Thanks a lot, > > Lilia > > > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From wmbmacam at lg.ehu.es Wed Oct 4 17:48:05 2006 From: wmbmacam at lg.ehu.es (=?UTF-8?B?TWlndWVsIE1hcnTDrW5leiBDYW5hbGVz?=) Date: Wed, 04 Oct 2006 17:48:05 +0200 Subject: [Pw_forum] Pseudopotential troubles Message-ID: <4523D7B5.9050609@lg.ehu.es> Dear people, I'm interested in studying calcium at high pressure, close to 1Mbar. For that, I downloaded the available pseudopotentials in the web and did a few quick convergence tests. I found that the LDA pseudopotential had a notable energy jump (4 mRy) when going from a cutoff of 150 rydbergs to 200. I wasn't expecting this, due to a not so small r_c. Could be my fault, though. The pbe pseudopotential seems to predict a too small fcc->bcc transition pressure. What's more, running a phonon calculation of the simple cubic structure (experimentally stable, and in my calculations preferred over fcc and bcc) yields some instabilities. I assumed this could be related to the pseudopotential, though the pbe pseudopotential predicts a nice bandstructure and stable phonons (just a quick look) at 0 GPa. Thus, I intended to "cook" my pseudopotential. Here's the catch. I can't produce a Ca pseudopotential. Doing a RRKJ pseudization with a 4s or 4p r_c under 2.66 a_0 results in "error: too many nodes". Changing rho0 to 0.01 or 0.02 or 0.1 or even 1.0 doesn't solve it. Any clues? 4s and 4p nodes are both under 2 bohrs. To use Ca.pbe-nsp-van.UPF rcuts I need to force Troulliers-Martins pseudization. I can put one of those rcuts under 2.66 by pseudizing that l channel, but I want both rcuts small because my interest lies in high pressure. I've noted that if I change the grid settings to the Mg example one, I can reduce the 4p rcut slightly, but have to increase the 4s rcut to 2.68. Anyway, I have another more profound question. I don't understand why this pseudopotential input is terribly flawed (its result is). I'd be glad if some of you more knowledgeable people kindly pointed my errors. &input title='Ca', zed=20.0, xmin=-6.0, dx=0.0200, rmax=80, rel=1, beta=0.3, rlderiv=2.7, eminld=-4.0, emaxld=4.0, deld=0.02d0, nld=4, config='[Ar] 3d0 4s2 4p0', iswitch=3, dft='PBE', / &inputp pseudotype=3, lloc=-1, rcloc=2.1, tm =.false., nlcc =.true., rho0 =0.01, file_pseudopw='Ca.pbe-van.UPF', file_screen='ld1ps.screen' / 3 3D 3 2 0.00 0.00 2.40 2.40 4S 1 0 2.00 0.00 2.68 2.68 4P 2 1 0.00 0.00 2.59 2.59 Another question I have is how to add additional projectors per orbital, such as the ones used in Paolo Gianozzi's pbe pseudopotential. Thank you very much, Miguel PS: I'm also thinking on including an empty 4s orbital. ?Any comments? ---------------------------------------- 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 ---------------------------------------- From degironc at sissa.it Wed Oct 4 17:59:08 2006 From: degironc at sissa.it (degironc) Date: Wed, 04 Oct 2006 17:59:08 +0200 Subject: [Pw_forum] Question about potential (pp tool) In-Reply-To: References: <4523C2AF.8010609@sissa.it> Message-ID: <4523DA4C.4070305@sissa.it> If you are reading a psedopotential in the OLD fomat (not the UPF format) there is an array Vloc and then a number of DeltaV_l, and then there is also a variable called lloc that specify which l-component is considered as local. Inside the PW code then Vloc_at(r) = Vloc(r) + DeltaV_lloc(r) [this is the quantity that enters in Vbare referred to in pp.x] and the NON-LOCAL part is built from DeltaV_l - DeltaV_lloc. If the pseudopotential were used in the semilocal form (not in the KB form) different choices for lloc would all coincide for 0<= l <=lmax ( max ang component in the pseudo) and only angular component correspondig to l > lmax would be affected... this should not have large effects (this form is not anymore implemented in PW). In the KB fully non-local form different choices of lloc therefore produce slightly different pseudopotentials that all gives the same results on the atomic configuration used for the KB projection and should behave similarly provided they do not have GHOST states. Usually lloc=lmax (or lmax+1) would be the most "physical" choice but this is not always possible due to the ghosts problem above. Using UPF format this little algebra has alredy been done in the pseudopotential file and Vloc_at above is given direclty in the PP_LOCAL field while the KB projections = DeltaV*Psi_atomic are given in the PP_BETA fields and lloc is not needed anymore. Sorry if this explanation confuses you again... stefano Lilia Boeri wrote: >dear Stefano, > >thanks a lot. Reading another message that you wrote me some time ago, >I understand that when a psp is read from PWSCF one of the l components is >assumed as "local" potential, while the other channels are considered >non-local and obtained from the first by projection. Is this correct? > >If this is only a numerical trick, the results should not depend on which >channel I choose as "local", is this correct? >thanks, >Lilia > >On Wed, 4 Oct 2006, degironc wrote: > > > >>It is the total local potential which inside the core region is meaningless. >>Only the region outside the core is "physical" . >>stefano >> >>Lilia Boeri wrote: >> >> >> >>>Dear PWSCF users, >>> >>>I am sorry for the somehow stupid question, but the more I think about it >>>the more I get confused... The question is: what is exactly the "total" >>>potential which is given by pp.x when plot_num is set to 1? >>>The manual says: >>>1=total potential V_bare+V_H + V_xc >>> >>>my question is: what is V_bare? Is it the local potential or some sort of >>>total average potential? >>> >>>Thanks a lot, >>>Lilia >>> >>> >>>_______________________________________________ >>>Pw_forum mailing list >>>Pw_forum at pwscf.org >>>http://www.democritos.it/mailman/listinfo/pw_forum >>> >>> >>> >>> >>_______________________________________________ >>Pw_forum mailing list >>Pw_forum at pwscf.org >>http://www.democritos.it/mailman/listinfo/pw_forum >> >> >> >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From L.Boeri at fkf.mpg.de Wed Oct 4 18:04:33 2006 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Wed, 4 Oct 2006 18:04:33 +0200 (CEST) Subject: [Pw_forum] Question about potential (pp tool) In-Reply-To: <4523DA4C.4070305@sissa.it> References: <4523C2AF.8010609@sissa.it> <4523DA4C.4070305@sissa.it> Message-ID: dear Stefano, on the contrary it is very useful, thanks a lot. One last question: is there any way to visualize what the other (i.e. non-local) part of the scf potential look like in PWSCF? thanks, Lilia On Wed, 4 Oct 2006, degironc wrote: > If you are reading a psedopotential in the OLD fomat (not the UPF format) > there is an array Vloc and then a number of > DeltaV_l, and then there is also a variable called lloc that specify which > l-component is considered as local. > Inside the PW code then Vloc_at(r) = Vloc(r) + DeltaV_lloc(r) [this is the > quantity that enters in Vbare referred to in pp.x] and the NON-LOCAL part is > built from DeltaV_l - DeltaV_lloc. > If the pseudopotential were used in the semilocal form (not in the KB form) > different choices for lloc would all coincide for > 0<= l <=lmax ( max ang component in the pseudo) and only angular component > correspondig to l > lmax would be affected... this should not have large > effects (this form is not anymore implemented in PW). > In the KB fully non-local form different choices of lloc therefore produce > slightly different pseudopotentials that all gives the same results on the > atomic configuration used for the KB projection and should behave similarly > provided they do not have GHOST states. Usually lloc=lmax (or lmax+1) would > be the most "physical" choice but this is not always possible due to the > ghosts problem above. > > Using UPF format this little algebra has alredy been done in the > pseudopotential file and Vloc_at above is given direclty in the PP_LOCAL field > while the KB projections = DeltaV*Psi_atomic are given in the PP_BETA fields > and lloc is not needed anymore. > > Sorry if this explanation confuses you again... > > stefano From stant at nc.rr.com Wed Oct 4 18:04:40 2006 From: stant at nc.rr.com (Stan Toporek) Date: Wed, 4 Oct 2006 12:04:40 -0400 Subject: [Pw_forum] Linking of pw.x reported wrong ElF class Message-ID: <20061004120440.b993kq8sk08kkc0o@webmail.spamcop.net> All, Ok, I am the solaris 10 guinnea pig. I tried to do a make and got the following error during linking: ld: fatal: file pwscf.o: wrong ELF class: ELFCLASS64 ld: fatal: File processing errors. No output written to pw.x *** Error code 1 make: Fatal error: Command failed for target `pw.x' Current working directory /raid/espresso-3.1.1/PW *** Error code 1 make: Fatal error: Command failed for target `pw' Since this is the core program I think I have a problem. Now if I do a which on ld I get /usr/ucb/ld should I be using another linker this is the one Sunstudio 11 appears to use. Thanks, Stan From giannozz at nest.sns.it Wed Oct 4 18:26:13 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 4 Oct 2006 18:26:13 +0200 Subject: [Pw_forum] Question about potential (pp tool) In-Reply-To: References: <4523C2AF.8010609@sissa.it> <4523DA4C.4070305@sissa.it> Message-ID: <379876CD-2909-4733-9618-EE7207CDBB76@nest.sns.it> On Oct 4, 2006, at 18:04 , Lilia Boeri wrote: > One last question: is there any way to visualize what the other > (i.e. non-local) part of the scf potential look like in PWSCF? short answer: no. Since the nonlocal potentail is actually V(r, r') = \sum_n f_n(r) f*_n(r') , one could visualize the f_n(r) functions. They are short-ranged functions, centered around atoms, not really interesting I guess Paolo From wierzbom at ts.infn.it Wed Oct 4 18:26:57 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Wed, 4 Oct 2006 18:26:57 +0200 (CEST) Subject: [Pw_forum] Question about potential (pp tool) In-Reply-To: <4523DA4C.4070305@sissa.it> References: <4523C2AF.8010609@sissa.it> <4523DA4C.4070305@sissa.it> Message-ID: On Wed, 4 Oct 2006, degironc wrote: > In the KB fully non-local form different choices of lloc therefore produce > slightly different pseudopotentials that all gives the same results on the > atomic configuration used for the KB projection and should behave similarly > provided they do not have GHOST states. Usually lloc=lmax (or lmax+1) would > be the most "physical" choice but this is not always possible due to the > ghosts problem above. > And for the transition metals lloc=0. For the choice of lloc, such that there are no ghost states in KB-form, see Fuch and Scheffler CPC vol. 119, 67-98 (1999), (pages 81-83 explain lloc). Gosia > Using UPF format this little algebra has alredy been done in the > pseudopotential file and Vloc_at above is given direclty in the PP_LOCAL > field while the KB projections = DeltaV*Psi_atomic are given in the PP_BETA > fields and lloc is not needed anymore. > > Sorry if this explanation confuses you again... > > stefano > > Lilia Boeri wrote: > >> dear Stefano, >> >> thanks a lot. Reading another message that you wrote me some time ago, I >> understand that when a psp is read from PWSCF one of the l components is >> assumed as "local" potential, while the other channels are considered >> non-local and obtained from the first by projection. Is this correct? >> >> If this is only a numerical trick, the results should not depend on which >> channel I choose as "local", is this correct? thanks, >> Lilia >> >> On Wed, 4 Oct 2006, degironc wrote: >> >> >>> It is the total local potential which inside the core region is >>> meaningless. >>> Only the region outside the core is "physical" . >>> stefano >>> >>> Lilia Boeri wrote: >>> >>> >>>> Dear PWSCF users, >>>> >>>> I am sorry for the somehow stupid question, but the more I think about it >>>> the more I get confused... The question is: what is exactly the "total" >>>> potential which is given by pp.x when plot_num is set to 1? >>>> The manual says: >>>> 1=total potential V_bare+V_H + V_xc >>>> >>>> my question is: what is V_bare? Is it the local potential or some sort of >>>> total average potential? >>>> >>>> Thanks a lot, >>>> Lilia >>>> >>>> >>>> _______________________________________________ >>>> Pw_forum mailing list >>>> Pw_forum at pwscf.org >>>> http://www.democritos.it/mailman/listinfo/pw_forum >>>> >>>> >>>> >>> _______________________________________________ >>> Pw_forum mailing list >>> Pw_forum at pwscf.org >>> http://www.democritos.it/mailman/listinfo/pw_forum >>> >>> >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From giannozz at nest.sns.it Wed Oct 4 18:28:53 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 4 Oct 2006 18:28:53 +0200 Subject: [Pw_forum] Linking of pw.x reported wrong ElF class In-Reply-To: <20061004120440.b993kq8sk08kkc0o@webmail.spamcop.net> References: <20061004120440.b993kq8sk08kkc0o@webmail.spamcop.net> Message-ID: On Oct 4, 2006, at 18:04 , Stan Toporek wrote: > ld: fatal: file pwscf.o: wrong ELF class: ELFCLASS64 since ELF = Electron Localization Function seems unlikely, I guess that ELF = Executable and Linkable Format. This message means that pwscf.o has the wrong format (apparently 64-bit), so it was either produced by the wrong compiler, or by the right compiler with wrong options, or is being linked by the wrong linker, or by the right linker with bad options (or any combination of the above) Paolo From ferretti.andrea at unimore.it Wed Oct 4 18:37:01 2006 From: ferretti.andrea at unimore.it (Andrea Ferretti) Date: Wed, 4 Oct 2006 18:37:01 +0200 (CEST) Subject: [Pw_forum] Linking of pw.x reported wrong ElF class In-Reply-To: References: <20061004120440.b993kq8sk08kkc0o@webmail.spamcop.net> Message-ID: time ago I got a similar problem on a SGI machine; the configure script set the 64bit option in FFLAGS (due to a preset in the script, I guess) while the loader was only 32bit able. after I removed the 64bit flag it worked fine andrea On Wed, 4 Oct 2006, Paolo Giannozzi wrote: > On Oct 4, 2006, at 18:04 , Stan Toporek wrote: > > > ld: fatal: file pwscf.o: wrong ELF class: ELFCLASS64 > > since ELF = Electron Localization Function seems unlikely, I guess > that ELF = Executable and Linkable Format. This message means > that pwscf.o has the wrong format (apparently 64-bit), so it was > either produced by the wrong compiler, or by the right compiler with > wrong options, or is being linked by the wrong linker, or by the right > linker with bad options (or any combination of the above) > > Paolo > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From stant at nc.rr.com Wed Oct 4 18:45:44 2006 From: stant at nc.rr.com (Stan Toporek) Date: Wed, 4 Oct 2006 12:45:44 -0400 Subject: [Pw_forum] Re: Problem linking pw.x on solaris 10 Message-ID: <20061004124544.w000sk8coskso8c8@webmail.spamcop.net> All, Sorry about that last post. I changed -xarch to v9 rather then v8plusb and got the ELF error sorry just trying to access the 64 bit stuff. I still do not have pw.x running but I am working on that. Sorry, Stan -- Stan Toporek,MS Martin Group Graduate Student and NMR Servant North Carolina State University Department of Chemistry Office Phone: 515-8924 ----- "The visible world is the invisible organization of Energy" --Physicist Heinz Pagels There are only 10 types of people in the world - those who understand binary, and those who don't. From stant at nc.rr.com Wed Oct 4 18:56:47 2006 From: stant at nc.rr.com (Stan Toporek) Date: Wed, 4 Oct 2006 12:56:47 -0400 Subject: [Pw_forum] Solaris 10 make failure in cplib.o Message-ID: <20061004125647.mxsg8488cw4wso4s@webmail.spamcop.net> All, Solaris 10 make all failed with: f90 -fast -O1 -nodepend -xvector=no -xchip=ultra3 -xarch=v8plusb -xlic_lib=sunpe rf -D__SUN -I../include -M. -M../Modules -M../PW -M../PH -M../iotk/src -M../CPV -c cplib.F90 -o cplib.o CALL gracsc( bec, nkbx, betae, cp, ngwx, i, csc, n ) ^ "cplib.F90", Line = 1210, Column = 41: WARNING: Procedure "GRACSC" is defined at line 966 (cplib.F90). The actual argument type does not agree with the type of dummy argument "CP". WRITE( stdout,'(''atomic state # '',i3,'': atom # '',i3, '' species # '',i2,'' wfc # '& ^ "cplib.F90", Line = 1893, Column = 121: ERROR: Encountered end of data before co mpletion of last edit descriptor processing. &',i2, '' (l='',i1,'' m='',i2,'')'')'& ^ "cplib.F90", Line = 1894, Column = 1: ERROR: Unexpected syntax: "," or ")" was e xpected but found "'". f90comp: 2958 SOURCE LINES f90comp: 2 ERRORS, 1 WARNINGS, 0 OTHER MESSAGES, 0 ANSI *** Error code 1 make: Fatal error: Command failed for target `cplib.o' Current working directory /raid/espresso-3.1.1/CPV *** Error code 1 make: Fatal error: Command failed for target `cp' What should I do? -- Stan Toporek,MS Martin Group Graduate Student and NMR Servant North Carolina State University Department of Chemistry Office Phone: 515-8924 ----- "The visible world is the invisible organization of Energy" --Physicist Heinz Pagels There are only 10 types of people in the world - those who understand binary, and those who don't. From stant at nc.rr.com Wed Oct 4 19:20:50 2006 From: stant at nc.rr.com (Stan Toporek) Date: Wed, 4 Oct 2006 13:20:50 -0400 Subject: [Pw_forum] SunStudio 11 for free In-Reply-To: <20061003213930.jdbs3cy04bkks8k8@my1.dal.ca> References: <20061003213930.jdbs3cy04bkks8k8@my1.dal.ca> Message-ID: <20061004132050.zr4444c00w00kgks@webmail.spamcop.net> All, For Sun/Solaris users if you are a SDN member you can get the latest Sun Compilers for free. Yes, free. Even get sunperf librarys and OpenMP 2.5 and an IDE. This might help with debugging on Sun system and maybe even a 64-bit version of espresso. Check out the web site here: http://developers.sun.com/prodtech/cc/downloads/index.jsp Back to working out the problems with running PWSCF on Solaris 10 under ultrasparc II processors. Hope this helps, Stan -- Stan Toporek,MS Martin Group Graduate Student and NMR Servant North Carolina State University Department of Chemistry Office Phone: 515-8924 ----- "The visible world is the invisible organization of Energy" --Physicist Heinz Pagels There are only 10 types of people in the world - those who understand binary, and those who don't. From giannozz at nest.sns.it Wed Oct 4 19:22:43 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 4 Oct 2006 19:22:43 +0200 Subject: [Pw_forum] Solaris 10 make failure in cplib.o In-Reply-To: <20061004125647.mxsg8488cw4wso4s@webmail.spamcop.net> References: <20061004125647.mxsg8488cw4wso4s@webmail.spamcop.net> Message-ID: On Oct 4, 2006, at 18:56 , Stan Toporek wrote: > Solaris 10 make all failed [...] What should I do? 0) if you do not need something, do not compile. If you type "make" you get a list of possibilities 1) find and understand the error message 2) understand where is the error occurring (hint: it is at the 'write' at line 1893 of file CPV/cplib.F90) 3) understand what causes the error (hint: compilers may have different ideas on how to split long commands into multiple lines) 4) remove what causes of the error, if possible 5) report success (failure is not an option) P. From stant at nc.rr.com Wed Oct 4 19:31:11 2006 From: stant at nc.rr.com (Stan Toporek) Date: Wed, 4 Oct 2006 13:31:11 -0400 Subject: [Pw_forum] Solaris 10 pw.x complies but will not run Message-ID: <20061004133111.thc0cko04g8k4c0c@webmail.spamcop.net> All, Thanks for the help on the compile problem. Ok. Next problem. pw.x complies and links but when I run example01 I get: bash-3.00$ ./run_example /raid/espresso-3.1.1/examples/example01 : starting This example shows how to use pw.x to calculate the total energy and the band structure of four simple systems: Si, Al, Cu, Ni. executables directory: /raid/espresso-3.1.1/bin pseudo directory: /raid/espresso-3.1.1/pseudo temporary directory: /raid/scr/stan checking that needed directories and files exist... done running pw.x as: /raid/espresso-3.1.1/bin/pw.x cleaning /raid/scr/stan... done running the scf calculation for Si..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Si..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done cleaning /raid/scr/stan... done running the scf calculation for Al..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Al..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done cleaning /raid/scr/stan... done running the scf calculation for Cu..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Cu..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done cleaning /raid/scr/stan... done running the scf calculation for Ni..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Ni..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done cleaning /raid/scr/stan... done running the scf calculation for Si..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Si..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done cleaning /raid/scr/stan... done running the scf calculation for Al..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Al..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done cleaning /raid/scr/stan... done running the scf calculation for Cu..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Cu..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done cleaning /raid/scr/stan... done running the scf calculation for Ni..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done running the band-structure calculation for Ni..../run_example: /raid/espresso-3.1.1/bin/pw.x: cannot execute done /raid/espresso-3.1.1/examples/example01 : done Do not jump to the conclusion that the permission are wrong on pw.x a ls -la gives: bash-3.00$ ls -la pw.x -rwxr-xr-x 1 stan staff 8371508 Oct 4 13:09 pw.x I do not know how to feed it a command line calculation so if I just execute pw.x I get: bash-3.00$ ./pw.x bash: ./pw.x: Invalid argument Which is normal. What is going on? Thanks, Stan -- Stan Toporek,MS Martin Group Graduate Student and NMR Servant North Carolina State University Department of Chemistry Office Phone: 515-8924 ----- "The visible world is the invisible organization of Energy" --Physicist Heinz Pagels There are only 10 types of people in the world - those who understand binary, and those who don't. From rudrabnrj at gmail.com Thu Oct 5 16:20:40 2006 From: rudrabnrj at gmail.com (rudra banerjee) Date: Thu, 5 Oct 2006 19:50:40 +0530 Subject: [Pw_forum] band structure Message-ID: <2e36f8e50610050720hb4500eal87d94eaae6160e84@mail.gmail.com> hello. i am new to PWscf. but i am getting some problem in understanding. the thing is i am trying band structure calculation with scf giving k-points. how we allow weightage to each of them? is there any rules? secondly, as in examples, they are taking so many points. but as we know, some special points are adequate(chadi & cohen). so why we need it? hope to have some help rudra -- have a good time! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061005/62b45758/attachment.htm From eyvaz_isaev at yahoo.com Thu Oct 5 19:47:28 2006 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Thu, 5 Oct 2006 10:47:28 -0700 (PDT) Subject: [Pw_forum] band structure In-Reply-To: <2e36f8e50610050720hb4500eal87d94eaae6160e84@mail.gmail.com> Message-ID: <20061005174728.32194.qmail@web60312.mail.yahoo.com> Hi Rudra, > i am new to PWscf. Welcome! > but i am getting some problem in understanding. This one always happens if you do something. > the thing is i am trying band structure calculation > with scf giving k-points. how we allow weightage to > each of them? is there any rules? The question was discussed before in the forum and it would be nice if you search in the forum archive. Very shortly, a weight of k-points for band structure calculations is useless and you can put them 1. > secondly, as in examples, they are taking so many > points. but as we know, some special points are > adequate(chadi & cohen). so why we need it? Sure, Chadi-Cohen's special points are very useful for integration over the Brillouin zone (BZ). But I am not sure that using the CC procedure you can generate any point in the BZ. Band structure, E(k) is a set of energies for k-points along high symmetry directions and the k-points are usually (except some directions, say \Gamma-L in the FCC case) distributed over the BZ edges. But CC special points are inside the BZ. Bests, Eyvaz. > hope to have some help > > rudra > > > -- > have a good time! > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From mmiao at wsu.edu Thu Oct 5 20:00:50 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Thu, 5 Oct 2006 11:00:50 -0700 Subject: [Pw_forum] Phonon calculation for UREA Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034F29@EXCHANGEVS-04.ad.wsu.edu> Hi, Thanks for helping me compile the program. Now I get into deeper trouble. :-) I am doing phonon calculation for urea. The geometry was optimized. But then program crash while I run ph.x. The input for phonon calculation is: phonons of Urea at Gamma &inputph tr2_ph=1.0d-14, prefix='UREA', epsil=.true., amess(1)=12.011, amess(2)=14.007, amess(3)=15.999, amess(4)=1.008, outdir='./temp/', fildyn='UREA.dynG', / 0.0 0.0 0.0 The ph.x crashed and give error messgae like: Program PHONON v.3.1.1 starts ... Today is 5Oct2006 at 10:14: 6 Parallel version (MPI) Number of processors in use: 4 R & G space division: proc/pool = 4 Ultrasoft (Vanderbilt) Pseudopotentials %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%% from phq_readin : error # 4324 reading inputph namelist %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%% stopping ... Can anybody show me what could be wrong? Thanks, Maosheng From L.Boeri at fkf.mpg.de Thu Oct 5 20:03:55 2006 From: L.Boeri at fkf.mpg.de (Lilia Boeri) Date: Thu, 5 Oct 2006 20:03:55 +0200 (CEST) Subject: [Pw_forum] Phonon calculation for UREA In-Reply-To: <13EE5E22F052EB4BBBAA85A648AEFEA8034F29@EXCHANGEVS-04.ad.wsu.edu> References: <13EE5E22F052EB4BBBAA85A648AEFEA8034F29@EXCHANGEVS-04.ad.wsu.edu> Message-ID: Hi! amass is spelt "amass", not "amess" best, Lilia On Thu, 5 Oct 2006, Miao, Maosheng wrote: > Hi, > > Thanks for helping me compile the program. Now I get into deeper > trouble. :-) I am doing phonon calculation for urea. The geometry was > optimized. But then program crash while I run ph.x. The input for phonon > calculation is: > > phonons of Urea at Gamma > &inputph > tr2_ph=1.0d-14, > prefix='UREA', > epsil=.true., > amess(1)=12.011, > amess(2)=14.007, > amess(3)=15.999, > amess(4)=1.008, > outdir='./temp/', > fildyn='UREA.dynG', > / > 0.0 0.0 0.0 > > The ph.x crashed and give error messgae like: > > > Program PHONON v.3.1.1 starts ... > Today is 5Oct2006 at 10:14: 6 > > Parallel version (MPI) > > Number of processors in use: 4 > R & G space division: proc/pool = 4 > > Ultrasoft (Vanderbilt) Pseudopotentials > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%% > from phq_readin : error # 4324 > reading inputph namelist > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%% > > stopping ... > > Can anybody show me what could be wrong? > > Thanks, > Maosheng > _______________________________________________ > 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 5 20:04:15 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 5 Oct 2006 20:04:15 +0200 Subject: [Pw_forum] Solaris 10 pw.x complies but will not run In-Reply-To: <20061004133111.thc0cko04g8k4c0c@webmail.spamcop.net> References: <20061004133111.thc0cko04g8k4c0c@webmail.spamcop.net> Message-ID: <3D070B79-6682-4BE7-A1F5-0CB68A9CF71F@nest.sns.it> On Oct 4, 2006, at 19:31 , Stan Toporek wrote: > [...] if I just execute pw.x I get: > bash-3.00$ ./pw.x > bash: ./pw.x: Invalid argument > > Which is normal. no, it is not normal ! > What is going on? your executable is either bad or corrupted or good for a different architecture, or for a different OS, or for a different planet . Try to compile and execute a trivial program first. Paolo From mmiao at wsu.edu Thu Oct 5 20:08:31 2006 From: mmiao at wsu.edu (Miao, Maosheng) Date: Thu, 5 Oct 2006 11:08:31 -0700 Subject: [Pw_forum] Phonon calculation for UREA Message-ID: <13EE5E22F052EB4BBBAA85A648AEFEA8034F2C@EXCHANGEVS-04.ad.wsu.edu> Hi Lilia, Who said it's not a mess? Thank you. It works now. Maosheng -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Lilia Boeri Sent: Thursday, October 05, 2006 11:04 AM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] Phonon calculation for UREA Hi! amass is spelt "amass", not "amess" best, Lilia On Thu, 5 Oct 2006, Miao, Maosheng wrote: > Hi, > > Thanks for helping me compile the program. Now I get into deeper > trouble. :-) I am doing phonon calculation for urea. The geometry was > optimized. But then program crash while I run ph.x. The input for > phonon calculation is: > > phonons of Urea at Gamma > &inputph > tr2_ph=1.0d-14, > prefix='UREA', > epsil=.true., > amess(1)=12.011, > amess(2)=14.007, > amess(3)=15.999, > amess(4)=1.008, > outdir='./temp/', > fildyn='UREA.dynG', > / > 0.0 0.0 0.0 > > The ph.x crashed and give error messgae like: > > > Program PHONON v.3.1.1 starts ... > Today is 5Oct2006 at 10:14: 6 > > Parallel version (MPI) > > Number of processors in use: 4 > R & G space division: proc/pool = 4 > > Ultrasoft (Vanderbilt) Pseudopotentials > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %% > %%%%%% > from phq_readin : error # 4324 > reading inputph namelist > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %% > %%%%%% > > stopping ... > > Can anybody show me what could be wrong? > > Thanks, > Maosheng > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From tsli at ucdavis.edu Fri Oct 6 01:40:25 2006 From: tsli at ucdavis.edu (Tianshu Li) Date: Thu, 05 Oct 2006 16:40:25 -0700 Subject: [Pw_forum] Compatibility among different versions Message-ID: <452597E9.7020406@ucdavis.edu> Dear QE users and developers, I am curious if there might be a compatibility problem in file I/O among different versions, because when I used the post-processing code, for instance, pp.x in the latest version (v.3.1.1) to read the .save file generated in the previous version (v.3.0), it complained it could not find the .save file. The message is like: from pp_check_file : error # 1 file **********************/**.save not found. I noticed that in the latest version the new function "pp_check_file" was added to the xml_io_base.f90. This might be what gives the problem, but I am not very sure of that. Thanks Tianshu Li From giannozz at nest.sns.it Sat Oct 7 01:16:29 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sat, 7 Oct 2006 01:16:29 +0200 Subject: [Pw_forum] Compatibility among different versions In-Reply-To: <452597E9.7020406@ucdavis.edu> References: <452597E9.7020406@ucdavis.edu> Message-ID: On Oct 6, 2006, at 1:40 , Tianshu Li wrote: > I am curious if there might be a compatibility problem in file I/O > among > different versions, because when I used the post-processing code, for > instance, pp.x in the latest version (v.3.1.1) to read the .save file > generated in the previous version (v.3.0), it complained it could not > find the .save file. this is definitely a compatibility problem! > The message is like: > > from pp_check_file : error # 1 > file **********************/**.save not found. > > I noticed that in the latest version the new function "pp_check_file" > was added to the xml_io_base.f90. This might be what gives the > problem, > but I am not very sure of that. Thanks in principle with the new format there shouldn't be any longer any problem of incompatibility, because the file format is extensible, i.e. if something more is needed, one just adds a new field, without breaking the rest. Or at least, so goes the theory. In practise: - one may realise that the way things are done is not smart and that it needs to be changed, so that a field name or the content of a field may no longer be what it used to be. This happens especially at the beginning, when the file format has to stabilize. - if something is added to the file format, the code reading it should be aware that something may be missing from files produced by earlier versions, For instance, this "pp_check_file" stuff was added to read a tag that is written at convergence. In the first version, the code simply stopped if there was no such tags, thus making all files produced by previous versions unreadable. Now this should be corrected. In the future we hope that there will be no incompatible changes. For the past ... the past is gone. The specific error you have looks very strange and I am not sure it is related to the file format. Does it work if you try to read files produced with the same version of the code? Paolo From tsli at ucdavis.edu Sat Oct 7 08:42:16 2006 From: tsli at ucdavis.edu (Tianshu Li) Date: Fri, 06 Oct 2006 23:42:16 -0700 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1173 - 1 msg In-Reply-To: <20061007053751.27571.26427.Mailman@democritos.sissa.it> References: <20061007053751.27571.26427.Mailman@democritos.sissa.it> Message-ID: <45274C48.8070205@ucdavis.edu> Hi Paolo, Thanks for your reply. In fact the code works perfectly if it reads the output generated from the same version. It only gives problem when there is a file I/O across the different versions. So this is just what you've described. Tianshu pw_forum-request at pwscf.org wrote: >From: Paolo Giannozzi >Subject: Re: [Pw_forum] Compatibility among different versions >Date: Sat, 7 Oct 2006 01:16:29 +0200 >To: pw_forum at pwscf.org >Reply-To: pw_forum at pwscf.org > >On Oct 6, 2006, at 1:40 , Tianshu Li wrote: > > > >>I am curious if there might be a compatibility problem in file I/O >>among >>different versions, because when I used the post-processing code, for >>instance, pp.x in the latest version (v.3.1.1) to read the .save file >>generated in the previous version (v.3.0), it complained it could not >>find the .save file. >> >> > >this is definitely a compatibility problem! > > > >>The message is like: >> >>from pp_check_file : error # 1 >>file **********************/**.save not found. >> >>I noticed that in the latest version the new function "pp_check_file" >>was added to the xml_io_base.f90. This might be what gives the >>problem, >>but I am not very sure of that. Thanks >> >> > >in principle with the new format there shouldn't be any longer any >problem of incompatibility, because the file format is extensible, i.e. >if something more is needed, one just adds a new field, without >breaking the rest. Or at least, so goes the theory. In practise: > >- one may realise that the way things are done is not smart and that >it needs to be changed, so that a field name or the content of a field >may no longer be what it used to be. This happens especially at the >beginning, when the file format has to stabilize. > >- if something is added to the file format, the code reading it should >be aware that something may be missing from files produced by earlier >versions, For instance, this "pp_check_file" stuff was added to read >a tag that is written at convergence. In the first version, the code >simply >stopped if there was no such tags, thus making all files produced by >previous versions unreadable. Now this should be corrected. > >In the future we hope that there will be no incompatible changes. >For the past ... the past is gone. The specific error you have looks >very strange and I am not sure it is related to the file format. Does >it work if you try to read files produced with the same version of >the code? > >Paolo > > > > >--__--__-- > >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > >End of Pw_forum Digest > > > -- Tianshu Li Chemistry Department University of California, Davis Tel: 1-530-754-9664 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061006/edeb075f/attachment.htm From babazade at gmail.com Sun Oct 8 12:40:27 2006 From: babazade at gmail.com (Amin Babazadeh) Date: Sun, 8 Oct 2006 14:10:27 +0330 Subject: [Pw_forum] about density Message-ID: dear users In espresso code we can calculate the DOS for s,p,or d orbitals but as i know in pwscf method we expand a plane wave function and use it to find eigenvalues and eigenstete and in the expanded plane wave we don't have a quantum numbers (i mean n & l ) or orbitals.So how we can say that this is a DOS diagram of s orbital or p orbital. Sincerely yours -- AMIN -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061008/cd391f9b/attachment.htm From baroni at sissa.it Sun Oct 8 14:30:12 2006 From: baroni at sissa.it (Stefano Baroni) Date: Sun, 8 Oct 2006 14:30:12 +0200 Subject: [Pw_forum] about density In-Reply-To: References: Message-ID: <0A4155DE-17CE-41C8-BA73-41BB238C528A@sissa.it> Hi Amin. It is a matter of definition that may or may be not useful, according to the use that one makes of it. The projected density of states reads: n(\epsilon) = \sum_n | \langle \psi_n | \phi_0 \rangle |^2 \delta (\epsilon-\epsilon_n) where \phi_0 is the orbitals you want to project onto, and \phi_n and \espilon_n are eigenpairs of the Kohn-Sham equation. The above expression, that can be seen as the imaginary part of the one- electron Green's function, is perfectly well defined for any localized orbital irrespective of the basis set used to expand the Kohn-Sham orbitals. In a nutshell, n(\epsilon) expresses the contribution of the localized orbital to the total density of states of the system under study. does this help? regards - SB On Oct 8, 2006, at 12:40 PM, Amin Babazadeh wrote: > dear users > In espresso code we can calculate the DOS for s,p,or d orbitals > but as i know in pwscf method we expand a plane wave function and > use it to find eigenvalues and eigenstete and in the expanded > plane wave we don't have a quantum numbers (i mean n & l ) or > orbitals.So how we can say that this is a DOS diagram of s orbital > or p orbital. > Sincerely yours > > -- > AMIN --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061008/489cbf63/attachment.htm From konstantin_kudin at yahoo.com Mon Oct 9 04:46:10 2006 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Sun, 8 Oct 2006 19:46:10 -0700 (PDT) Subject: [Pw_forum] Raman spectra for (almost) metallic systems In-Reply-To: <0A4155DE-17CE-41C8-BA73-41BB238C528A@sissa.it> Message-ID: <20061009024610.32388.qmail@web57703.mail.re3.yahoo.com> Hi all, My question is if there is any hope to a get a meaningful non-resonant Raman spectra from the PH program for something almost metallic, like graphite. Basically, how are these Raman tensors (d^3 E/ dx de1 de2) are computed, will they totally break down for something with a small gap? Or is there a bit of hope there? Thanks! Kostya __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From giannozz at nest.sns.it Mon Oct 9 04:56:25 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 9 Oct 2006 04:56:25 +0200 Subject: [Pw_forum] Pseudopotential troubles In-Reply-To: <4523D7B5.9050609@lg.ehu.es> References: <4523D7B5.9050609@lg.ehu.es> Message-ID: On Oct 4, 2006, at 17:48 , Miguel Mart?nez Canales wrote: > Anyway, I have another more profound question. I don't understand > why this pseudopotential input is terribly flawed (its result is). > > &input > title='Ca', > zed=20.0, > xmin=-6.0, > dx=0.0200, > rmax=80, > rel=1, > beta=0.3, > rlderiv=2.7, > eminld=-4.0, > emaxld=4.0, > deld=0.02d0, > nld=4, > config='[Ar] 3d0 4s2 4p0', > iswitch=3, > dft='PBE', > / you do not need to specify the radial grid and other optional variables > &inputp > pseudotype=3, that is, ultrasoft > lloc=-1, > rcloc=2.1, rcloc should be larger than any rc > tm =.false., > nlcc =.true., > rho0 =0.01, > file_pseudopw='Ca.pbe-van.UPF', > file_screen='ld1ps.screen' > / > 3 > 3D 3 2 0.00 0.00 2.40 2.40 > 4S 1 0 2.00 0.00 2.68 2.68 > 4P 2 1 0.00 0.00 2.59 2.59 you asked for a US PP but rc for norm-conserving and ultrasoft pseudization are the same. These values for rc seem to me too large for 3d, too small for 4s and 4p. If you plan to use Ca for high-pressure studies, you should consider using 3s and 3p semicore states in valence. Example: &input atom='Ca', rel=1, config='[Ar] 3d0 4s0 4p0', iswitch=3, dft='PBE' / &inputp pseudotype=1, lloc=2 nlcc =.true., file_pseudopw='Ca.pbe-nsp-1.UPF', rho0=0.01 / 3 3S 1 0 2.00 0.00 1.20 1.20 3P 2 1 6.00 0.00 1.30 1.30 3D 3 2 0.00 0.00 1.40 1.40 (notice that some versions of the atomic code yield a wrong Zval in a case like this: it should be checked and changed in the *UPF file) > Another question I have is how to add additional projectors per > orbital, > such as the ones used in Paolo Gianozzi's pbe pseudopotential. ??? Paolo From jasonsun98 at hotmail.com Mon Oct 9 08:32:52 2006 From: jasonsun98 at hotmail.com (sun jian) Date: Mon, 09 Oct 2006 06:32:52 +0000 Subject: [Pw_forum] about disordered structures Message-ID: Dear all, whether pwscf can be used to calculate disordered crystals within virtural crystal approximation ? If it can, how to specify the lattice, i.e. atom positions and compositions ? Thank you very much in advance ! Best regards, =============================================== Jian SUN Physics Dept. of Nanjing University National Lab. of Solid State Microstructures 22 Hankou Road, Gulou District Nanjing, Jiangsu Province 210093 China =============================================== From degironc at sissa.it Mon Oct 9 09:49:13 2006 From: degironc at sissa.it (Stefano de Gironcoli) Date: Mon, 9 Oct 2006 09:49:13 +0200 (CEST) Subject: [Pw_forum] about disordered structures In-Reply-To: References: Message-ID: The whole idea of virtual crystal approximation is that one replaces all atom of a given disordered sublattice with a pseudopotential which is the proper average of the two species that are distributed in that sublattice. In the upftools directory of the QE distribution there is a simple code called virtual.x that reads two pseudopotentials in the upf format, a mixing fraction, x, and produces a third pseudopotential as (1-x)V1+xV2 . Then you can use this virtual-atom pseudo as usual in the other QE codes. This is a feature that as very seldom been used... so test that your results make sense before trusting them (even more than usual). Please report any doubt/problem/bug. Thank you, Stefano de Gironcoli On Mon, 9 Oct 2006, sun jian wrote: > Dear all, > > whether pwscf can be used to calculate disordered crystals within virtural > crystal approximation ? If it can, how to specify the lattice, i.e. atom > positions and compositions ? > > Thank you very much in advance ! > > Best regards, > =============================================== > 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 eminekb at yahoo.com Mon Oct 9 10:46:40 2006 From: eminekb at yahoo.com (e kb) Date: Mon, 9 Oct 2006 01:46:40 -0700 (PDT) Subject: [Pw_forum] Pseudopotential with semicore states Message-ID: <20061009084640.99696.qmail@web34809.mail.mud.yahoo.com> Dear all, I am trying to generate a pseudopotential with semicore states, INPUT_LD1 file was very useful (thanks) but there is this sentence that i still couldnt understand what is exactly meant: (at Section 3.1 ) "...use semicore states ONLY to produce the PP, use semicore AND valence states (if occupied) to make the unscreening." It might be a very easy question, but i really couldnt get. Why should i use ONLY semicore states to produce PP, why not the -occupied- valance states at that step? Thanks, Emine Kucukbenli --------------------------------- Get your email and more, right on the new Yahoo.com -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061009/cc00e489/attachment.htm From degironc at sissa.it Mon Oct 9 11:11:12 2006 From: degironc at sissa.it (Stefano de Gironcoli) Date: Mon, 9 Oct 2006 11:11:12 +0200 (CEST) Subject: [Pw_forum] Pseudopotential with semicore states In-Reply-To: <20061009084640.99696.qmail@web34809.mail.mud.yahoo.com> References: <20061009084640.99696.qmail@web34809.mail.mud.yahoo.com> Message-ID: Paolo or some other may correct me but this is how I think it is. In order to generate the screened pseudopotential the KS equation must be inverted (for ONE pseudowavefunction per channel) and this is done using the semicore states that are nodeless. In order to unscreen the potential one need the pseudocharge that is computed from all occupied states (both semicore and valence states). Stefano de Gironcoli On Mon, 9 Oct 2006, e kb wrote: > Dear all, > I am trying to generate a pseudopotential with semicore states, > INPUT_LD1 file was very useful (thanks) > but there is this sentence that i still couldnt understand what is exactly meant: > (at Section 3.1 ) > "...use semicore states ONLY to produce the PP, use semicore > AND valence states (if occupied) to make the unscreening." > > It might be a very easy question, but i really couldnt get. > Why should i use ONLY semicore states to produce PP, > why not the -occupied- valance states at that step? > > Thanks, > Emine Kucukbenli > > > --------------------------------- > Get your email and more, right on the new Yahoo.com From wmbmacam at lg.ehu.es Mon Oct 9 11:27:30 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Mon, 09 Oct 2006 11:27:30 +0200 Subject: [Pw_forum] Pseudopotential troubles In-Reply-To: References: <4523D7B5.9050609@lg.ehu.es> Message-ID: <452A1602.3080805@lg.ehu.es> Dear Paolo, First, thanks for your answer. > you do not need to specify the radial grid and other optional variables Indeed. It's just that I was puzzled to see that different radial grids allowed me different rc's for RRKJ pseudization. Due to the 3d orbital I am indeed more interested in an ultrasoft pseudopotential than in a Trouliers-Martins one. > rcloc should be larger than any rc Thanks for this one. > you asked for a US PP but rc for norm-conserving and ultrasoft > pseudization are the same. Yes, I know. When I increase rcutus, ld1.x complains for having only one reference energy. I would add additional energies per channel but I don't see an obvious way to do this after reading INPIT_LD1. ?Should I add an empty 3s orbital with the pseudization energy .neq. 0.00? >> Another question I have is how to add additional projectors per orbital, >> such as the ones used in Paolo Gianozzi's pbe pseudopotential. > > ??? OK, it seems I didn't explain myself. I should have written projector per l. If you grep Projectors out of a pseudo file you get, for example: $ grep Projector ~/pseudo/Ca.pbe-nsp-van.UPF 4 6 Number of Wavefunctions, Number of Projectors This is straightforward to do in David Vanderbilt's code, as you directly specify the number of beta functions the pseudopotential is going to have. Again, to do this, I would add another let's say 3s line to &inputp. But I fear this won't give the desired effect. Thanks again for your time and attention, Miguel From eminekb at yahoo.com Mon Oct 9 15:27:25 2006 From: eminekb at yahoo.com (e kb) Date: Mon, 9 Oct 2006 06:27:25 -0700 (PDT) Subject: [Pw_forum] Pseudopotential with semicore states In-Reply-To: Message-ID: <20061009132725.31725.qmail@web34810.mail.mud.yahoo.com> Dear de Gironcoli and espresso users, Thank you for replying Although I still have questionmarks in my mind about the explanation, I will work more on this. Thank you again, Emine Kucukbenli Stefano de Gironcoli wrote: Paolo or some other may correct me but this is how I think it is. In order to generate the screened pseudopotential the KS equation must be inverted (for ONE pseudowavefunction per channel) and this is done using the semicore states that are nodeless. In order to unscreen the potential one need the pseudocharge that is computed from all occupied states (both semicore and valence states). Stefano de Gironcoli --------------------------------- Stay in the know. Pulse on the new Yahoo.com. Check it out. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061009/b44aa406/attachment.htm From babazade at gmail.com Mon Oct 9 15:45:42 2006 From: babazade at gmail.com (Amin Babazadeh) Date: Mon, 9 Oct 2006 17:15:42 +0330 Subject: [Pw_forum] about density In-Reply-To: <0A4155DE-17CE-41C8-BA73-41BB238C528A@sissa.it> References: <0A4155DE-17CE-41C8-BA73-41BB238C528A@sissa.it> Message-ID: Dear Stefano Thanks a lot for your help.It helps me, Regards amin On 10/8/06, Stefano Baroni wrote: > > Hi Amin. It is a matter of definition that may or may be not useful, > according to the use that one makes of it. The projected density of states > reads: > > n(\epsilon) = \sum_n | \langle \psi_n | \phi_0 \rangle |^2 > \delta(\epsilon-\epsilon_n) > > where \phi_0 is the orbitals you want to project onto, and \phi_n and > \espilon_n are eigenpairs of the Kohn-Sham equation. The above expression, > that can be seen as the imaginary part of the one-electron Green's function, > is perfectly well defined for any localized orbital irrespective of the > basis set used to expand the Kohn-Sham orbitals. In a nutshell, n(\epsilon) > expresses the contribution of the localized orbital to the total density of > states of the system under study. > > does this help? > > regards - SB > > > > On Oct 8, 2006, at 12:40 PM, Amin Babazadeh wrote: > > dear users > In espresso code we can calculate the DOS for s,p,or d orbitals but as i > know in pwscf method we expand a plane wave function and use it to find > eigenvalues and eigenstete and in the expanded plane wave we don't have a > quantum numbers (i mean n & l ) or orbitals.So how we can say that this is > a DOS diagram of s orbital or p orbital. > Sincerely yours > > -- > AMIN > > > --- > Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - > Trieste > [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) > > Please, if possible, don't send me MS Word or PowerPoint attachments > Why? See: http://www.gnu.org/philosophy/no-word-attachments.html > > > > -- AMIN -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061009/465fae2e/attachment.htm From liyanpcl at yahoo.com.cn Mon Oct 9 16:15:39 2006 From: liyanpcl at yahoo.com.cn (li yan) Date: Mon, 9 Oct 2006 22:15:39 +0800 (CST) Subject: [Pw_forum] Problem of the underestimation of the frequency caused by matdyn.x Message-ID: <20061009141540.86000.qmail@web15611.mail.cnb.yahoo.com> Dear all, I calculated the phonon frequencies of FeO for a list of generic q vectors with q2r.x and matdyn.x. In the FeO.dyn files , the highest frequency is about 43 THz, while it is about 22 THz in the matdyn.modes file and FeO.freq file. What problems may occur in the the process? --------------------------------- ??????-3.5G???20M?? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061009/e9db5aa6/attachment.htm From giannozz at nest.sns.it Mon Oct 9 16:30:33 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 9 Oct 2006 16:30:33 +0200 Subject: [Pw_forum] Problem of the underestimation of the frequency caused by matdyn.x In-Reply-To: <20061009141540.86000.qmail@web15611.mail.cnb.yahoo.com> References: <20061009141540.86000.qmail@web15611.mail.cnb.yahoo.com> Message-ID: On Oct 9, 2006, at 16:15 , li yan wrote: > I calculated the phonon frequencies of FeO for a list of > generic > q vectors with q2r.x and matdyn.x. In the FeO.dyn files , the > highest frequency is about 43 THz, while it is about 22 THz in the > matdyn.modes file and FeO.freq file. What problems may occur in > the the process? wrong masses? Paolo From vranjan at ncsu.edu Mon Oct 9 16:40:41 2006 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Mon, 9 Oct 2006 10:40:41 -0400 (EDT) Subject: [Pw_forum] born effective charges Message-ID: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> Hi ! I am trying to calculate born effective charges for a system consisting of more than hundred atoms. I am doing it using linear response, i.e. I relax the structure, then I do a scf calculation and then I run a gamma point phonon calculation. After the programme prints the dielectric constants, it must calculate the born effective charges. With the given computer resources at my disposal, it cannot complete this part. So, I must restart the calculation from the previous step. I use "recover.true." flag. But it never seems to go beyond the dielectric constant calculation part. Is there a way I can do this calculation ? Thanks, Vivek PS : I know that I can calculate polarization using berry phase calculation. Then I can move atoms one-by-one and calculate effective charges by finite difference. But that will be doing many calculations, so I want to avoid this method. From wierzbom at ts.infn.it Mon Oct 9 16:44:25 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Mon, 9 Oct 2006 16:44:25 +0200 (CEST) Subject: [Pw_forum] Problem of the underestimation of the frequency caused by matdyn.x In-Reply-To: References: <20061009141540.86000.qmail@web15611.mail.cnb.yahoo.com> Message-ID: On Mon, 9 Oct 2006, Paolo Giannozzi wrote: > On Oct 9, 2006, at 16:15 , li yan wrote: > >> I calculated the phonon frequencies of FeO for a list of generic >> q vectors with q2r.x and matdyn.x. In the FeO.dyn files , the highest >> frequency is about 43 THz, while it is about 22 THz in the matdyn.modes >> file and FeO.freq file. What problems may occur in the the process? > > wrong masses? Paolo > What q-points and modes of FeO? Are you sure about units, THz? Gosia > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From baroni at sissa.it Mon Oct 9 16:50:32 2006 From: baroni at sissa.it (Stefano Baroni) Date: Mon, 9 Oct 2006 16:50:32 +0200 Subject: [Pw_forum] born effective charges In-Reply-To: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> References: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> Message-ID: In your case it would be much more effective to calculate the effective charges as the forces linearly induced by an applied homogeneous electric field (just one electric-field calculation), rather than as the polarization linearly induced by a nuclear displacement. The two definitions do of course coincide (they both amount to calculating the mixed second derivative of the crystal energy with respect to the electric field and a nuclear displacement), but the calculation of the force response to an electric field only requires just 1 calculation, rather that 100 (in your case)! S. On Oct 9, 2006, at 4:40 PM, Vivek Ranjan wrote: > Hi ! > > I am trying to calculate born effective charges for a system > consisting of > more than hundred atoms. I am doing it using linear response, i.e. > I relax > the structure, then I do a scf calculation and then I run a gamma > point > phonon calculation. After the programme prints the dielectric > constants, > it must calculate the born effective charges. With the given computer > resources at my disposal, it cannot complete this part. So, I must > restart > the calculation from the previous step. I use "recover.true." flag. > But it > never seems to go beyond the dielectric constant calculation part. Is > there a way I can do this calculation ? > > Thanks, > Vivek > > PS : I know that I can calculate polarization using berry phase > calculation. Then I can move atoms one-by-one and calculate effective > charges by finite difference. But that will be doing many > calculations, so > I want to avoid this method. > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061009/8a6f23b8/attachment.htm From liyanpcl at yahoo.com.cn Tue Oct 10 04:33:25 2006 From: liyanpcl at yahoo.com.cn (li yan) Date: Tue, 10 Oct 2006 10:33:25 +0800 (CST) Subject: [Pw_forum] =?gb2312?q?Re=A3=BA=20Re:=20[Pw=5Fforum]=20Problem=20of=20the=20underesti?= =?gb2312?q?mation=20of=20the=20frequency=20caused=20by=20matdyn.x?= In-Reply-To: Message-ID: <20061010023325.77758.qmail@web15614.mail.cnb.yahoo.com> i'm sorry to have make a mistake about my problem. The system is CaC6 instead of FeO . i select one of the dyn files, the the frequencis at q point are: q = ( 0.000000000 0.000000000 0.286236050 ) omega( 1) = 1.432232 [THz] = 47.774442 [cm-1] omega( 2) = 1.432232 [THz] = 47.774442 [cm-1] omega( 3) = 3.262337 [THz] = 108.820586 [cm-1] omega( 4) = 3.262337 [THz] = 108.820586 [cm-1] omega( 5) = 3.587688 [THz] = 119.673169 [cm-1] omega( 6) = 8.461922 [THz] = 282.261207 [cm-1] omega( 7) = 12.447993 [THz] = 415.223105 [cm-1] omega( 8) = 12.447993 [THz] = 415.223105 [cm-1] omega( 9) = 13.116535 [THz] = 437.523408 [cm-1] omega(10) = 13.116535 [THz] = 437.523408 [cm-1] omega(11) = 19.159806 [THz] = 639.106575 [cm-1] omega(12) = 28.691164 [THz] = 957.040582 [cm-1] omega(13) = 29.038383 [THz] = 968.622620 [cm-1] omega(14) = 33.223897 [THz] = 1108.237267 [cm-1] omega(15) = 33.223897 [THz] = 1108.237267 [cm-1] omega(16) = 33.281138 [THz] = 1110.146637 [cm-1] omega(17) = 33.281138 [THz] = 1110.146637 [cm-1] omega(18) = 41.604039 [THz] = 1387.770602 [cm-1] omega(19) = 42.074591 [THz] = 1403.466608 [cm-1] omega(20) = 43.376807 [THz] = 1446.904149 [cm-1] omega(21) = 43.376807 [THz] = 1446.904149 [cm-1] while in the matdyn.modes file, the frequencies of the modes at the same point becom q = 0.0000 0.0000 0.2862 omega( 1) = 0.867379 [THz] = 28.932849 [cm-1] omega( 2) = 0.867380 [THz] = 28.932870 [cm-1] omega( 3) = 2.308488 [THz] = 77.003380 [cm-1] omega( 4) = 4.689393 [THz] = 156.422366 [cm-1] omega( 5) = 4.689394 [THz] = 156.422387 [cm-1] omega( 6) = 6.821788 [THz] = 227.551869 [cm-1] omega( 7) = 6.821788 [THz] = 227.551870 [cm-1] omega( 8) = 8.276312 [THz] = 276.069874 [cm-1] omega( 9) = 8.276312 [THz] = 276.069882 [cm-1] omega(10) = 10.491723 [THz] = 349.968534 [cm-1] omega(11) = 13.134275 [THz] = 438.115167 [cm-1] omega(12) = 15.696771 [THz] = 523.591396 [cm-1] omega(13) = 15.903656 [THz] = 530.492373 [cm-1] omega(14) = 18.187294 [THz] = 606.666874 [cm-1] omega(15) = 18.187294 [THz] = 606.666875 [cm-1] omega(16) = 18.222317 [THz] = 607.835092 [cm-1] omega(17) = 18.222317 [THz] = 607.835098 [cm-1] omega(18) = 22.768578 [THz] = 759.483060 [cm-1] omega(19) = 23.037730 [THz] = 768.461058 [cm-1] omega(20) = 23.744563 [THz] = 792.038631 [cm-1] omega(21) = 23.744563 [THz] = 792.038632 [cm-1] that is to say that all the frequencies are underestimated after using q2r.x and matdyn.x. what's the prpblem with it? Thank you very much. ka ??? On Mon, 9 Oct 2006, Paolo Giannozzi wrote: > On Oct 9, 2006, at 16:15 , li yan wrote: > >> I calculated the phonon frequencies of FeO for a list of generic >> q vectors with q2r.x and matdyn.x. In the FeO.dyn files , the highest >> frequency is about 43 THz, while it is about 22 THz in the matdyn.modes >> file and FeO.freq file. What problems may occur in the the process? > > wrong masses? Paolo > What q-points and modes of FeO? Are you sure about units, THz? Gosia > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum --------------------------------- Mp3???-??????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061010/1a2c6245/attachment.htm From yanmingma at sohu.com Tue Oct 10 09:43:46 2006 From: yanmingma at sohu.com (yanmingma) Date: Tue, 10 Oct 2006 09:43:46 +0200 Subject: [Pw_forum] =?gb2312?B?UmU6IFtQd19mb3J1bV0gUmWjuiBSZTogW1B3X2ZvcnVtXSBQcm9ibGVtIG9mIHRoZSB1bmRlcmVzdGltYXRpb24gb2YgdGhlIGZyZXF1ZW5jeSBjYXVzZWQgYnkgbWF0ZHluLng=?= References: <20061010023325.77758.qmail@web15614.mail.cnb.yahoo.com> Message-ID: <200610100943401563012@sohu.com> Dear Liyan, Please check the corresponding masses for Ca and C as suggested by Paolo. Regards yanmingma 2006-10-10 ???? li yan ????? 2006-10-10 04:34:12 ???? pw_forum at pwscf.org ??? ??? [Pw_forum] Re? Re: [Pw_forum] Problem of the underestimation of the frequency caused by matdyn.x i'm sorry to have make a mistake about my problem. The system is CaC6 instead of FeO . i select one of the dyn files, the the frequencis at q point are: q = ( 0.000000000 0.000000000 0.286236050 ) omega( 1) = 1.432232 [THz] = 47.774442 [cm-1] omega( 2) = 1.432232 [THz] = 47.774442 [cm-1] omega( 3) = 3.262337 [THz] = 108.820586 [cm-1] omega( 4) = 3.262337 [THz] = 108.820586 [cm-1] omega( 5) = 3.587688 [THz] = 119.673169 [cm-1] omega( 6) = 8.461922 [THz] = 282.261207 [cm-1] omega( 7) = 12.447993 [THz] = 415.223105 [cm-1] omega( 8) = 12.447993 [THz] = 415.223105 [cm-1] omega( 9) = 13.116535 [THz] = 437.523408 [cm-1] omega(10) = 13.116535 [THz] = 437.523408 [cm-1] omega(11) = 19.159806 [THz] = 639.106575 [cm-1] omega(12) = 28.691164 [THz] = 957.040582 [cm-1] omega(13) = 29.038383 [THz] = 968.622620 [cm-1] omega(14) = 33.223897 [THz] = 1108.237267 [cm-1] omega(15) = 33.223897 [THz] = 1108.237267 [cm-1] omega(16) = 33.281138 [THz] = 1110.146637 [cm-1] omega(17) = 33.281138 [THz] = 1110.146637 [cm-1] omega(18) = 41.604039 [THz] = 1387.770602 [cm-1] omega(19) = 42.074591 [THz] = 1403.466608 [cm-1] omega(20) = 43.376807 [THz] = 1446.904149 [cm-1] omega(21) = 43.376807 [THz] = 1446.904149 [cm-1] while in the matdyn.modes file, the frequencies of the modes at the same point becom q = 0.0000 0.0000 0.2862 omega( 1) = 0.867379 [THz] = 28.932849 [cm-1] omega( 2) = 0.867380 [THz] = 28.932870 [cm-1] omega( 3) = 2.308488 [THz] = 77.003380 [cm-1] omega( 4) = 4.689393 [THz] = 156.422366 [cm-1] omega( 5) = 4.689394 [THz] = 156.422387 [cm-1] omega( 6) = 6.821788 [THz] = 227.551869 [cm-1] omega( 7) = 6.821788 [THz] = 227.551870 [cm-1] omega( 8) = 8.276312 [THz] = 276.069874 [cm-1] omega( 9) = 8.276312 [THz] = 276.069882 [cm-1] omega(10) = 10.491723 [THz] = 349.968534 [cm-1] omega(11) = 13.134275 [THz] = 438.115167 [cm-1] omega(12) = 15.696771 [THz] = 523.591396 [cm-1] omega(13) = 15.903656 [THz] = 530.492373 [cm-1] omega(14) = 18.187294 [THz] = 606.666874 [cm-1] omega(15) = 18.187294 [THz] = 606.666875 [cm-1] omega(16) = 18.222317 [THz] = 607.835092 [cm-1] omega(17) = 18.222317 [THz] = 607.835098 [cm-1] omega(18) = 22.768578 [THz] = 759.483060 [cm-1] omega(19) = 23.037730 [THz] = 768.461058 [cm-1] omega(20) = 23.744563 [THz] = 792.038631 [cm-1] omega(21) = 23.744563 [THz] = 792.038632 [cm-1] that is to say that all the frequencies are underestimated after using q2r.x and matdyn.x. what's the prpblem with it? Thank you very much. ka ??? On Mon, 9 Oct 2006, Paolo Giannozzi wrote: > On Oct 9, 2006, at 16:15 , li yan wrote: > >> I calculated the phonon frequencies of FeO for a list of generic >> q vectors with q2r.x and matdyn.x. In the FeO.dyn files , the highest >> frequency is about 43 THz, while it is about 22 THz in the matdyn.modes >> file and FeO.freq file. What problems may occur in the the process? > > wrong masses? Paolo > What q-points and modes of FeO? Are you sure about units, THz? Gosia > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum Mp3???-??????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061010/cb370057/attachment.htm From sodovar at gmail.com Tue Oct 10 10:47:39 2006 From: sodovar at gmail.com (Ruslan Minibaev) Date: Tue, 10 Oct 2006 12:47:39 +0400 Subject: [Pw_forum] about projwfc.x Message-ID: Dear all! I perform calculation of oxide surface. I've made following steps: 1) relax mode - OK 2) scf mode - OK 3) nscf mode - OK4) total dos - OK On the next step I should separate surface states from the total DOS of a slab. If I understood correctly it can be done with use of projwfc.x Input file for projwfc.x: &inputpp outdir='./' filpdos='PDOS.dos', Emin=-30.0, Emax=25.0, DeltaE=0.05 / This mode failed: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% task # 0 from read_restart_wfc : error # 1 Data Section not present in restart file %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% What's the reason of the error? Best regards, Ruslan Minibaev. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061010/b173af04/attachment.htm From babazade at gmail.com Tue Oct 10 11:28:04 2006 From: babazade at gmail.com (Amin Babazadeh) Date: Tue, 10 Oct 2006 12:58:04 +0330 Subject: [Pw_forum] about projwfc.x In-Reply-To: References: Message-ID: Dear Ruslan 1.Check your outdir and be sure that it is correct. 2.for making pdos file you can use the file that is in the example08 (the file name is ni.pdos.in) in that file you can see Gaussian delta and some another commend. I thought it will help you. Regards. amin Are On 10/10/06, Ruslan Minibaev wrote: > > Dear all! > > I perform calculation of oxide surface. > I've made following steps: > 1) relax mode - OK > 2) scf mode - OK > 3) nscf mode - OK4) > total dos - OK > On the next step I should separate surface states from the total DOS of > a slab. > If I understood correctly it can be done with use of projwfc.x > > Input file for projwfc.x: > &inputpp > outdir='./' > filpdos='PDOS.dos', > Emin=-30.0, Emax=25.0, DeltaE=0.05 > / > This mode failed: > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > task # 0 > from read_restart_wfc : error # 1 > Data Section not present in restart file > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > What's the reason of the error? > > Best regards, > Ruslan Minibaev. > -- AMIN -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061010/02645322/attachment.htm From sodovar at gmail.com Tue Oct 10 11:41:24 2006 From: sodovar at gmail.com (Ruslan Minibaev) Date: Tue, 10 Oct 2006 13:41:24 +0400 Subject: [Pw_forum] Re: about projwfc Message-ID: Dear Amin Babazadeh! Thank you for quick answer! Best regards, Ruslan Minibaev -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061010/bc53f59a/attachment.htm From ezadshojaee at hotmail.com Tue Oct 10 11:53:18 2006 From: ezadshojaee at hotmail.com (Ezad Shojaee) Date: Tue, 10 Oct 2006 09:53:18 +0000 Subject: [Pw_forum] phonon at a specifiec point Message-ID: hi performing phonon calculation at a speciefic point, it seems that the program writes the values of phonon frequencies in the output file , at the end of all. if i want to know the value of first mode during the calculation where should i search for ? thanx _________________________________________________________________ Don't just search. Find. Check out the new MSN Search! http://search.msn.com/ From wierzbom at ts.infn.it Tue Oct 10 14:11:38 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Tue, 10 Oct 2006 14:11:38 +0200 (CEST) Subject: [Pw_forum] Re: **Suspected SPAM** [Pw_forum] =?gb2312?q?Re=A3=BA=20Re:=20[Pw=5Fforum]=20Problem=20of=20the=20underesti?= =?gb2312?q?mation=20of=20the=20frequency=20caused=20by=20matdyn.x?= In-Reply-To: <20061010023325.77758.qmail@web15614.mail.cnb.yahoo.com> References: <20061010023325.77758.qmail@web15614.mail.cnb.yahoo.com> Message-ID: On Tue, 10 Oct 2006, li yan wrote: Dear Li Yan, What was the array of q-points which you have used for the Fourier transformation q2r and matdyn? Please can you give us a list of all q-points (where one of them is the point below: 0.0, 0.0, 0.286). Please tell us the input files to q2r and matdyn. Regards, Gosia > > i'm sorry to have make a mistake about my problem. The system is CaC6 instead of FeO . i select one of the dyn files, the the frequencis at q point are: > q = ( 0.000000000 0.000000000 0.286236050 ) > omega( 1) = 1.432232 [THz] = 47.774442 [cm-1] > omega( 2) = 1.432232 [THz] = 47.774442 [cm-1] > omega( 3) = 3.262337 [THz] = 108.820586 [cm-1] > omega( 4) = 3.262337 [THz] = 108.820586 [cm-1] > omega( 5) = 3.587688 [THz] = 119.673169 [cm-1] > omega( 6) = 8.461922 [THz] = 282.261207 [cm-1] > omega( 7) = 12.447993 [THz] = 415.223105 [cm-1] > omega( 8) = 12.447993 [THz] = 415.223105 [cm-1] > omega( 9) = 13.116535 [THz] = 437.523408 [cm-1] > omega(10) = 13.116535 [THz] = 437.523408 [cm-1] > omega(11) = 19.159806 [THz] = 639.106575 [cm-1] > omega(12) = 28.691164 [THz] = 957.040582 [cm-1] > omega(13) = 29.038383 [THz] = 968.622620 [cm-1] > omega(14) = 33.223897 [THz] = 1108.237267 [cm-1] > omega(15) = 33.223897 [THz] = 1108.237267 [cm-1] > omega(16) = 33.281138 [THz] = 1110.146637 [cm-1] > omega(17) = 33.281138 [THz] = 1110.146637 [cm-1] > omega(18) = 41.604039 [THz] = 1387.770602 [cm-1] > omega(19) = 42.074591 [THz] = 1403.466608 [cm-1] > omega(20) = 43.376807 [THz] = 1446.904149 [cm-1] > omega(21) = 43.376807 [THz] = 1446.904149 [cm-1] > while in the matdyn.modes file, the frequencies of the modes at the same point becom > q = 0.0000 0.0000 0.2862 > omega( 1) = 0.867379 [THz] = 28.932849 [cm-1] > omega( 2) = 0.867380 [THz] = 28.932870 [cm-1] > omega( 3) = 2.308488 [THz] = 77.003380 [cm-1] > omega( 4) = 4.689393 [THz] = 156.422366 [cm-1] > omega( 5) = 4.689394 [THz] = 156.422387 [cm-1] > omega( 6) = 6.821788 [THz] = 227.551869 [cm-1] > omega( 7) = 6.821788 [THz] = 227.551870 [cm-1] > omega( 8) = 8.276312 [THz] = 276.069874 [cm-1] > omega( 9) = 8.276312 [THz] = 276.069882 [cm-1] > omega(10) = 10.491723 [THz] = 349.968534 [cm-1] > omega(11) = 13.134275 [THz] = 438.115167 [cm-1] > omega(12) = 15.696771 [THz] = 523.591396 [cm-1] > omega(13) = 15.903656 [THz] = 530.492373 [cm-1] > omega(14) = 18.187294 [THz] = 606.666874 [cm-1] > omega(15) = 18.187294 [THz] = 606.666875 [cm-1] > omega(16) = 18.222317 [THz] = 607.835092 [cm-1] > omega(17) = 18.222317 [THz] = 607.835098 [cm-1] > omega(18) = 22.768578 [THz] = 759.483060 [cm-1] > omega(19) = 23.037730 [THz] = 768.461058 [cm-1] > omega(20) = 23.744563 [THz] = 792.038631 [cm-1] > omega(21) = 23.744563 [THz] = 792.038632 [cm-1] > that is to say that all the frequencies are underestimated after using q2r.x and matdyn.x. what's the prpblem with it? > Thank you very much. > ka ?????? > On Mon, 9 Oct 2006, Paolo Giannozzi wrote: > >> On Oct 9, 2006, at 16:15 , li yan wrote: >> >>> I calculated the phonon frequencies of FeO for a list of generic >>> q vectors with q2r.x and matdyn.x. In the FeO.dyn files , the highest >>> frequency is about 43 THz, while it is about 22 THz in the matdyn.modes >>> file and FeO.freq file. What problems may occur in the the process? >> >> wrong masses? Paolo >> > > What q-points and modes of FeO? Are you sure about units, THz? > > Gosia > > >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > > > --------------------------------- > Mp3??????-?????????????? From jiaanyan at gmail.com Tue Oct 10 20:31:00 2006 From: jiaanyan at gmail.com (Jia-An Yan) Date: Tue, 10 Oct 2006 14:31:00 -0400 Subject: [Pw_forum] error in the el-ph calculation Message-ID: <6581de880610101131r1c8e39dei498e84e9cf112d25@mail.gmail.com> Hi everyone, I am running the el-phonon calculations for the Al in the example07. When I get to the al.elph.in part, the program always stops at the third q-points with an error message as follows: %%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% task # 2 from potinit : error # 1 starting and expected charges differ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% and in the al.elph.out it says: The potential is recalculated from file : al.save/charge-density.xml %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from potinit : error # 1 starting and expected charges differ %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% I use the original input file in the example. I don't know what causes such error. Can anyone here help me? Thanks! Sincerely, Jia-An Yan -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061010/62c49594/attachment.htm From lucky8121 at sohu.com Wed Oct 11 03:14:56 2006 From: lucky8121 at sohu.com (lucky8121) Date: Wed, 11 Oct 2006 09:14:56 +0800 Subject: [Pw_forum] relaxation of even-layer slab Message-ID: <20061011011233.8A97300B403E@smtp127.sohu.com> Dear users, I am a new user of PWSCF. I am doing the structure relaxation of metal slabs. The relaxations (interlayer spacings) on the two sides of the odd-layer slab are symmetric. However, I met problem with the relaxation of even-layer slabs. The initial positions of the Cu(111) slab were: Cu 0.000000000 0.577350269190 -1.224744871392 Cu 0.000000000 0.000000000000 -0.408248290464 Cu 0.500000000 0.288675134595 0.408248290464 Cu 0.000000000 0.577350269190 1.224744871392 After relaxation (forc_conv_thr = 4.0D-4), the final positions are: Cu 0.000000000 0.577350269 -1.207031882 Cu 0.000000000 0.000000000 -0.404842681 Cu 0.500000000 0.288675135 0.404671325 Cu 0.000000000 0.577350269 1.207203237 As you can see, the relaxations on the two sides of the slab are different. The same case occured whether I set "nosym" as ".true." or ".false.". This is strange. I wonder whether someone else has met the same problem and how he or she solved it. Thanks. By the way, I have used a vacuum of 12.7 angstrom. Best regards lucky lucky8121 at sohu.com 2006-10-11 From baroni at sissa.it Wed Oct 11 04:03:01 2006 From: baroni at sissa.it (Stefano Baroni) Date: Wed, 11 Oct 2006 04:03:01 +0200 Subject: [Pw_forum] relaxation of even-layer slab In-Reply-To: <20061011011233.8A97300B403E@smtp127.sohu.com> References: <20061011011233.8A97300B403E@smtp127.sohu.com> Message-ID: <1C55A641-5FED-4488-93E0-A2C805151AEB@sissa.it> Dear Lucky: the non-symmetric part of the relaxation is very small and possibly not very meaningful. Nevertheless, I agree that if reflection were an exact symmetry of your system, this asymmetry should vanish. One possible explanation could be that reflection symmetry is (slightly) violated by your numerical setup, e.g. the real-space fft grid could be not symmetryc. Stefano On Oct 11, 2006, at 3:14 AM, lucky8121 wrote: > Dear users, > > I am a new user of PWSCF. I am doing the structure relaxation of > metal slabs. > The relaxations (interlayer spacings) on the two sides of the odd- > layer slab are symmetric. However, I met problem with the > relaxation of even-layer slabs. > > The initial positions of the Cu(111) slab were: > Cu 0.000000000 0.577350269190 -1.224744871392 > Cu 0.000000000 0.000000000000 -0.408248290464 > Cu 0.500000000 0.288675134595 0.408248290464 > Cu 0.000000000 0.577350269190 1.224744871392 > > After relaxation (forc_conv_thr = 4.0D-4), the final positions are: > Cu 0.000000000 0.577350269 -1.207031882 > Cu 0.000000000 0.000000000 -0.404842681 > Cu 0.500000000 0.288675135 0.404671325 > Cu 0.000000000 0.577350269 1.207203237 > > As you can see, the relaxations on the two sides of the slab are > different. The same case occured whether I set "nosym" as ".true." > or ".false.". This is strange. I wonder whether someone else has > met the same problem and how he or she solved it. Thanks. > > By the way, I have used a vacuum of 12.7 angstrom. > > Best regards > > lucky > lucky8121 at sohu.com > 2006-10-11 > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061011/2e5aeb6c/attachment.htm From serenechan at charteredsemi.com Wed Oct 11 06:00:29 2006 From: serenechan at charteredsemi.com (serenechan at charteredsemi.com) Date: Wed, 11 Oct 2006 12:00:29 +0800 Subject: [Pw_forum] CHAN Serene is out of the office. Message-ID: I will be out of the office starting 10/11/2006 and will not return until 10/16/2006. Please approach my colleague Wang Xinzi for all tapeout related matters. From wierzbom at ts.infn.it Wed Oct 11 10:27:56 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Wed, 11 Oct 2006 10:27:56 +0200 (CEST) Subject: [Pw_forum] error in the el-ph calculation In-Reply-To: <6581de880610101131r1c8e39dei498e84e9cf112d25@mail.gmail.com> References: <6581de880610101131r1c8e39dei498e84e9cf112d25@mail.gmail.com> Message-ID: On Tue, 10 Oct 2006, Jia-An Yan wrote: Hi Jia-An Yan, What espresso version do you use? There is some change in this part between espresso-3.0 and espresso-3.1. Can you clean tmp directory and repeat everything using espresso-3.1? Gosia > Hi everyone, > > I am running the el-phonon calculations for the Al in the example07. When I > get to the al.elph.in part, the program always stops at the third q-points > with an error message as follows: > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > task # 2 > from potinit : error # 1 > starting and expected charges differ > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > and in the al.elph.out it says: > > The potential is recalculated from file : > al.save/charge-density.xml > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from potinit : error # 1 > starting and expected charges differ > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > I use the original input file in the example. I don't know what causes such > error. > Can anyone here help me? Thanks! > > Sincerely, > Jia-An Yan > From wmbmacam at lg.ehu.es Wed Oct 11 12:15:53 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Wed, 11 Oct 2006 12:15:53 +0200 Subject: [Pw_forum] More pseudopotential troubles In-Reply-To: <452A1602.3080805@lg.ehu.es> References: <4523D7B5.9050609@lg.ehu.es> <452A1602.3080805@lg.ehu.es> Message-ID: <452CC459.3020408@lg.ehu.es> Dear people, Although your previous comments on Ca pseudopotential were very helpful, now I'm finding another problem. I may be wrong to think that the input attached generates an US pseudopotential with semicore states, using l=2 as the pseudization channel and two reference energies for the s and p channels. Anyway, when I try to run this input &input atom='Ca', beta = 0.2, rel=1, config='[Ar] 3d0 4s2 4p0', iswitch=3, dft='PBE' / &inputp pseudotype=3, lloc=2, nlcc =.true., file_pseudopw='Ca.pbe-nsp-1.UPF', rho0=0.01 / 5 3S 1 0 2.00 0.00 1.10 1.30 3S 2 0 0.00 -0.27 2.80 2.90 3P 2 1 6.00 0.00 1.20 1.30 3P 3 1 0.00 -0.10 2.80 2.90 3D 3 2 0.00 0.00 2.90 2.90 &test nconf = 1, file_pseudo='Ca.pbe-nsp-1.UPF', ecutmin = 20.0, ecutmax = 60.0, / 3 3S 2 0 2.00 0.00 2.80 2.90 3P 3 1 0.00 0.00 2.80 2.90 3D 3 2 0.00 0.00 2.90 2.90 I get the following error: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from ascheqps : error # 1 too many attempts %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... Increasing beta doesn't help. Neither does increasing the maximum number of iterations in ascheqps.f90. I'm currently looking at that module to understand what's going wrong. Nevertheless, if you saw something stupid in my input, pleas tell me. Thanks in advance, Miguel PS: In the excited state for the s channel, 3S 2 0, do I have to write the pseudization energy? Or is 0.00 the corresponding energy to the excited state? From giannozz at nest.sns.it Wed Oct 11 17:46:57 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 11 Oct 2006 17:46:57 +0200 Subject: [Pw_forum] born effective charges In-Reply-To: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> References: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> Message-ID: On Oct 9, 2006, at 16:40 , Vivek Ranjan wrote: > I am trying to calculate born effective charges for a system > consisting of > more than hundred atoms. I am doing it using linear response, i.e. > I relax > the structure, then I do a scf calculation and then I run a gamma > point > phonon calculation. After the programme prints the dielectric > constants, > it must calculate the born effective charges. With the given computer > resources at my disposal, it cannot complete this part. So, I must > restart > the calculation from the previous step. I use "recover.true." flag. > But it > never seems to go beyond the dielectric constant calculation part. Is > there a way I can do this calculation ? last time I tried - admittedly not very recently - it was working. I rewrote part of the restart procedure exactly for the same purpose. Have you tried to check if it works by restarting from a previous run with a small system? Paolo From lucky8121 at sohu.com Thu Oct 12 04:38:52 2006 From: lucky8121 at sohu.com (lucky8121) Date: Thu, 12 Oct 2006 10:38:52 +0800 Subject: [Pw_forum] relaxation of even-layer slab Message-ID: <20061012023627.9A425002182E@smtp126.sohu.com> Dear Prof. Baroni? ?Thanks for your prompt reply. (1) The four-layer slab I did get warnings about the setup of the four-layer slab in the output file: warning: symmetry operation # 4 not allowed. fractional translation: 0.3333333 -0.3333333 0.0000000 in crystal coordinates warning: symmetry operation # 11 not allowed. fractional translation: 0.3333333 -0.3333333 0.0000000 in crystal coordinates warning: symmetry operation # 12 not allowed. fractional translation: 0.3333333 -0.3333333 0.0000000 in crystal coordinates warning: symmetry operation # 13 not allowed. fractional translation: 0.3333333 -0.3333333 0.0000000 in crystal coordinates warning: symmetry operation # 19 not allowed. fractional translation: 0.3333333 -0.3333333 0.0000000 in crystal coordinates warning: symmetry operation # 20 not allowed. fractional translation: 0.3333333 -0.3333333 0.0000000 in crystal coordinates The real-space FFT grids in this calculation were: G cutoff = 146.8225 ( 47267 G-vectors) FFT grid: ( 25, 25,180) G cutoff = 73.4112 ( 16847 G-vectors) smooth grid: ( 18, 18,125) (2) The six-layer slab Then I performed a calculation of a six-layer slab. The initial positions were: Cu 0.5000000 0.2886751 -1.632993 Cu 0.0000000 0.5773505 -0.816497 Cu 0.0000000 0.0000000 0.000000 Cu 0.5000000 0.2886751 0.816497 Cu 0.0000000 0.5773503 1.632993 Cu 0.0000000 0.0000000 2.449490 After relaxation, the final positions are: Cu 0.500000000 0.288675100 -1.626940097 Cu 0.000000000 0.577350500 -0.819973440 Cu 0.000000000 0.000000000 -0.003573650 Cu 0.500000000 0.288675100 0.818345337 Cu 0.000000000 0.577350300 1.634784508 Cu 0.000000000 0.000000000 2.446847342 The interlayer spacings on the two sides of the slab are 0.812062834 and 0.806966657, respectively. This time the asymmetry becomes larger than the case of four-layer slab. I noticed that in the output file there are warnings as below: warning: symmetry operation # 4 not allowed. fractional translation: 0.3333334 -0.3333333 -0.0909088 in crystal coordinates warning: symmetry operation # 11 not allowed. fractional translation: 0.3333334 -0.3333333 -0.0909088 in crystal coordinates warning: symmetry operation # 12 not allowed. fractional translation: 0.3333334 -0.3333333 -0.0909088 in crystal coordinates warning: symmetry operation # 13 not allowed. fractional translation: 0.3333334 -0.3333333 -0.0909088 in crystal coordinates warning: symmetry operation # 19 not allowed. fractional translation: 0.3333334 -0.3333333 -0.0909088 in crystal coordinates warning: symmetry operation # 20 not allowed. fractional translation: 0.3333334 -0.3333333 -0.0909088 in crystal coordinates And the FFT grids are: G cutoff = 146.8225 ( 57757 G-vectors) FFT grid: ( 25, 25,225) G cutoff = 73.4112 ( 20595 G-vectors) smooth grid: ( 18, 18,160) Since I have six atom layers and six vacuume layers, it seems that the code translated the FFT grids along the z direction about eleventh of the z-axis length to judge whether the grid (nr3) can coincide with itself. It failed due to 225 can't be divided by 11. I have tried to set "nr1=25, nr2=25, nr3 =220" manually (also nr3=231 and 242), but these FFT grids were not allowed. A solution to this may be to change the FFT library which allows factor of 11, and recompile the PWSCF code. Another solution to this problem may be to reduce or increase the vacuum layers so that the FFT grids are appropriate for the symmetric operations. But since I want to see the dependence of relaxations of surface layers on the slab thickness, I have to fix the vacuum thickness for different slabs. Is there anyway to get symmetric relaxations of different even-layer slabs? For example, is it possible to shift the FFT grid at the setup so that it coincides with symmetry of different slabs? Best regards, Lucky Dear Lucky: the non-symmetric part of the relaxation is very small and possibly not very meaningful. Nevertheless, I agree that if reflection were an exact symmetry of your system, this asymmetry should vanish. One possible explanation could be that reflection symmetry is (slightly) violated by your numerical setup, e.g. the real-space fft grid could be not symmetryc.?/DIV> Stefano On Oct 11, 2006, at 3:14 AM, lucky8121 wrote: Dear users, I am a new user of PWSCF. I am doing the structure relaxation of metal slabs. The relaxations (interlayer spacings) on the two sides of the odd-layer slab are symmetric. However, I met problem with the relaxation of even-layer slabs. The initial positions of the Cu(111) slab were: Cu ???0.000000000 ?0.577350269190?-1.224744871392 Cu ???0.000000000 ?0.000000000000?-0.408248290464 Cu ???0.500000000 ?0.288675134595 ?0.408248290464 Cu ???0.000000000 ?0.577350269190 ?1.224744871392 After relaxation (forc_conv_thr = 4.0D-4), the final positions are: Cu ???0.000000000 ?0.577350269?-1.207031882 Cu ???0.000000000 ?0.000000000?-0.404842681 Cu ???0.500000000 ?0.288675135 ?0.404671325 Cu ???0.000000000 ?0.577350269 ?1.207203237 As you can see, the relaxations on the two sides of the slab are different. The same case occured whether I set "nosym" as ".true." or ".false.". This is strange. I wonder whether someone else has met the same problem and how he or she solved it. Thanks. By the way, I have used a vacuum of 12.7 angstrom. Best regards lucky lucky8121 at sohu.com 2006-10-11 _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA?&?DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't?send me MS Word or PowerPoint attachments Why? See:?http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061012/1ffeb861/attachment.htm From degironc at sissa.it Thu Oct 12 10:14:25 2006 From: degironc at sissa.it (Stefano de Gironcoli) Date: Thu, 12 Oct 2006 10:14:25 +0200 (CEST) Subject: [Pw_forum] relaxation of even-layer slab In-Reply-To: <20061012023627.9A425002182E@smtp126.sohu.com> References: <20061012023627.9A425002182E@smtp126.sohu.com> Message-ID: you could define the positions simmetrically w.r.t the origin, in this way there will be no fractional translation along z but only some f.t. along x and y, and take care that the nr1 and nr2 are multiples of 3 so that also these f.t are allowed. In any case the fact that a symmetry that should be there is not enforced (due to grid problems) affects the number of k-points used but should not change the physics of the system. If you see diffrences between the two surfaces this can be due to 1) insufficient self consistency that lead to some noise in the forces, hence in the relaxed geometry. 2) charge and potential FFT grid which is not dense enough, so that the energy slightly depends on the grid origin making the two surfaces inequivalent. stefano On Thu, 12 Oct 2006, lucky8121 wrote: > Dear Prof. Baroni?? > > ??Thanks for your prompt reply. > > (1) The four-layer slab > I did get warnings about the setup of the four-layer slab in the output file: > warning: symmetry operation # 4 not allowed. fractional translation: > 0.3333333 -0.3333333 0.0000000 in crystal coordinates > warning: symmetry operation # 11 not allowed. fractional translation: > 0.3333333 -0.3333333 0.0000000 in crystal coordinates > warning: symmetry operation # 12 not allowed. fractional translation: > 0.3333333 -0.3333333 0.0000000 in crystal coordinates > warning: symmetry operation # 13 not allowed. fractional translation: > 0.3333333 -0.3333333 0.0000000 in crystal coordinates > warning: symmetry operation # 19 not allowed. fractional translation: > 0.3333333 -0.3333333 0.0000000 in crystal coordinates > warning: symmetry operation # 20 not allowed. fractional translation: > 0.3333333 -0.3333333 0.0000000 in crystal coordinates > > The real-space FFT grids in this calculation were: > G cutoff = 146.8225 ( 47267 G-vectors) FFT grid: ( 25, 25,180) > G cutoff = 73.4112 ( 16847 G-vectors) smooth grid: ( 18, 18,125) > > (2) The six-layer slab > Then I performed a calculation of a six-layer slab. The initial positions were: > Cu 0.5000000 0.2886751 -1.632993 > Cu 0.0000000 0.5773505 -0.816497 > Cu 0.0000000 0.0000000 0.000000 > Cu 0.5000000 0.2886751 0.816497 > Cu 0.0000000 0.5773503 1.632993 > Cu 0.0000000 0.0000000 2.449490 > > After relaxation, the final positions are: > Cu 0.500000000 0.288675100 -1.626940097 > Cu 0.000000000 0.577350500 -0.819973440 > Cu 0.000000000 0.000000000 -0.003573650 > Cu 0.500000000 0.288675100 0.818345337 > Cu 0.000000000 0.577350300 1.634784508 > Cu 0.000000000 0.000000000 2.446847342 > > The interlayer spacings on the two sides of the slab are 0.812062834 and 0.806966657, respectively. This time the asymmetry becomes larger than the case of four-layer slab. > > I noticed that in the output file there are warnings as below: > warning: symmetry operation # 4 not allowed. fractional translation: > 0.3333334 -0.3333333 -0.0909088 in crystal coordinates > warning: symmetry operation # 11 not allowed. fractional translation: > 0.3333334 -0.3333333 -0.0909088 in crystal coordinates > warning: symmetry operation # 12 not allowed. fractional translation: > 0.3333334 -0.3333333 -0.0909088 in crystal coordinates > warning: symmetry operation # 13 not allowed. fractional translation: > 0.3333334 -0.3333333 -0.0909088 in crystal coordinates > warning: symmetry operation # 19 not allowed. fractional translation: > 0.3333334 -0.3333333 -0.0909088 in crystal coordinates > warning: symmetry operation # 20 not allowed. fractional translation: > 0.3333334 -0.3333333 -0.0909088 in crystal coordinates > And the FFT grids are: > G cutoff = 146.8225 ( 57757 G-vectors) FFT grid: ( 25, 25,225) > G cutoff = 73.4112 ( 20595 G-vectors) smooth grid: ( 18, 18,160) > > Since I have six atom layers and six vacuume layers, it seems that the code translated the FFT grids along the z direction about eleventh of the z-axis length to judge whether the grid (nr3) can coincide with itself. It failed due to 225 can't be divided by 11. I have tried to set "nr1=25, nr2=25, nr3 =220" manually (also nr3=231 and 242), but these FFT grids were not allowed. > A solution to this may be to change the FFT library which allows factor of 11, and recompile the PWSCF code. Another solution to this problem may be to reduce or increase the vacuum layers so that the FFT grids are appropriate for the symmetric operations. But since I want to see the dependence of relaxations of surface layers on the slab thickness, I have to fix the vacuum thickness for different slabs. Is there anyway to get symmetric relaxations of different even-layer slabs? For example, is it possible to shift the FFT grid at the setup so that it coincides with symmetry of different slabs? > > Best regards, > Lucky > > > Dear Lucky: > > > the non-symmetric part of the relaxation is very small and possibly not very meaningful. Nevertheless, I agree that if reflection were an exact symmetry of your system, this asymmetry should vanish. One possible explanation could be that reflection symmetry is (slightly) violated by your numerical setup, e.g. the real-space fft grid could be not symmetryc.?/DIV> > > > Stefano > > > On Oct 11, 2006, at 3:14 AM, lucky8121 wrote: > > > Dear users, > > > I am a new user of PWSCF. I am doing the structure relaxation of metal slabs. > The relaxations (interlayer spacings) on the two sides of the odd-layer slab are symmetric. However, I met problem with the relaxation of even-layer slabs. > > > The initial positions of the Cu(111) slab were: > Cu ???0.000000000 ?0.577350269190?-1.224744871392 > Cu ???0.000000000 ?0.000000000000?-0.408248290464 > Cu ???0.500000000 ?0.288675134595 ?0.408248290464 > Cu ???0.000000000 ?0.577350269190 ?1.224744871392 > > > After relaxation (forc_conv_thr = 4.0D-4), the final positions are: > Cu ???0.000000000 ?0.577350269?-1.207031882 > Cu ???0.000000000 ?0.000000000?-0.404842681 > Cu ???0.500000000 ?0.288675135 ?0.404671325 > Cu ???0.000000000 ?0.577350269 ?1.207203237 > > > As you can see, the relaxations on the two sides of the slab are different. The same case occured whether I set "nosym" as ".true." or ".false.". This is strange. I wonder whether someone else has met the same problem and how he or she solved it. Thanks. > > > By the way, I have used a vacuum of 12.7 angstrom. > > > Best regards > > > lucky > lucky8121 at sohu.com > 2006-10-11 > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > > --- > Stefano Baroni - SISSA?&?DEMOCRITOS National Simulation Center - Trieste > [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) > > > Please, if possible, don't?send me MS Word or PowerPoint attachments > Why? See:?http://www.gnu.org/philosophy/no-word-attachments.html > From babazade at gmail.com Thu Oct 12 11:00:41 2006 From: babazade at gmail.com (Amin Babazadeh) Date: Thu, 12 Oct 2006 12:30:41 +0330 Subject: [Pw_forum] about generating pseudopotential Message-ID: dear users When we use GGA approximation we can use many exchange-correlation potential for example PW91 exch-corr GGA ,PBE (Perdew-Burke-Ernzerhof), EV-GGA (Engel-Vosko),....... I found some of the pseudopotential in the http://www.pwscf.org/pseudo.htmthat have been generated with some of the exch-corr functions but i need the pseudopotential with EV-GGA (Engel-Vosko) exch-corr potential. I want to generate this potential but i don't know if i can do it whit using USPP ``*U*ltra*S*oft *P*seudo*P*otential'' code or not. I haven't worked whit this code until yet and i want to know that if this code in capable of doing my job i start learning it. Exactly what i need is I want to use the function that Engel and Vosko proposed for exchange-correlation potential {Exact exchange-only potential and ...., Phys.Rev B47 13164 (1993)} instead of PW91. I will appreciate if some one help me and tell me if it is possible to change exchange correlation potential in USPP or recommend me a code for generating pseudopotential. Sincerely yours -- AMIN -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061012/ef81632c/attachment.htm From giannozz at nest.sns.it Thu Oct 12 11:12:14 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 12 Oct 2006 11:12:14 +0200 Subject: [Pw_forum] about projwfc.x In-Reply-To: References: Message-ID: On Oct 10, 2006, at 10:47 , Ruslan Minibaev wrote: > from read_restart_wfc : error # 1 > Data Section not present in restart file > What's the reason of the error? it is written there : "Data Section not present in restart file" :-) Seriously: what version are you using? there is no such error message in recent versions. Paolo From wierzbom at ts.infn.it Thu Oct 12 11:16:01 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Thu, 12 Oct 2006 11:16:01 +0200 (CEST) Subject: [Pw_forum] about generating pseudopotential In-Reply-To: References: Message-ID: On Thu, 12 Oct 2006, Amin Babazadeh wrote: > dear users > > When we use GGA approximation we can use many exchange-correlation potential > for example PW91 exch-corr GGA ,PBE (Perdew-Burke-Ernzerhof), EV-GGA > (Engel-Vosko),....... > I found some of the pseudopotential in the > http://www.pwscf.org/pseudo.htmthat have been generated with some of > the exch-corr > functions but i need the pseudopotential with EV-GGA (Engel-Vosko) exch-corr > potential. I want to generate this potential but i don't know if i can do it > whit using USPP ``*U*ltra*S*oft *P*seudo*P*otential'' code or not. I > haven't worked whit this code until yet and i want to know that if this code > in capable of doing my job i start learning it. > > Exactly what i need is > I want to use the function that Engel and Vosko proposed for > exchange-correlation potential {Exact exchange-only potential and ...., > Phys.Rev B47 13164 (1993)} instead of PW91. > The ultrasoft code uspp-735 can only generate USPP with the DFT functionals: exfact: specify the type of exchange-correlation (CA means ceperley-alder in perdew-zunger parametrization) 0 LDA: CA -1 LDA: wigner -2 LDA: hedin-lundquist -3 LDA: gunnarson-lundquist 1 GGA: CA + Becke88 + LYP 2 GGA: CA + Becke88 3 GGA: CA + Becke88 + Perdew86 4 GGA: PW(91) 5 GGA: PBE(96) To change the functional you need to set the proper exfact parameter in the input_ae. For the exact exchange you need other tools. See EXX_example in the last espresso version and ask Stefano de Gironcoli for more details. Gosia > I will appreciate if some one help me and tell me if it is possible to > change exchange correlation potential in USPP or recommend me a code for > generating pseudopotential. > Sincerely yours > From giannozz at nest.sns.it Thu Oct 12 11:18:10 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 12 Oct 2006 11:18:10 +0200 Subject: [Pw_forum] phonon at a specifiec point In-Reply-To: References: Message-ID: <3FC95AC1-9C3E-4F3F-8B28-443948FAABE1@nest.sns.it> On Oct 10, 2006, at 11:53 , Ezad Shojaee wrote: > performing phonon calculation at a speciefic point, it seems that > the program writes the values of phonon frequencies in the output > file , at the end of all. if i want to know the value of first mode > during the calculation where should i search for ? what do you mean by "the value of first mode during the calculation"??? You can have the values of the dynamical matrix elements, but phonon frequencies are not available util the end of the calculation and the diagonalization of the entire dynamical matrix Paolo From lucky8121 at sohu.com Thu Oct 12 14:02:15 2006 From: lucky8121 at sohu.com (lucky8121) Date: Thu, 12 Oct 2006 20:02:15 +0800 Subject: [Pw_forum] relaxation of even-layer slab Message-ID: <20061012115952.52A060271412@smtp127.sohu.com> Dear Stefano, Thanks a lot for the help from both Stefano de Gironcoli and Stefano Baroni. I think I know how to do the calculation now. I have just joined in the member list of pw_forum for two days. Its activity impresses me very much. :) Best wishes, Lucky > >you could define the positions simmetrically w.r.t the origin, in this way >there will be no fractional translation along z but only some f.t. along x >and y, and take care that the nr1 and nr2 are multiples of 3 so that also >these f.t are allowed. > >In any case the fact that a symmetry that should be there is not enforced >(due to grid problems) affects the number of k-points used but should not >change the physics of the system. >If you see diffrences between the two surfaces this can be due to >1) insufficient self consistency that lead to some noise in the forces, >hence in the relaxed geometry. >2) charge and potential FFT grid which is not dense enough, so that the >energy slightly depends on the grid origin making the two surfaces >inequivalent. > >stefano > > > >On Thu, 12 Oct 2006, lucky8121 wrote: > >> Dear Prof. Baroni? >> >> ?Thanks for your prompt reply. >> >> (1) The four-layer slab >> I did get warnings about the setup of the four-layer slab in the output file: >> warning: symmetry operation # 4 not allowed. fractional translation: >> 0.3333333 -0.3333333 0.0000000 in crystal coordinates >> warning: symmetry operation # 11 not allowed. fractional translation: >> 0.3333333 -0.3333333 0.0000000 in crystal coordinates >> warning: symmetry operation # 12 not allowed. fractional translation: >> 0.3333333 -0.3333333 0.0000000 in crystal coordinates >> warning: symmetry operation # 13 not allowed. fractional translation: >> 0.3333333 -0.3333333 0.0000000 in crystal coordinates >> warning: symmetry operation # 19 not allowed. fractional translation: >> 0.3333333 -0.3333333 0.0000000 in crystal coordinates >> warning: symmetry operation # 20 not allowed. fractional translation: >> 0.3333333 -0.3333333 0.0000000 in crystal coordinates >> >> The real-space FFT grids in this calculation were: >> G cutoff = 146.8225 ( 47267 G-vectors) FFT grid: ( 25, 25,180) >> G cutoff = 73.4112 ( 16847 G-vectors) smooth grid: ( 18, 18,125) >> >> (2) The six-layer slab >> Then I performed a calculation of a six-layer slab. The initial positions were: >> Cu 0.5000000 0.2886751 -1.632993 >> Cu 0.0000000 0.5773505 -0.816497 >> Cu 0.0000000 0.0000000 0.000000 >> Cu 0.5000000 0.2886751 0.816497 >> Cu 0.0000000 0.5773503 1.632993 >> Cu 0.0000000 0.0000000 2.449490 >> >> After relaxation, the final positions are: >> Cu 0.500000000 0.288675100 -1.626940097 >> Cu 0.000000000 0.577350500 -0.819973440 >> Cu 0.000000000 0.000000000 -0.003573650 >> Cu 0.500000000 0.288675100 0.818345337 >> Cu 0.000000000 0.577350300 1.634784508 >> Cu 0.000000000 0.000000000 2.446847342 >> >> The interlayer spacings on the two sides of the slab are 0.812062834 and 0.806966657, respectively. This time the asymmetry becomes larger than the case of four-layer slab. >> >> I noticed that in the output file there are warnings as below: >> warning: symmetry operation # 4 not allowed. fractional translation: >> 0.3333334 -0.3333333 -0.0909088 in crystal coordinates >> warning: symmetry operation # 11 not allowed. fractional translation: >> 0.3333334 -0.3333333 -0.0909088 in crystal coordinates >> warning: symmetry operation # 12 not allowed. fractional translation: >> 0.3333334 -0.3333333 -0.0909088 in crystal coordinates >> warning: symmetry operation # 13 not allowed. fractional translation: >> 0.3333334 -0.3333333 -0.0909088 in crystal coordinates >> warning: symmetry operation # 19 not allowed. fractional translation: >> 0.3333334 -0.3333333 -0.0909088 in crystal coordinates >> warning: symmetry operation # 20 not allowed. fractional translation: >> 0.3333334 -0.3333333 -0.0909088 in crystal coordinates >> And the FFT grids are: >> G cutoff = 146.8225 ( 57757 G-vectors) FFT grid: ( 25, 25,225) >> G cutoff = 73.4112 ( 20595 G-vectors) smooth grid: ( 18, 18,160) >> >> Since I have six atom layers and six vacuume layers, it seems that the code translated the FFT grids along the z direction about eleventh of the z-axis length to judge whether the grid (nr3) can coincide with itself. It failed due to 225 can't be divided by 11. I have tried to set "nr1=25, nr2=25, nr3 =220" manually (also nr3=231 and 242), but these FFT grids were not allowed. >> A solution to this may be to change the FFT library which allows factor of 11, and recompile the PWSCF code. Another solution to this problem may be to reduce or increase the vacuum layers so that the FFT grids are appropriate for the symmetric operations. But since I want to see the dependence of relaxations of surface layers on the slab thickness, I have to fix the vacuum thickness for different slabs. Is there anyway to get symmetric relaxations of different even-layer slabs? For example, is it possible to shift the FFT grid at the setup so that it coincides with symmetry of different slabs? >> >> Best regards, >> Lucky >> >> >> Dear Lucky: >> >> >> the non-symmetric part of the relaxation is very small and possibly not very meaningful. Nevertheless, I agree that if reflection were an exact symmetry of your system, this asymmetry should vanish. One possible explanation could be that reflection symmetry is (slightly) violated by your numerical setup, e.g. the real-space fft grid could be not symmetryc.?/DIV> >> >> >> Stefano >> >> >> On Oct 11, 2006, at 3:14 AM, lucky8121 wrote: >> >> >> Dear users, >> >> >> I am a new user of PWSCF. I am doing the structure relaxation of metal slabs. >> The relaxations (interlayer spacings) on the two sides of the odd-layer slab are symmetric. However, I met problem with the relaxation of even-layer slabs. >> >> >> The initial positions of the Cu(111) slab were: >> Cu ???0.000000000 ?0.577350269190?-1.224744871392 >> Cu ???0.000000000 ?0.000000000000?-0.408248290464 >> Cu ???0.500000000 ?0.288675134595 ?0.408248290464 >> Cu ???0.000000000 ?0.577350269190 ?1.224744871392 >> >> >> After relaxation (forc_conv_thr = 4.0D-4), the final positions are: >> Cu ???0.000000000 ?0.577350269?-1.207031882 >> Cu ???0.000000000 ?0.000000000?-0.404842681 >> Cu ???0.500000000 ?0.288675135 ?0.404671325 >> Cu ???0.000000000 ?0.577350269 ?1.207203237 >> >> >> As you can see, the relaxations on the two sides of the slab are different. The same case occured whether I set "nosym" as ".true." or ".false.". This is strange. I wonder whether someone else has met the same problem and how he or she solved it. Thanks. >> >> >> By the way, I have used a vacuum of 12.7 angstrom. >> >> >> Best regards >> >> >> lucky >> lucky8121 at sohu.com >> 2006-10-11 >> >> >> >> >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> >> >> --- >> Stefano Baroni - SISSA?&?DEMOCRITOS National Simulation Center - Trieste >> [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) >> >> >> Please, if possible, don't?send me MS Word or PowerPoint attachments >> Why? See:?http://www.gnu.org/philosophy/no-word-attachments.html >> From hsd22 at hermes.cam.ac.uk Thu Oct 12 17:38:24 2006 From: hsd22 at hermes.cam.ac.uk (H.S.Domingos) Date: Thu, 12 Oct 2006 16:38:24 +0100 (BST) Subject: [Pw_forum] nscf Message-ID: Dear All, I would like to ask if using nscf is the same as using scf with electron_maxstep = 1 and startingpot = 'file from previous scf or vc-relax run' ? Thank you very much. Best regards, Helder ======================================================================= | Dr. Helder S. Domingos | | | | INESC Microsyst & Nanotechnol, Lisbon, P-1000 Portugal | | and | | R&D unit for Molecular Chemical Physics | | Chemistry Department, University of Coimbra | ======================================================================= From giannozz at nest.sns.it Thu Oct 12 18:59:01 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 12 Oct 2006 18:59:01 +0200 Subject: [Pw_forum] born effective charges In-Reply-To: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> References: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> Message-ID: <617FA167-CF4F-47A6-9B09-33EAC7EC898F@nest.sns.it> On Oct 9, 2006, at 16:40 , Vivek Ranjan wrote: > After the programme prints the dielectric constants, it must calculate > the born effective charges. With the given computer resources at my > disposal, it cannot complete this part. So, I must restart the > calculation > from the previous step. I use "recover.true." flag. But it never seems > to go beyond the dielectric constant calculation part. there is something I don't understand here. The calculation of the dielectric response takes something of the order of 3 times the time needed for the scf calculation on the same system. Once this is done, the calculation of effective charges takes close to nothing. What is the part that you cannot complete? Use 'epsil=.true., trans=.false. to skip the calculation of phonon modes (which will take a lot of time if you have a lot of atoms) Paolo From vranjan at ncsu.edu Thu Oct 12 19:16:37 2006 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Thu, 12 Oct 2006 13:16:37 -0400 (EDT) Subject: [Fwd: Re: [Pw_forum] born effective charges] In-Reply-To: <617FA167-CF4F-47A6-9B09-33EAC7EC898F@nest.sns.it> References: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> <617FA167-CF4F-47A6-9B09-33EAC7EC898F@nest.sns.it> Message-ID: <46342.152.14.74.117.1160673397.squirrel@webmail.ncsu.edu> thats exactly what i felt. the programme does not print anything after it prints the dielectric constant. then i figured out from the forum that some ppl found their programme crash if npool > 1. so, i also set npool to 1 and then one of the calculations went beyond the calculation of dielectric constant, but the other still has the problem (i have two slightly different systems). so, i am still in the process of understanding what could be done. presently, i feel its not a restart problem.it is some problem with the calculation itself, and the programme does not print anything, it stays there without doing anything till the allocated time is over. thanks, vivek > > On Oct 9, 2006, at 16:40 , Vivek Ranjan wrote: > >> After the programme prints the dielectric constants, it must calculate >> the born effective charges. With the given computer resources at my >> disposal, it cannot complete this part. So, I must restart the >> calculation >> from the previous step. I use "recover.true." flag. But it never seems >> to go beyond the dielectric constant calculation part. > > there is something I don't understand here. The calculation of the > dielectric response takes something of the order of 3 times the time > needed for the scf calculation on the same system. Once this is > done, the calculation of effective charges takes close to nothing. > What is the part that you cannot complete? Use 'epsil=.true., > trans=.false. to skip the calculation of phonon modes (which > will take a lot of time if you have a lot of atoms) > > Paolo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________________________________ 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 -------------------------------------------------------------------------- ____________________________________________________ From giannozz at nest.sns.it Thu Oct 12 22:40:46 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 12 Oct 2006 22:40:46 +0200 Subject: [Pw_forum] born effective charges In-Reply-To: <46342.152.14.74.117.1160673397.squirrel@webmail.ncsu.edu> References: <40109.152.14.74.117.1160404841.squirrel@webmail.ncsu.edu> <617FA167-CF4F-47A6-9B09-33EAC7EC898F@nest.sns.it> <46342.152.14.74.117.1160673397.squirrel@webmail.ncsu.edu> Message-ID: On Oct 12, 2006, at 19:16 , Vivek Ranjan wrote: > [...] the programme does not print anything after it prints the > dielectric > constant. then i figured out from the forum that some ppl found their > programme crash if npool > 1. you cannot use more than one pool with one k-point > so, i also set npool to 1 and then one of the calculations went beyond > the calculation of dielectric constant, but the other still has the > problem > (i have two slightly different systems). please provide the test job that gives a problem Paolo From giannozz at nest.sns.it Thu Oct 12 22:42:01 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 12 Oct 2006 22:42:01 +0200 Subject: [Pw_forum] nscf In-Reply-To: References: Message-ID: On Oct 12, 2006, at 17:38 , H.S.Domingos wrote: > I would like to ask if using nscf is the same as using scf with > electron_maxstep = 1 > and startingpot = 'file from previous scf or vc-relax run' ? more or less, but the threshold used in the diagonalization is not the same in the two cases Paolo From giannozz at nest.sns.it Thu Oct 12 22:48:44 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 12 Oct 2006 22:48:44 +0200 Subject: [Pw_forum] about generating pseudopotential In-Reply-To: References: Message-ID: On Oct 12, 2006, at 11:00 , Amin Babazadeh wrote: > I want to use the function that Engel and Vosko proposed for > exchange-correlation potential {Exact exchange-only potential > and ...., > Phys.Rev B47 13164 (1993)} instead of PW91. > > I will appreciate if some one help me and tell me if it is possible > to change > exchange correlation potential in USPP or recommend me a code for > generating pseudopotential. I am not aware of any pseudopotential-generation or electronic-structure code using this EV-GGA. if this Engel-Vosko stuff is just a variant of GGA with a different functional form (I am too lazy to read the reference) its implementation in quantum-espresso is a straightforward but annoying exercise: see Modules/functionals.f90 Paolo From hsd22 at hermes.cam.ac.uk Fri Oct 13 14:17:15 2006 From: hsd22 at hermes.cam.ac.uk (H.S.Domingos) Date: Fri, 13 Oct 2006 13:17:15 +0100 (BST) Subject: [Pw_forum] nscf In-Reply-To: References: Message-ID: Dear All, I was wondering if there is a way of obtaining the Efermi in an expedite way and without doing an nscf calculation ? Best regards, Helder ======================================================================= | Dr. Helder S. Domingos | | | | INESC Microsyst & Nanotechnol, Lisbon, P-1000 Portugal | | and | | R&D unit for Molecular Chemical Physics | | Chemistry Department, University of Coimbra | ======================================================================= From babazade at gmail.com Fri Oct 13 14:25:08 2006 From: babazade at gmail.com (Amin Babazadeh) Date: Fri, 13 Oct 2006 15:55:08 +0330 Subject: [Pw_forum] about generating pseudopotential In-Reply-To: References: Message-ID: Dear Dr Paolo & Malgorzata Thanks a lot for your helps.I will try to do it. Sincerely yours Amin On 10/13/06, Paolo Giannozzi wrote: > > On Oct 12, 2006, at 11:00 , Amin Babazadeh wrote: > > > I want to use the function that Engel and Vosko proposed for > > exchange-correlation potential {Exact exchange-only potential > > and ...., > > Phys.Rev B47 13164 (1993)} instead of PW91. > > > > I will appreciate if some one help me and tell me if it is possible > > to change > > exchange correlation potential in USPP or recommend me a code for > > generating pseudopotential. > > I am not aware of any pseudopotential-generation or electronic-structure > code using this EV-GGA. if this Engel-Vosko stuff is just a variant > of GGA > with a different functional form (I am too lazy to read the > reference) its > implementation in quantum-espresso is a straightforward but annoying > exercise: see Modules/functionals.f90 > > Paolo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -- AMIN -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061013/66fbc2ac/attachment.htm From giannozz at nest.sns.it Fri Oct 13 14:31:28 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 13 Oct 2006 14:31:28 +0200 Subject: [Pw_forum] nscf In-Reply-To: References: Message-ID: <7D5AF1F1-AF52-400C-8B4E-B753D4EA3A65@nest.sns.it> On Oct 13, 2006, at 14:17 , H.S.Domingos wrote: > I was wondering if there is a way of obtaining the Efermi in an > expedite > way and without doing an nscf calculation ? Efermi is printed at the end of a scf calculation, if smearing or tetrahedra are used. Paolo From baroni at sissa.it Fri Oct 13 16:08:41 2006 From: baroni at sissa.it (Stefano Baroni) Date: Fri, 13 Oct 2006 16:08:41 +0200 Subject: [Pw_forum] nscf In-Reply-To: <7D5AF1F1-AF52-400C-8B4E-B753D4EA3A65@nest.sns.it> References: <7D5AF1F1-AF52-400C-8B4E-B753D4EA3A65@nest.sns.it> Message-ID: <6B6E84F9-F246-49E1-933A-D80E034B80CB@sissa.it> May I add that efermi is _obviously_ calculated in a SCF run. otherwise how could the code figure out which Bloch states to include in and which to exclude from the sum over occupied states needed to calculate everything (charge density, kinetic energy, etc.) ... take care - Stefano On Oct 13, 2006, at 2:31 PM, Paolo Giannozzi wrote: > > On Oct 13, 2006, at 14:17 , H.S.Domingos wrote: > >> I was wondering if there is a way of obtaining the Efermi in an >> expedite >> way and without doing an nscf calculation ? > > Efermi is printed at the end of a scf calculation, if smearing or > tetrahedra > are used. > > Paolo > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061013/f6f564df/attachment.htm From luch001 at 126.com Fri Oct 13 16:40:26 2006 From: luch001 at 126.com (luch001) Date: Fri, 13 Oct 2006 22:40:26 +0800 (CST) Subject: [Pw_forum] problem with pseudopotential Message-ID: <452FA55A.00009E.23748@bj126app12.126.com> Dear all, I am confused with some questions about the pseudopotential (PP). Above all, my research is mainly on the materials properties under high compression, including the phase transition under pressure. My questions are as following: I want to determine the stable structure by searching the global minimum of the ab initio free energy. Therefore, I should get reliable free energy with acceptable parameters (e.g., k-mesh and Ecut). In other words, the convergence of the PP should be good enough, namely a bit softer PP. But simultaneously I should guarantee the pseudo cores do not overlap too much, that is to say the rc should be small enough. This is paradox. How can I give attentions to the two things mentioned above ? Overlap to what extent is reasonable ? Any reply is appreciated! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061013/401f7e2b/attachment.htm From hsd22 at hermes.cam.ac.uk Fri Oct 13 17:39:57 2006 From: hsd22 at hermes.cam.ac.uk (H.S.Domingos) Date: Fri, 13 Oct 2006 16:39:57 +0100 (BST) Subject: [Pw_forum] nscf In-Reply-To: <6B6E84F9-F246-49E1-933A-D80E034B80CB@sissa.it> References: <7D5AF1F1-AF52-400C-8B4E-B753D4EA3A65@nest.sns.it> <6B6E84F9-F246-49E1-933A-D80E034B80CB@sissa.it> Message-ID: Dear All, I am sorry to insist on a naive question but I dont seem to be able to get the numerical value printed out for the fermi energy. How do you do that ? best regards, Helder ======================================================================= | Dr. Helder S. Domingos | | | | INESC Microsyst & Nanotechnol, Lisbon, P-1000 Portugal | | and | | R&D unit for Molecular Chemical Physics | | Chemistry Department, University of Coimbra | ======================================================================= From wmbmacam at lg.ehu.es Fri Oct 13 17:53:36 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Fri, 13 Oct 2006 17:53:36 +0200 Subject: [Pw_forum] nscf In-Reply-To: References: <7D5AF1F1-AF52-400C-8B4E-B753D4EA3A65@nest.sns.it ><6B6E84F9-F246-49E1-933A-D80E034B80CB@sissa.it> Message-ID: <452FB680.3090700@lg.ehu.es> Dear Helder, As far as I know, which is little, the Fermi energy is printed by default and, as Paolo pointed out, if you use tetrahedra or smearing as occupation method. You can find it just above the system energy. The command: $ grep Fermi your.system.out will yield the fermi energy. Keep in mind that it is also calculated on a non self-consistent run, so don't mistake them. > I am sorry to insist on a naive question but I dont seem to be able to > get the numerical value printed out for the fermi energy. How do you do > that ? -- ---------------------------------------- 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 ---------------------------------------- "The problem with Renault is that they dont have the skills to convince FIA that the mass damper is just a brake cooler." Annonymous From eyvaz_isaev at yahoo.com Fri Oct 13 18:02:05 2006 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 13 Oct 2006 09:02:05 -0700 (PDT) Subject: [Pw_forum] nscf In-Reply-To: Message-ID: <20061013160205.76165.qmail@web60313.mail.yahoo.com> Hi, grep Fermi *.scf.out But if you have a semiconductor or insulator without specifying an occupation method (and degauss, if a smearing technique is used) the fermi energy will not be printed, because in this case you (the program) know how many electron bands are occupied, N_valence_elec/2. In this case the highest filled band is assumed to be the Fermi level. In the case of the tetrahedra method used for semiconductors the Fermi level is placed in the band gap. Bests, Eyvaz. --- "H.S.Domingos" wrote: > Dear All, > > I am sorry to insist on a naive question but I dont > seem to be able to > get the numerical value printed out for the fermi > energy. How do you do > that ? > > > best regards, > > Helder > > > ======================================================================= > | Dr. Helder S. Domingos > | > | > | > | INESC Microsyst & Nanotechnol, Lisbon, P-1000 > Portugal | > | and > | > | R&D unit for Molecular Chemical Physics > | > | Chemistry Department, University of Coimbra > | > ======================================================================= > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From giannozz at nest.sns.it Fri Oct 13 18:33:58 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 13 Oct 2006 18:33:58 +0200 Subject: [Pw_forum] problem with pseudopotential In-Reply-To: <452FA55A.00009E.23748@bj126app12.126.com> References: <452FA55A.00009E.23748@bj126app12.126.com> Message-ID: On Oct 13, 2006, at 16:40 , luch001 wrote: > I want to determine the stable structure by searching the global > minimum of the > ab initio free energy. Therefore, I should get reliable free energy > with acceptable parameters (e.g., k-mesh and Ecut). In other words, > the convergence of the PP > should be good enough, namely a bit softer PP. But simultaneously I > should > guarantee the pseudo cores do not overlap too much, that is to say > the rc should > be small enough. This is paradox. it is not a paradox, it is life. You have to strike a compromise between two conflicting requirements. Anyway: I don't think that PP cores overlap, even at very high pressure, except for some large-core elements. If this happens, use smaller cores, or move semicore states into valence Paolo From luch001 at 126.com Sat Oct 14 03:22:28 2006 From: luch001 at 126.com (luch001) Date: Sat, 14 Oct 2006 09:22:28 +0800 (CST) Subject: [Pw_forum] problem with pseudopotential Message-ID: <45303BD4.0000FD.10365@bj126app6.126.com> Dear Paolo, Thank you very much for your patient help. I still have some questions: (1) If I download the pseudopotential (PP) from the http://www.pwscf.org, how can I know the detail of the PP (e.g., Rcut) ? Here is a example of the PP for Ge offered at the website, Generated using unknown code Author: X. Gonze et al Generation date: 1990 Info: Ge BHS modified for use as KB with l=2 as local part 1 The Pseudo was generated with a Scalar-Relativistic Calculation 0.00000000000E+00 Local Potential cutoff radius nl pn l occ Rcut Rcut US E pseu 4S 0 0 2.00 0.00000000000 0.00000000000 0.00000000000 4P 0 1 1.50 0.00000000000 0.00000000000 0.00000000000 My question is how much is Rcut ? (2) Take NaCl as an example, the overlap of the pseudo core is that the sum of the radius of Na and Cl is larger than the bond length between the Na and Cl. Is my comprehension of overlap right ? I have seen that the overlap do not too much somewhere. Does this mean a bit overlap is permitted ? Overlapping to what extent will lead to unreliable results ? Regards -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061014/c63cde49/attachment.htm From ezadshojaee at hotmail.com Sat Oct 14 10:37:20 2006 From: ezadshojaee at hotmail.com (Ezad Shojaee) Date: Sat, 14 Oct 2006 08:37:20 +0000 Subject: [Pw_forum] phonon calculation Message-ID: hi calculating phonon frequencies is a huge time-consuming process. i want to know that is it possible to stop the calculation & "continue" it after a while? if it is, what should i do to save the conclusions till know and how should i continue it? thanx _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ From ding at sissa.it Sat Oct 14 10:50:58 2006 From: ding at sissa.it (Xunlei Ding) Date: Sat, 14 Oct 2006 10:50:58 +0200 Subject: [Pw_forum] phonon calculation In-Reply-To: References: Message-ID: <4530A4F2.9060409@sissa.it> Maybe you just need to set recover = .true. in the input file when you continue a work from a previous one. Ding Ezad Shojaee wrote: > hi > calculating phonon frequencies is a huge time-consuming process. i > want to know that is it possible to stop the calculation & > "continue" it after a while? if it is, what should i do to save the > conclusions till know and how should i continue it? > thanx > > _________________________________________________________________ > Express yourself instantly with MSN Messenger! Download today it's > FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From shuixblue at yahoo.com.cn Sat Oct 14 11:46:22 2006 From: shuixblue at yahoo.com.cn (shangyi) Date: Sat, 14 Oct 2006 17:46:22 +0800 (CST) Subject: [Pw_forum] Help: the error of computaion single atom Message-ID: <20061014094622.95865.qmail@web15207.mail.cnb.yahoo.com> Dear all: I'm a newer of pwscf(3.1.1) and troubled by the error of computation single N atom energy! I computed the N atom without spin using the following input: ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ &control calculation = 'scf' restart_mode='from_scratch', prefix='N.atom', pseudo_dir = '/home/pub/pw/pseudo/', outdir='/home/shyma/tmp/' / &system ibrav= 1, celldm(1) =20, nat= 1, ntyp= 1, ecutwfc =40, / &electrons diagonalization='cg' mixing_mode = 'plain' mixing_beta = 0.7 conv_thr = 1.0d-8 / ATOMIC_SPECIES N 14.0067 N.pbe-rrkjus.UPF ATOMIC_POSITIONS N 0 0 0 K_POINTS AUTOMATIC 1 1 1 0 0 0 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ and it stop with the error: WARNING: integrated charge= 4.30000000, expected= 5.00000000 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from electrons : error # 1 charge is wrong %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% I have searched the maillist, but the information I got didn't solve my problem, so what's wrong with my input file and how to correct it? Another question is that I added the keyword "nspin=2, multiplicity=3" when I want to computate the N atom with spin, it stoped and displayed "some starting_magnetization MUST be set". The INPUT_PW says that "multiplicity fixes the final value of the magnetization". So is it need the starting_magnetization value when I have already set the multiplicity? If it's needed, what's the meaning of starting_magnetization and how ot set its value? Any advice is appreciated! Best wishes! shuix --------------------------------- Mp3???-??????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061014/b061e79b/attachment.htm From ding at sissa.it Sat Oct 14 12:28:10 2006 From: ding at sissa.it (Xunlei Ding) Date: Sat, 14 Oct 2006 12:28:10 +0200 Subject: [Pw_forum] Help: the error of computaion single atom In-Reply-To: <20061014094622.95865.qmail@web15207.mail.cnb.yahoo.com> References: <20061014094622.95865.qmail@web15207.mail.cnb.yahoo.com> Message-ID: <4530BBBA.9090805@sissa.it> Hi, If you are calculating for free N atom, I think you may have made some mistakes. The first question: N atom has 5 valence electrons. So it can't be calculated with spin=1, because it has at least one unpaired electrons. The second question: Total_magnatization= (number of unpaired electrons)/ (total number of electrons), and starting_magnetization is the initial value of it. For spin=2 calculation, the starting_magnetization must be set with a nonzero value. As for N, it has 5 valence electrons in which 3 electrons are unpaired. So it is reliable to set starting_magnetization(1)=3/5=0.6 . Actually other nonzero value will also work. And, for N atom, multiplicity=4, not 3. btw: maybe nosym is needed in calculation of only one atom. Yours, Ding shangyi wrote: > Dear all: > I'm a newer of pwscf(3.1.1) and troubled by the error of computation > single N atom energy! I computed the N atom without spin using the > following input: > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > &control > calculation = 'scf' > restart_mode='from_scratch', > prefix='N.atom', > pseudo_dir = '/home/pub/pw/pseudo/', > outdir='/home/shyma/tmp/' > / > &system > ibrav= 1, celldm(1) =20, nat= 1, ntyp= 1, > ecutwfc =40, > / > &electrons > diagonalization='cg' > mixing_mode = 'plain' > mixing_beta = 0.7 > conv_thr = 1.0d-8 > / > ATOMIC_SPECIES > N 14.0067 N.pbe-rrkjus.UPF > ATOMIC_POSITIONS > N 0 0 0 > K_POINTS AUTOMATIC > 1 1 1 0 0 0 > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > and it stop with the error: > WARNING: integrated charge= 4.30000000, expected= 5.00000000 > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from electrons : error # 1 > charge is wrong > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > I have searched the maillist, but the information I got didn't solve > my problem, so what's wrong with my input file and how to correct it? > Another question is that I added the keyword "nspin=2, multiplicity=3" > when I want to computate the N atom with spin, it stoped and displayed > "some starting_magnetization MUST be set". The INPUT_PW says that > "multiplicity fixes the final value of the magnetization". > So is it need the starting_magnetization value when I have already set > the multiplicity? If it's needed, what's the meaning of > starting_magnetization and how ot set its value? > Any advice is appreciated! > Best wishes! > shuix > > ------------------------------------------------------------------------ > Mp3???-??????? > From eyvaz_isaev at yahoo.com Sat Oct 14 16:16:46 2006 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Sat, 14 Oct 2006 07:16:46 -0700 (PDT) Subject: [Pw_forum] Help: the error of computaion single atom In-Reply-To: <4530BBBA.9090805@sissa.it> Message-ID: <20061014141646.19467.qmail@web60322.mail.yahoo.com> Hi, Sorry interfering, but it seems you forgot a very important keyword, "occupations=XXXX". Your system is considered as a semiconductor until you did not specify an occupations technique. You can use either tetrahedra or smearing technique with additional keywords. By the way the problem was discussed before and presumably it is a good idea to check the forum archive before submitting a question. Bests, Eyvaz. --- Xunlei Ding wrote: > Hi, > > If you are calculating for free N atom, I think you > may have made some > mistakes. > > The first question: > N atom has 5 valence electrons. So it can't be > calculated with spin=1, > because it has at least one unpaired electrons. > > The second question: > Total_magnatization= (number of unpaired electrons)/ > (total number of > electrons), and starting_magnetization is the > initial value of it. > For spin=2 calculation, the starting_magnetization > must be set with a > nonzero value. > As for N, it has 5 valence electrons in which 3 > electrons are unpaired. > So it is reliable to set > starting_magnetization(1)=3/5=0.6 . Actually > other nonzero value will also work. > And, for N atom, multiplicity=4, not 3. > > btw: maybe nosym is needed in calculation of only > one atom. > > Yours, > Ding > > shangyi wrote: > > > Dear all: > > I'm a newer of pwscf(3.1.1) and troubled by the > error of computation > > single N atom energy! I computed the N atom > without spin using the > > following input: > > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > &control > > calculation = 'scf' > > restart_mode='from_scratch', > > prefix='N.atom', > > pseudo_dir = '/home/pub/pw/pseudo/', > > outdir='/home/shyma/tmp/' > > / > > &system > > ibrav= 1, celldm(1) =20, nat= 1, ntyp= 1, > > ecutwfc =40, > > / > > &electrons > > diagonalization='cg' > > mixing_mode = 'plain' > > mixing_beta = 0.7 > > conv_thr = 1.0d-8 > > / > > ATOMIC_SPECIES > > N 14.0067 N.pbe-rrkjus.UPF > > ATOMIC_POSITIONS > > N 0 0 0 > > K_POINTS AUTOMATIC > > 1 1 1 0 0 0 > > > ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ > > and it stop with the error: > > WARNING: integrated charge= 4.30000000, expected= > 5.00000000 > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > from electrons : error # 1 > > charge is wrong > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > I have searched the maillist, but the information > I got didn't solve > > my problem, so what's wrong with my input file and > how to correct it? > > Another question is that I added the keyword > "nspin=2, multiplicity=3" > > when I want to computate the N atom with spin, it > stoped and displayed > > "some starting_magnetization MUST be set". The > INPUT_PW says that > > "multiplicity fixes the final value of the > magnetization". > > So is it need the starting_magnetization value > when I have already set > > the multiplicity? If it's needed, what's the > meaning of > > starting_magnetization and how ot set its value? > > Any advice is appreciated! > > Best wishes! > > shuix > > > > > ------------------------------------------------------------------------ > > Mp3??????-?????????????? > > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From luch001 at 126.com Sun Oct 15 04:09:19 2006 From: luch001 at 126.com (luch001) Date: Sun, 15 Oct 2006 10:09:19 +0800 (CST) Subject: [Pw_forum] electron-phonon coupling calculation on s upercell Message-ID: <4531984F.0000B4.13720@bj126app6.126.com> Dear all, I encounter difficulty in performing electron-phonon coupling (EPC) on a supercell. Take NaCl as a example, I construct a 2*2*2 supercell in the unit of the NaCl primitive cell, which has 16 atoms. But the supercell has a P1 space group. We all kown that the EPC is very time-consuming. Time to perform a EPC on such a supercell will be unbearable. What can I do to handle such a problem? My second question is how much the k-mesh should choose for such a supercell when do the EPC calculation ? Could someone give some empirical advice ? Any reply will be appreciated! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/74f2fe7b/attachment.htm From liyanpcl at yahoo.com.cn Sun Oct 15 04:51:11 2006 From: liyanpcl at yahoo.com.cn (li yan) Date: Sun, 15 Oct 2006 10:51:11 +0800 (CST) Subject: [Pw_forum] ultrasoft pseudopotentials for Cs and I Message-ID: <20061015025111.18801.qmail@web15615.mail.cnb.yahoo.com> deer all, Does anyone know where I can find the ultra-soft pseudopotentials for Cs and I ? best regards --------------------------------- ????????-3.5G???20M??? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/76233fed/attachment.htm From baroni at sissa.it Sun Oct 15 07:31:06 2006 From: baroni at sissa.it (Stefano Baroni) Date: Sun, 15 Oct 2006 07:31:06 +0200 Subject: [Pw_forum] ultrasoft pseudopotentials for Cs and I In-Reply-To: <20061015025111.18801.qmail@web15615.mail.cnb.yahoo.com> References: <20061015025111.18801.qmail@web15615.mail.cnb.yahoo.com> Message-ID: <04AB9E71-AD01-4474-A2EF-9653F412107A@sissa.it> norm-conserving PP's are rather soft in this case because the cores avery very large, even including semi-core states in the valence ... SB On Oct 15, 2006, at 4:51 AM, li yan wrote: > deer all, > Does anyone know where I can find the ultra-soft > pseudopotentials for Cs and I ? > best regards > > ????????-3.5G???20M??? --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/92e6c6e7/attachment.htm From luch001 at 126.com Sun Oct 15 12:12:01 2006 From: luch001 at 126.com (luch001) Date: Sun, 15 Oct 2006 18:12:01 +0800 (CST) Subject: [Pw_forum] electron phonon coupling of supercells Message-ID: <45320971.000006.09906@bj126app6.126.com> Dear all, I encounter difficulty in performing electron-phonon coupling (EPC) on a supercell. Take NaCl as a example, I construct a 2*2*2 supercell in the unit of the NaCl primitive cell, which has 16 atoms. But the supercell has a P1 space group. We all kown that the EPC is very time-consuming. Time to perform a EPC on such a supercell will be unbearable. What can I do to handle such a problem? My second question is how much the k-mesh should choose for such a supercell when do the EPC calculation ? Could someone give some empirical advice ? Any reply will be appreciated! -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/9de5c948/attachment.htm From marzari at MIT.EDU Sun Oct 15 12:20:58 2006 From: marzari at MIT.EDU (Nicola Marzari) Date: Sun, 15 Oct 2006 06:20:58 -0400 Subject: [Pw_forum] electron phonon coupling of supercells In-Reply-To: <45320971.000006.09906@bj126app6.126.com> References: <45320971.000006.09906@bj126app6.126.com> Message-ID: <45320B8A.10701@mit.edu> > I encounter difficulty in performing electron-phonon > coupling (EPC) on a supercell. Take NaCl as a example, > I construct a 2*2*2 supercell in the unit of the NaCl primitive > cell, which has 16 atoms. But the supercell has a P1 space > group. We all kown that the EPC is very time-consuming. Time to > perform a EPC on such a supercell will be unbearable. What can > I do to handle such a problem? Use the unit cell instead ? If you need the supercell, then there is no easy solution (apart from the ones gone below). > My second question is how much the k-mesh should choose > for such a supercell when do the EPC calculation ? Could someone > give some empirical advice ? 1) Start with a coarse sampling, and increase until it is converged. 2) Use Wannier functions as an interpolator: http://arxiv.org/abs/cond-mat/0507089 (also available in PRL) and especially http://arxiv.org/abs/cond-mat/0608257. There is more very pretty but yet unpublished work from the Berkeley group - wait a little for that to appear. 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 2586534 marzari at mit.edu http://quasiamore.mit.edu From luch001 at 126.com Sun Oct 15 16:17:22 2006 From: luch001 at 126.com (luch001) Date: Sun, 15 Oct 2006 22:17:22 +0800 (CST) Subject: [Pw_forum] pseudopotential problems Message-ID: <453242F2.000048.04512@bj126app11.126.com> Dear all, I encounter difficulties with pseudopotential (PP)(1)If I download the PP from the http://www.pwscf.org how can I know the detail of the PP (e.g., Rcut) ? Here is a example of the PP for Ge offered at the website (http://www.pwscf.org/pseudo.htm)Generated using unknown codeAuthor: X. Gonze et al Generation date: 1990Info: Ge BHS modified for use as KB with l=2 as local part 1 The Pseudo was generated with a Scalar-Relativistic Calculation 0.00000000000E+00 Local Potential cutoff radiusnl pn l occ Rcut Rcut US E pseu4S 0 0 2.00 0.00000000000 0.00000000000 0.000000000004P 0 1 1.50 0.00000000000 0.00000000000 0.00000000000My question is how much is Rcut ? (2)Take NaCl as an example, the overlap of the pseudo core is that the sum of the radius of Na and Cl is larger than the bond length between the Na and Cl. Is my comprehension of overlap right ? I have seen that the overlap do not too much somewhere. Does this mean a bit overlap is permitted ? Overlapping to what extent will lead to unreliable results ? Any reply will be appreciated. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/421e6e56/attachment.htm From luch001 at 126.com Sun Oct 15 16:53:32 2006 From: luch001 at 126.com (luch001) Date: Sun, 15 Oct 2006 22:53:32 +0800 (CST) Subject: [Pw_forum] electron phonon coupling of supercells Message-ID: <45324B6C.000139.07161@bj126app11.126.com> Dear all, Maybe I have not illustrate problem clearly. I want to perform EPC calculation on a supercell. Take NaCl as a example, I construct a 2*2*2 supercell in the unit of primitive cell, whose space group is P1 with the lowest symmetry. We all know that the EPC is a much time consuming process. Introducing symmetry operation to the system can speed the calculation greatly. My question is how can I introduce symmetry to the supercell I mentioned above in the precondition of obtaining physical results. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/a8c80ef6/attachment.htm From eyvaz_isaev at yahoo.com Sun Oct 15 19:03:40 2006 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Sun, 15 Oct 2006 10:03:40 -0700 (PDT) Subject: [Pw_forum] electron phonon coupling of supercells Message-ID: <20061015170340.82088.qmail@web60319.mail.yahoo.com> Hi, Maybe I have not illustrate problem clearly. I want to perform EPC calculation on a supercell. Take NaCl as a example, The example looks meaningless, NaCl is an insulator. You can try an Al supercell based on primitive unit cell.. I construct a 2*2*2 supercell in the unit of primitive cell, whose space group is P1 with the lowest symmetry. We all know that the EPC is a much time consuming process. Introducing symmetry operation to the system can speed the calculation greatly. My question is how can I introduce symmetry to the supercell I mentioned above in the precondition of obtaining physical results. If I understood correctly, you would like to know the space group of your 2x2x2 supercell and then use it in EPC calculations. There are several ways to do it. Say, VASP, CRYSTAL03 (but not sure about LMTO47, because it requires only irreducible atomic positions and space group number, too). ISOTROPY code seems to be very useful for this purpose: http://stokes.byu.edu/isotropy.html Hope it helps. Bests, Eyvaz. 3G ? ? ? ? ??? ? ? ? ? ? ? ? ? ? ? ? ? ? ? 3G ? ? ? ? ? ? ?280 ? ? ? ? ? ? ? ? ? ? ? ? ? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/d23b4ff8/attachment.htm From baroni at sissa.it Sun Oct 15 19:05:55 2006 From: baroni at sissa.it (Stefano Baroni) Date: Sun, 15 Oct 2006 19:05:55 +0200 Subject: [Pw_forum] electron phonon coupling of supercells In-Reply-To: <20061015170340.82088.qmail@web60319.mail.yahoo.com> References: <20061015170340.82088.qmail@web60319.mail.yahoo.com> Message-ID: <5D7BEE39-149D-4AB6-918E-C370949E172B@sissa.it> On Oct 15, 2006, at 7:03 PM, Eyvaz Isaev wrote: > Hi, > > Maybe I have not illustrate problem clearly. > I want to perform EPC calculation on a supercell. > Take NaCl as a example, > > The example looks meaningless, NaCl is an insulator. don't phonons interact with electrons in insulators? SB --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/708f1a63/attachment.htm From eyvaz_isaev at yahoo.com Sun Oct 15 20:31:48 2006 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Sun, 15 Oct 2006 11:31:48 -0700 (PDT) Subject: [Pw_forum] electron phonon coupling of supercells Message-ID: <20061015183148.28193.qmail@web60315.mail.yahoo.com> On Oct 15, 2006, at 7:03 PM, Eyvaz Isaev wrote: Hi, Maybe I have not illustrate problem clearly. I want to perform EPC calculation on a supercell. Take NaCl as a example, The example looks meaningless, NaCl is an insulator. don't phonons interact with electrons in insulators? Hi Stefano, Of course, they do, but it seemed to me that there was an intention to calculate \lambda. It might be that I missed something but I never hurd that NaCl is a superconductor . Just in this context I suggested that it is "meaningless". Bests, Eyvaz. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061015/acead636/attachment.htm From giannozz at nest.sns.it Sun Oct 15 23:06:56 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 15 Oct 2006 23:06:56 +0200 Subject: [Pw_forum] electron phonon coupling of supercells In-Reply-To: <45324B6C.000139.07161@bj126app11.126.com> References: <45324B6C.000139.07161@bj126app11.126.com> Message-ID: On Oct 15, 2006, at 16:53 , luch001 wrote: > I want to perform EPC calculation on a supercell. > Take NaCl as a example, I construct a 2*2*2 supercell > in the unit of primitive cell, whose space group is P1 > with the lowest symmetry. We all know that the EPC > is a much time consuming process. Introducing symmetry > operation to the system can speed the calculation greatly. point symmetry is automatically calculated and exploited, irrespective whether the unit cell is what one usually calls "a supercell", or not. Translational symmetry is dictated by the cell (or supercell). Of course the larger the cell, the larger the computer time needed. Paolo From giannozz at nest.sns.it Sun Oct 15 23:19:18 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 15 Oct 2006 23:19:18 +0200 Subject: [Pw_forum] pseudopotential problems In-Reply-To: <453242F2.000048.04512@bj126app11.126.com> References: <453242F2.000048.04512@bj126app11.126.com> Message-ID: On Oct 15, 2006, at 16:17 , luch001 wrote: > (1)If I download the PP from the http://www.pwscf.org how can I > know the detail > of the PP (e.g., Rcut) ? if it is not written there, look for references there . PPs by Xavier Gonze et al are described in some ~1990 obscure internal report that is quite hard to find (I think I have a paper copy somewhere). Maybe somebody in the list knows an easier to find reference. Paolo From giannozz at nest.sns.it Sun Oct 15 23:27:55 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 15 Oct 2006 23:27:55 +0200 Subject: [Pw_forum] Help: the error of computaion single atom In-Reply-To: <20061014094622.95865.qmail@web15207.mail.cnb.yahoo.com> References: <20061014094622.95865.qmail@web15207.mail.cnb.yahoo.com> Message-ID: <1DAED3A8-5BA4-4F2C-B16D-B269104155A0@nest.sns.it> On Oct 14, 2006, at 11:46 , shangyi wrote: > Another question is that I added the keyword "nspin=2, > multiplicity=3" when I > want to computate the N atom with spin, it stoped and displayed "some > starting_magnetization MUST be set". The INPUT_PW says that > "multiplicity > fixes the final value of the magnetization". there is an inconsistency between what the documentation says and what the code does. "multiplicity" and "tot_spin" are actually not working. This will be fixed sooner or later Paolo From giannozz at nest.sns.it Sun Oct 15 23:35:08 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 15 Oct 2006 23:35:08 +0200 Subject: [Pw_forum] phonon calculation In-Reply-To: References: Message-ID: On Oct 14, 2006, at 10:37 , Ezad Shojaee wrote: > i want to know that is it possible to stop the calculation & > "continue" it > after a while? if it is, what should i do to save the conclusions > till know > and how should i continue it? there is an option to stop the code after a prescribed amount of cpu time (max_seconds) and one to restart from an interrupted run (recover) Paolo From giannozz at nest.sns.it Sun Oct 15 23:42:03 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 15 Oct 2006 23:42:03 +0200 Subject: [Pw_forum] More pseudopotential troubles In-Reply-To: <452CC459.3020408@lg.ehu.es> References: <4523D7B5.9050609@lg.ehu.es> <452A1602.3080805@lg.ehu.es> <452CC459.3020408@lg.ehu.es> Message-ID: <56AA0834-AA06-443A-A895-950D75F4E186@nest.sns.it> On Oct 11, 2006, at 12:15 , Miguel Mart?nez Canales wrote: > Although your previous comments on Ca pseudopotential were very > helpful, > now I'm finding another problem. I may be wrong to think that the > input attached > generates an US pseudopotential with semicore states, using l=2 as the > pseudization channel and two reference energies for the s and p > channels. you should first try something simple before trying something difficult. A US pseudopotential is not easy to generate. The "atomic/" code uses Rabe-Rape-Kaxiras-Joannopoulos pseudization: very good when it works, but it is not easy to find a set of parameters for which it works. You can use Martins-Troullier pseudization, though. Moreover the "atomic/" code does not accept pseudostates with nodes, so you cannot use two bound states to pseudize if you have semicore states. You have to select a bound state, plus an unbound state having an energy that is not too far away from the energy of the nodeless (when pseudized) bound state Paolo From sodovar at gmail.com Mon Oct 16 15:03:18 2006 From: sodovar at gmail.com (Ruslan Minibaev) Date: Mon, 16 Oct 2006 17:03:18 +0400 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1182 - 7 msgs In-Reply-To: <20061012153902.30853.59247.Mailman@democritos.sissa.it> References: <20061012153902.30853.59247.Mailman@democritos.sissa.it> Message-ID: > > From: Paolo Giannozzi > Subject: Re: [Pw_forum] about projwfc.x > Date: Thu, 12 Oct 2006 11:12:14 +0200 > To: pw_forum at pwscf.org > Reply-To: pw_forum at pwscf.org > > On Oct 10, 2006, at 10:47 , Ruslan Minibaev wrote: > > > from read_restart_wfc : error # 1 > > Data Section not present in restart file > > > What's the reason of the error? > > it is written there : "Data Section not present in restart file" :-) > > Seriously: what version are you using? there is no such error message > in recent versions. > > Paolo Im using version 3.0 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061016/977d3f19/attachment.htm From wmbmacam at lg.ehu.es Mon Oct 16 15:25:32 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Mon, 16 Oct 2006 15:25:32 +0200 Subject: [Pw_forum] More pseudopotential troubles In-Reply-To: <56AA0834-AA06-443A-A895-950D75F4E186@nest.sns.it> References: <4523D7B5.9050609@lg.ehu.es> <452A1602.3080805@lg.ehu.es> <452CC459.3020408@lg.ehu.es> <56AA0834-AA06-443A-A895-950D75F4E186@nest.sns.it> Message-ID: <4533884C.7040508@lg.ehu.es> Dear Paolo, Thank you very much for your answer. It really helped my understanding of the atomic/ code. Anyway, I never really thought you could use Troulliers-Martins pseudization in US pseudopotentials. Paolo Giannozzi wrote: > > On Oct 11, 2006, at 12:15 , Miguel Mart?nez Canales wrote: > >> Although your previous comments on Ca pseudopotential were very helpful, >> now I'm finding another problem. I may be wrong to think that the >> input attached >> generates an US pseudopotential with semicore states, using l=2 as the >> pseudization channel and two reference energies for the s and p channels. > > you should first try something simple before trying something difficult. > A US pseudopotential is not easy to generate. The "atomic/" code > uses Rabe-Rape-Kaxiras-Joannopoulos pseudization: very good when > it works, but it is not easy to find a set of parameters for which it > works. > You can use Martins-Troullier pseudization, though. Moreover the > "atomic/" code does not accept pseudostates with nodes, so you cannot > use two bound states to pseudize if you have semicore states. You have > to select a bound state, plus an unbound state having an energy that is > not too far away from the energy of the nodeless (when pseudized) > bound state > > Paolo > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > From zucco at dipteris.unige.it Mon Oct 16 15:47:06 2006 From: zucco at dipteris.unige.it (Marino Vetuschi Zuccolini) Date: Mon, 16 Oct 2006 15:47:06 +0200 Subject: [Pw_forum] Masses in .dyn file Message-ID: <94f02882084230007aef7801d508f1b5@dipteris.unige.it> Hello to all, browsing the .dyn file about MgO, the header of the file is reported below the message. The numbers that in the code are written by ph.x (and read by others tools) after atomic symbols as atomic masses, evidently they are not, and change as the masses in &inputph section statement change. Are they related to reduced masses? I try to understand the meaning from the source but the result was not so good... Thanks in advance m. excerpt from MgO.dyn file Dynamical matrix file 2 2 2 7.9340000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 1 'Mg ' 22150.7811076348844 2 'O ' 14580.9646961962790 ... ******************************************************* 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 -- out of service 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: 1371 bytes Desc: not available Url : /pipermail/attachments/20061016/6232a4d3/attachment.bin From eyvaz_isaev at yahoo.com Mon Oct 16 16:11:05 2006 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Mon, 16 Oct 2006 07:11:05 -0700 (PDT) Subject: [Pw_forum] Masses in .dyn file In-Reply-To: <94f02882084230007aef7801d508f1b5@dipteris.unige.it> Message-ID: <20061016141105.48413.qmail@web60314.mail.yahoo.com> Hi Marino, > 2 2 2 7.9340000 0.0000000 0.0000000 > 0.0000000 0.0000000 > 0.0000000 > 1 'Mg ' 22150.7811076348844 > 2 'O ' 14580.9646961962790 > Atomic masses taken from matdyn.in file (which are from The Periodic Table) were converted to atomic units. Please find amconv parameter in matdyn.f90 to know what is a conversion unit. By the way did you find fhqa.f90 useful? Bests, Eyvaz. __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From giannozz at nest.sns.it Mon Oct 16 16:15:27 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 16 Oct 2006 16:15:27 +0200 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1182 - 7 msgs In-Reply-To: References: <20061012153902.30853.59247.Mailman@democritos.sissa.it> Message-ID: <271D6DF8-0169-4506-93F3-4435596A162F@nest.sns.it> On Oct 16, 2006, at 15:03 , Ruslan Minibaev wrote: > I'm using version 3.0 in parallel ? then maybe you forgot to save wavefunctions into a single directory (wf_collect=.true.) Paolo From katalin.gaal-nagy at physik.uni-regensburg.de Mon Oct 16 15:49:16 2006 From: katalin.gaal-nagy at physik.uni-regensburg.de (Katalin Gaal-Nagy) Date: Mon, 16 Oct 2006 15:49:16 +0200 (CEST) Subject: [Pw_forum] matdyn.x: write out complete dynamical matrices In-Reply-To: <4533884C.7040508@lg.ehu.es> References: <4523D7B5.9050609@lg.ehu.es> <452A1602.3080805@lg.ehu.es> <452CC459.3020408@lg.ehu.es> <56AA0834-AA06-443A-A895-950D75F4E186@nest.sns.it> <4533884C.7040508@lg.ehu.es> Message-ID: Dear all, I would like to modify the program matdyn.x in a way, that not only the frequencies are written, but also the dynamical matrices for all q in the star(q) like the program ph.x writes them to a file. As far as I looked through the routines, it seems to be sufficient to exchange the part where the routines dyndiag and writemodes are called to a call of the routine dynmatrix. Am I overlooking something? For example the calculation of the q in the star, or is this part done inside of the routine dynmatrix? Thanks for your help! All the best, Katalin From zucco at dipteris.unige.it Mon Oct 16 16:26:54 2006 From: zucco at dipteris.unige.it (Marino Vetuschi Zuccolini) Date: Mon, 16 Oct 2006 16:26:54 +0200 Subject: [Pw_forum] Masses in .dyn file In-Reply-To: <20061016141105.48413.qmail@web60314.mail.yahoo.com> References: <20061016141105.48413.qmail@web60314.mail.yahoo.com> Message-ID: Dear Eyvaz, as the question I posed reveal I'm walking thru a stony road, where for each answer you find at least one other question is at the horizon. To answer more directly I will use your code when I understand more or less clearly what I am doing. I continue with MgO exercise and step by step I will arrive to an end.. I suppose I contact you soon. Many thanks to you, and to all m. On 16 Oct 2006, at 4:11 PM, Eyvaz Isaev wrote: > Hi Marino, > >> 2 2 2 7.9340000 0.0000000 0.0000000 >> 0.0000000 0.0000000 >> 0.0000000 >> 1 'Mg ' 22150.7811076348844 >> 2 'O ' 14580.9646961962790 >> > Atomic masses taken from matdyn.in file (which are > from The Periodic Table) were converted to atomic > units. Please find amconv parameter in matdyn.f90 to > know what is a conversion unit. > > By the way did you find fhqa.f90 useful? > > Bests, > Eyvaz. > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > > ******************************************************* 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 -- out of service 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: 1766 bytes Desc: not available Url : /pipermail/attachments/20061016/5e0a3619/attachment.bin From shuixblue at yahoo.com.cn Tue Oct 17 14:14:09 2006 From: shuixblue at yahoo.com.cn (shangyi) Date: Tue, 17 Oct 2006 20:14:09 +0800 (CST) Subject: [Pw_forum] Re: Help: the error of computaion single atom In-Reply-To: <20061015053201.2415.49189.Mailman@democritos.sissa.it> Message-ID: <20061017121409.90531.qmail@web15209.mail.cnb.yahoo.com> Dear Ding and Eyvaz: Thanks for your advice! The calculation of free N atom with spin is OK with your help! And I test the calculation with and without some key-words "nosym and occupation": 1. without "nosym=.true. and occupation='tetrahedra'" the energy is -19.49051771 ryd 2. with "nosym=.true." the energy is -19.49051771 ryd 3. with "occupation='tetrahedra'" the energy is -19.49051770 ryd 4. with "nosym=.true. and occupation='tetrahedra'" the energy is -19.49051771 ryd so there is no difference among the energy results with or without those keywords! Maybe these key-words are optional in the atom calculation case. Thanks! Best wishes! --------------------------------- ????????-3.5G???20M??? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061017/7af5f481/attachment.htm From wmbmacam at lg.ehu.es Tue Oct 17 19:39:51 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Tue, 17 Oct 2006 19:39:51 +0200 Subject: [Pw_forum] More pseudopotential troubles In-Reply-To: <4533884C.7040508@lg.ehu.es> References: <4523D7B5.9050609@lg.ehu.es> <452A1602.3080805@lg.ehu.es> <452CC459.3020408@lg.ehu.es> <56AA0834-AA06-443A-A895-950D75F4E186@nest.sns.it> <4533884C.7040508@lg.ehu.es> Message-ID: <45351567.4030809@lg.ehu.es> Dear people, I'm now doing things step by step, so that I don't crash into the deep abyss of pseudopotentials, and I have found a new (for me) "error": ns= 3 l= 1 sum= 0.000034911 r(ikk) 1.899836304 from gener_pseudo : info # -1 chi too large beyond r_c plus two very long arrays of values I'll later put the input used to get this. The thing is: when I run said input, I don't get an error but neither do I get a pseudopotential. Examining atomic/gener_pseudo.f90 one finds (line 167 onwards, espresso 3.1.1): if (sum > 2.e-6_dp) then write(6, '(5x,''ns='',i4,'' l='',i4, '' sum='',f15.9, & & '' r(ikk) '',f15.9)') ns, lam, sum, r(ikk(ns)) call infomsg ('gener_pseudo ','chi too large beyond r_c', -1) do n=ikk(ns),mesh write(6,*) r(n),gi(n,1) enddo stop endif Seeing that it doesn't generate any pseudopotential, shouldn't it be "call errore(...)"? Please excuse my ignorance if this is something obvious. The input used to generate this was this: &input atom='Ca', beta = 0.2, rel=1, config='[Ar] 3d0 4s0 4p0', iswitch=3, dft='PBE' / &inputp pseudotype=3, lloc=2, nlcc =.true., rho0=0.01, file_pseudopw='Ca.pbe-van.UPF', / 4 3S 1 0 2.00 0.00 1.10 1.10 3P 2 1 6.00 0.00 1.30 1.60 3P 2 1 0.00 0.05 1.30 1.60 3D 3 2 0.00 0.00 1.80 1.80 I'd also be grateful is someone explained to me where the issue is or pointed me to a reference, as searching the archives for "chi too large" returns nothing useful. And changing rc's or rcutus' changes slightly the value of "sum" but it isn't enough. The same with pseudization energy. Thanks in advance, Miguel PS: Removing the second reference energy for the p channel and changing rcutus to 1.30 (thus generating a NC PP) does work... but the estimated cutoff is 110 rydbergs, which is a tad too high for running a phonon calculation. From rudrabnrj at gmail.com Wed Oct 18 07:17:24 2006 From: rudrabnrj at gmail.com (rudra banerjee) Date: Wed, 18 Oct 2006 10:47:24 +0530 Subject: [Pw_forum] unsubscribe me Message-ID: <2e36f8e50610172217m26a8b633pd88976fa99a276ba@mail.gmail.com> sir, as i have lost my password of this forum, i am unable u be benifited from this. so this is my request to the moderator to unsubscribe me from this group. hope you ll do that at soonest -- have a good time! rudra -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061018/4406c00b/attachment.htm From giannozz at nest.sns.it Wed Oct 18 09:25:22 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 18 Oct 2006 09:25:22 +0200 Subject: [Pw_forum] unsubscribe me In-Reply-To: <2e36f8e50610172217m26a8b633pd88976fa99a276ba@mail.gmail.com> References: <2e36f8e50610172217m26a8b633pd88976fa99a276ba@mail.gmail.com> Message-ID: <9DB1CB3C-5728-4B78-BCBE-0E3B6F9CFC65@nest.sns.it> On Oct 18, 2006, at 7:17 , rudra banerjee wrote: > as i have lost my password of this forum, i am unable u be benifited > from this. so this is my request to the moderator to unsubscribe me > from this group. hope you ll do that at soonest follow the 'subscribe' link in the pwscf web site. At the end of the page there are instructions on (among other things) how to recover your password P. From giannozz at nest.sns.it Wed Oct 18 10:07:29 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 18 Oct 2006 10:07:29 +0200 Subject: [Pw_forum] More pseudopotential troubles In-Reply-To: <45351567.4030809@lg.ehu.es> References: <4523D7B5.9050609@lg.ehu.es> <452A1602.3080805@lg.ehu.es> <452CC459.3020408@lg.ehu.es> <56AA0834-AA06-443A-A895-950D75F4E186@nest.sns.it> <4533884C.7040508@lg.ehu.es> <45351567.4030809@lg.ehu.es> Message-ID: <45A5B021-EC4C-4115-9D55-D6658ED5AA0A@nest.sns.it> On Oct 17, 2006, at 19:39 , Miguel Mart?nez Canales wrote: > ns= 3 l= 1 sum= 0.000034911 r(ikk) 1.899836304 > from gener_pseudo : info # -1 > chi too large beyond r_c this occasionally happens for no apparent good reason. I think it basically means that the PP generation failed. Try different electronic configuration / matching radii / pseudization algorithm / pseudization energies / phase of the moon Paolo From ezadshojaee at hotmail.com Wed Oct 18 13:10:16 2006 From: ezadshojaee at hotmail.com (Ezad Shojaee) Date: Wed, 18 Oct 2006 11:10:16 +0000 Subject: [Pw_forum] sufficient grid for phonon dispersion Message-ID: hi sorry if it is a silly question if we used a grid for calculating phonon dispersion and we find that it is'nt sufficient, what should we do? defining another grid ( with more points ) and performing for all of it again?! is there any way to add some new points information to the previous grid ? thanx _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ From ezadshojaee at hotmail.com Wed Oct 18 13:11:15 2006 From: ezadshojaee at hotmail.com (Ezad Shojaee) Date: Wed, 18 Oct 2006 11:11:15 +0000 Subject: [Pw_forum] sufficient grid for phonon dispersion Message-ID: hi sorry if it is a silly question if we used a grid for calculating phonon dispersion and we find that it is'nt sufficient, what should we do? defining another grid ( with more points ) and performing for all of it again?! is there any way to add some new points information to the previous grid ? thanx _________________________________________________________________ Express yourself instantly with MSN Messenger! Download today it's FREE! http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ From lanhaiping at gmail.com Wed Oct 18 15:05:47 2006 From: lanhaiping at gmail.com (lan haiping) Date: Wed, 18 Oct 2006 21:05:47 +0800 Subject: [Pw_forum] polarization calculation for system of surface states. Message-ID: Hi, All! I performed a calculation on a single layer of H1Ti3O7, i.e .a proton deficiency system, which would made system be metallic due to surface states. I wanna ask a stupid question: is it possible to use berryphase or DFPT method to examine the polarization of this system ? i had read some related publications, but didnot reach most idea of modern theory of polarization. when i run a ph.x calculation of this system , the programme complained about its metallic. Regards, hai-ping -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061018/144f5f8e/attachment.htm From lanhaiping at gmail.com Wed Oct 18 15:12:49 2006 From: lanhaiping at gmail.com (lan haiping) Date: Wed, 18 Oct 2006 21:12:49 +0800 Subject: [Pw_forum] sufficient grid for phonon dispersion In-Reply-To: References: Message-ID: hi, Ezad. some related problems have answered in this forum, you can try to find it. stefano had answered my silly question (i think almost the same) : "By and large, (squared) phonon frequencies are second derivatives of the crystal energy for a given kinetic-energy cutoff. If a cutoff is OK for geometry optimization, it is also so for phonon calculation (see below) " regards hai-ping On 10/18/06, Ezad Shojaee wrote: > > hi > sorry if it is a silly question > if we used a grid for calculating phonon dispersion and we find that it > is'nt sufficient, what should we do? defining another grid ( with more > points ) and performing for all of it again?! > is there any way to add some new points information to the previous grid ? > thanx > > _________________________________________________________________ > Express yourself instantly with MSN Messenger! Download today it's FREE! > http://messenger.msn.click-url.com/go/onm00200471ave/direct/01/ > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061018/b70c7b11/attachment.htm From giannozz at nest.sns.it Wed Oct 18 15:22:37 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 18 Oct 2006 15:22:37 +0200 Subject: [Pw_forum] sufficient grid for phonon dispersion In-Reply-To: References: Message-ID: <80E33678-E6C6-4DA0-95AA-A3E4F59075FB@nest.sns.it> On Oct 18, 2006, at 13:10 , Ezad Shojaee wrote: > if we used a grid for calculating phonon dispersion and we find > that it > isn't sufficient, what should we do? defining another grid ( with more > points ) and performing for all of it again?! is there any way to add > some new points information to the previous grid ? which grid are you talking about ? if it is the grid of phonon wavevectors used to calculate the interatomic forces constants in real space: in principle you could, but in practice it is not implemented. It shouldn't be that difficult: one has to skip calculations for wavevectors that were already calculated in the previous run Paolo From wmbmacam at lg.ehu.es Wed Oct 18 15:32:44 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Wed, 18 Oct 2006 15:32:44 +0200 Subject: [Pw_forum] sufficient grid for phonon dispersion In-Reply-To: References: Message-ID: <45362CFC.2020009@lg.ehu.es> Hi Ezad Shojaee wrote: > hi > sorry if it is a silly question > if we used a grid for calculating phonon dispersion and we find that it > is'nt sufficient, what should we do? defining another grid ( with more > points ) and performing for all of it again?! My probably naive way of doing it would be scf run -> nscf run -> phonon calculation at the desired q. Of course, with all the parameters being the same with the exception of the k-point sampling. After that, I would substitute the file containing the omegas and the eigenvectors with the new one. As you see, pretty na?ve. But faster than calculating every single point. From nguyenhalvt at yahoo.com Wed Oct 18 17:30:02 2006 From: nguyenhalvt at yahoo.com (Nguyen Ngoc Ha) Date: Wed, 18 Oct 2006 08:30:02 -0700 (PDT) Subject: [Pw_forum] Error diis diagonalization? Message-ID: <20061018153002.79368.qmail@web31007.mail.mud.yahoo.com> Dear PWscf Users! I used diis diagonalization to calculte singpoint with large cell, but result is a message: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from iosys : error # 1 diis diagonalization disabled %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Using cg or david diagonalization are well! Please help me the error! Best regards! Nguyen, Ngoc Ha Hanoi University of Education Faculty of Chemistry Department of Physical Chemistry Tel: Office: 04/8330842 Home: 04/7681083 Mobile: 0912129517 __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From giannozz at nest.sns.it Wed Oct 18 17:46:01 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 18 Oct 2006 17:46:01 +0200 Subject: [Pw_forum] Error diis diagonalization? In-Reply-To: <20061018153002.79368.qmail@web31007.mail.mud.yahoo.com> References: <20061018153002.79368.qmail@web31007.mail.mud.yahoo.com> Message-ID: <8BFC4374-E948-481B-990E-362074713606@nest.sns.it> On Oct 18, 2006, at 17:30 , Nguyen Ngoc Ha wrote: > from iosys : error # 1 > diis diagonalization disabled > Using cg or david diagonalization are well! so use them! it is easier than fixing the diis diagonaliser: an algorithm that has a great potential for speedup, but never worked in a satisfactory way. Paolo From lanhaiping at gmail.com Wed Oct 18 20:45:08 2006 From: lanhaiping at gmail.com (lan haiping) Date: Thu, 19 Oct 2006 02:45:08 +0800 Subject: [Pw_forum] Error diis diagonalization? In-Reply-To: <20061018153002.79368.qmail@web31007.mail.mud.yahoo.com> References: <20061018153002.79368.qmail@web31007.mail.mud.yahoo.com> Message-ID: you can find in INPUT_PW : 'diis' : DIIS-like diagonalization - PRESENTLY DISABLED regards, h.p On 10/18/06, Nguyen Ngoc Ha wrote: > > Dear PWscf Users! > I used diis diagonalization to calculte singpoint with > large cell, but result is a message: > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from iosys : error # 1 > diis diagonalization disabled > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > Using cg or david diagonalization are well! > Please help me the error! > Best regards! > > Nguyen, Ngoc Ha > Hanoi University of Education > Faculty of Chemistry > Department of Physical Chemistry > Tel: Office: 04/8330842 > Home: 04/7681083 > Mobile: 0912129517 > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/7946e0d9/attachment.htm From rudrabnrj at gmail.com Thu Oct 19 06:17:44 2006 From: rudrabnrj at gmail.com (rudra banerjee) Date: Thu, 19 Oct 2006 09:47:44 +0530 Subject: [Pw_forum] unsubscribe me In-Reply-To: <9DB1CB3C-5728-4B78-BCBE-0E3B6F9CFC65@nest.sns.it> References: <2e36f8e50610172217m26a8b633pd88976fa99a276ba@mail.gmail.com> <9DB1CB3C-5728-4B78-BCBE-0E3B6F9CFC65@nest.sns.it> Message-ID: <2e36f8e50610182117s6cc80540mbf02418fbf01a4a2@mail.gmail.com> thanks paolo On 10/18/06, Paolo Giannozzi wrote: > > > On Oct 18, 2006, at 7:17 , rudra banerjee wrote: > > > as i have lost my password of this forum, i am unable u be benifited > > from this. so this is my request to the moderator to unsubscribe me > > from this group. hope you ll do that at soonest > > follow the 'subscribe' link in the pwscf web site. At the end of the > page there are instructions on (among other things) how to recover > your password > > P. > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -- have a good time! rudra -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/7e2d1a3b/attachment.htm From baroni at sissa.it Thu Oct 19 08:35:08 2006 From: baroni at sissa.it (Stefano Baroni) Date: Thu, 19 Oct 2006 08:35:08 +0200 Subject: [Pw_forum] polarization calculation for system of surface states. In-Reply-To: References: Message-ID: <119CC9AB-668F-4748-89E9-0DA853548CBF@sissa.it> On Oct 18, 2006, at 3:05 PM, lan haiping wrote: > Hi, All! > > I performed a calculation on a single layer of H1Ti3O7, i.e .a > proton deficiency system, which would made system be metallic due > to surface states. I wanna ask a stupid question: is it possible > to use berryphase or DFPT method to examine the polarization of > this system ? let me try a studpid answer: 1) D=E+4\pi P 2) D=0 => E=-4\pi P 3) P /=0 => E /=0, which cannot be in the bulk of a conductor does the above make ansy sense? > i had read some related publications, but didnot reach most idea > of modern theory of polarization. > when i run a ph.x calculation of this system , the programme > complained about its metallic. the reason of the complains does not lie in any sophisticated "modern theory of polarization", but just in elementary (and old!) electrostatics.. Hope this helps - Stefano --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/a6a9cf22/attachment.htm From lanhaiping at gmail.com Thu Oct 19 11:17:30 2006 From: lanhaiping at gmail.com (lan haiping) Date: Thu, 19 Oct 2006 17:17:30 +0800 Subject: [Pw_forum] polarization calculation for system of surface states. In-Reply-To: <119CC9AB-668F-4748-89E9-0DA853548CBF@sissa.it> References: <119CC9AB-668F-4748-89E9-0DA853548CBF@sissa.it> Message-ID: thank you, stefano. my confusion may be due to this surface states. though there are some surface states , i thought it should be an insulator due to large bandgap. Is it mean that we should take systems with surface states as metal? regards, hai-ping On 10/19/06, Stefano Baroni wrote: > > > On Oct 18, 2006, at 3:05 PM, lan haiping wrote: > > Hi, All! > > I performed a calculation on a single layer of H1Ti3O7, i.e .a proton > deficiency system, which would made system be metallic due to surface > states. I wanna ask a stupid question: is it possible to use berryphase or > DFPT method to examine the polarization of this system ? > > > > > let me try a studpid answer: > 1) D=E+4\pi P > 2) D=0 => E=-4\pi P > 3) P /=0 => E /=0, which cannot be in the bulk of a conductor > > > does the above make ansy sense? > > > > i had read some related publications, but didnot reach most idea of > modern theory of polarization. > when i run a ph.x calculation of this system , the programme complained > about its metallic. > > > > > the reason of the complains does not lie in any sophisticated "modern > theory of polarization", but just in elementary (and old!) electrostatics.. > > > Hope this helps - Stefano > > --- > Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste > [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) > > Please, if possible, don't send me MS Word or PowerPoint attachments > Why? See: http://www.gnu.org/philosophy/no-word-attachments.html > > > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/5d9582f5/attachment.htm From sodovar at gmail.com Thu Oct 19 12:46:47 2006 From: sodovar at gmail.com (Ruslan Minibaev) Date: Thu, 19 Oct 2006 14:46:47 +0400 Subject: [Pw_forum] RE: about projwfc.x Message-ID: Dear, all >>* I'm using version 3.0 * >in parallel ? then maybe you forgot to save wavefunctions into a single >directory (wf_collect=.true.) >Paolo I put "wf_collect" in my input file and i get pdos.out: ** ** Program POST-PROC v.3.0 starts ... Today is 19Oct2006 at 13:20: 1 Parallel version (MPI) Number of processors in use: 27 R & G space division: proc/pool = 27 Reading file pwscf.save ... only dimensions read complete Reading file pwscf.save ... all except wavefuctions read complete Planes per process (thick) : nr3 =192 npp = 8 ncplane = 6400 Planes per process (smooth): nr3s=144 npps= 6 ncplanes= 3600 Proc/ planes cols G planes cols G columns G Pool (dense grid) (smooth grid) (wavefct grid) 1 8 162 19872 6 95 8923 26 1290 2 8 162 19872 6 95 8943 25 1281 3 8 162 19872 6 95 8939 25 1281 4 7 162 19872 6 95 8929 25 1281 5 7 162 19872 6 95 8923 25 1281 6 7 162 19872 6 95 8923 26 1288 7 7 162 19872 6 95 8917 26 1288 8 7 162 19872 6 94 8906 26 1288 9 7 162 19872 6 94 8906 26 1288 10 7 162 19872 5 94 8914 26 1288 11 7 162 19872 5 94 8918 26 1288 12 7 161 19863 5 95 8935 26 1288 13 7 162 19872 5 95 8939 25 1281 14 7 162 19872 5 95 8939 26 1290 15 7 162 19872 5 95 8939 26 1290 16 7 162 19872 5 95 8937 26 1290 17 7 162 19872 5 95 8939 26 1290 18 7 162 19872 5 95 8939 26 1286 19 7 162 19870 5 95 8937 26 1286 20 7 162 19870 5 95 8933 26 1286 21 7 162 19870 5 95 8929 26 1286 22 7 162 19870 5 95 8937 26 1286 23 7 162 19870 5 95 8935 26 1286 24 7 162 19870 5 95 8931 26 1288 25 7 162 19870 5 95 8931 26 1288 26 7 162 19870 5 95 8923 26 1288 27 7 162 19870 5 95 8921 26 1288 0 192 4373 536517 144 2561 241085 697 34739 nbndx = 386 nbnd = 386 natomwfc = 440 npwx = 1159 nelec = 644.00 nkb = 688 ngl = 9374 Check: negative/imaginary core charge= -0.000015 0.000000 Gaussian broadening (read from input): ngauss,degauss= 1 0.020000 Calling projwave .... ** ON ENTRY TO ZHPEV PARAMETER NUMBER 1 HAD AN ILLEGAL VALUE ** ** How to solve this problem? Best regards, Ruslan Minibaev -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/a054b7b0/attachment.htm From giannozz at nest.sns.it Thu Oct 19 13:14:31 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 19 Oct 2006 13:14:31 +0200 Subject: [Pw_forum] RE: about projwfc.x In-Reply-To: References: Message-ID: On Oct 19, 2006, at 12:46 , Ruslan Minibaev wrote: > ** ON ENTRY TO ZHPEV PARAMETER NUMBER 1 HAD AN ILLEGAL VALUE you are calling the essl version of zhpev with the calling sequence for the lapack version of zhpev (or vice versa). Load essl after lapack (or vice versa) Paolo From ashish.mhadeshwar at gmail.com Thu Oct 19 13:34:52 2006 From: ashish.mhadeshwar at gmail.com (Ashish Mhadeshwar) Date: Thu, 19 Oct 2006 07:34:52 -0400 Subject: [Pw_forum] unsubscribe me In-Reply-To: <2e36f8e50610182117s6cc80540mbf02418fbf01a4a2@mail.gmail.com> References: <2e36f8e50610172217m26a8b633pd88976fa99a276ba@mail.gmail.com> <9DB1CB3C-5728-4B78-BCBE-0E3B6F9CFC65@nest.sns.it> <2e36f8e50610182117s6cc80540mbf02418fbf01a4a2@mail.gmail.com> Message-ID: Kindly unsubscribe me as well. I am not using the software. Thank you. Ashish. On 10/19/06, rudra banerjee wrote: > > thanks paolo > > On 10/18/06, Paolo Giannozzi wrote: > > > > > > On Oct 18, 2006, at 7:17 , rudra banerjee wrote: > > > > > as i have lost my password of this forum, i am unable u be benifited > > > from this. so this is my request to the moderator to unsubscribe me > > > from this group. hope you ll do that at soonest > > > > follow the 'subscribe' link in the pwscf web site. At the end of the > > page there are instructions on (among other things) how to recover > > your password > > > > P. > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > -- > have a good time! > > rudra -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/8f8190f9/attachment.htm From giannozz at nest.sns.it Thu Oct 19 14:16:41 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 19 Oct 2006 14:16:41 +0200 Subject: [Pw_forum] unsubscribe me In-Reply-To: References: <2e36f8e50610172217m26a8b633pd88976fa99a276ba@mail.gmail.com> <9DB1CB3C-5728-4B78-BCBE-0E3B6F9CFC65@nest.sns.it> <2e36f8e50610182117s6cc80540mbf02418fbf01a4a2@mail.gmail.com> Message-ID: <9752A753-A82E-4210-84F1-6F3349E5C9B3@nest.sns.it> On Oct 19, 2006, at 13:34 , Ashish Mhadeshwar wrote: > Kindly unsubscribe me as well. I am not using the software. kindly follow the instructions From brandonw at MIT.EDU Thu Oct 19 19:04:10 2006 From: brandonw at MIT.EDU (Brandon Wood) Date: Thu, 19 Oct 2006 13:04:10 -0400 Subject: [Pw_forum] CP breakdown for long simulations? Message-ID: <4537B00A.9020604@mit.edu> Hi all, I'm running the CVS version of CP, and I'm wondering if anyone else has come across this problem. After ~10,000-15,000 timesteps of good data for a well-behaved, insulating system, the total integrated electronic charge density is suddenly no longer conserved and begins to fluctuate. As a result, the total energy drops rapidly and the simulation goes haywire. It seems like it is due to compounding of small numerical errors. Could there be a problem with the orthonormalization scheme that carries over from timestep to timestep but only manifests itself after sufficient time, for instance? Thanks, Brandon ------------------------------------- Brandon C. Wood Quasiamore Group Department of Materials Science Massachusetts Institute of Technology Office: 13-4078 Email: brandonw at mit.edu Phone: (617) 452-2455 ------------------------------------- From jess.kondor at gmail.com Thu Oct 19 23:39:24 2006 From: jess.kondor at gmail.com (Jess Kondor) Date: Thu, 19 Oct 2006 16:39:24 -0500 Subject: [Pw_forum] constraint variable cell relaxation Message-ID: <9d7293e10610191439le0ab4f8k4db12dfa930d0d38@mail.gmail.com> Dear All, Sorry, if this question was asked, but I did not find in the 'archive' the similar one. Is it possible to perform constraint variable cell relaxation? For example, to fix "a" and "b" parameters and relax ''c''? Thanks, JK -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/a4841d41/attachment.htm From marzari at MIT.EDU Fri Oct 20 00:34:26 2006 From: marzari at MIT.EDU (Nicola Marzari) Date: Thu, 19 Oct 2006 18:34:26 -0400 Subject: Fwd: [Pw_forum] CP breakdown for long simulations? In-Reply-To: <7368B4AC-538D-4BD6-B9A9-D6D1B72D2DA8@sns.it> References: <000901c6f3b7$207a5290$0300a8c0@int.cineca.it> <7368B4AC-538D-4BD6-B9A9-D6D1B72D2DA8@sns.it> Message-ID: <4537FD72.1070905@mit.edu> Thanks to Carlo Cavazzoni for the explanation ! The reason is that the lambda of the ortho constraints is calculated independently on each processor, and that can induce a numerical drift in different processors. The lambda is broadcast to all other processors every time there is a wavefunction write, but if one has a low frequency of restarts the drift can raise its ugly head. (Not to mention that our frequency of restarts was 100, that is quite frequent). Suggestion: since *especially* on massively-parallel run (like we do here - 4 cpus, and sometimes even 8 cpus) you do not want to save the wfcs very often, and since the ortho is very important, can't it be broadcast at every iteration, from the root processor to all others ? nicola >>>> Hi all, >>>> >>>> I'm running the CVS version of CP, and I'm wondering if anyone else >>>> has come across this problem. After ~10,000-15,000 timesteps of >>>> good data for a well-behaved, insulating system, the total >>>> integrated electronic charge density is suddenly no longer >>>> conserved and begins to fluctuate. As a result, the total energy >>>> drops rapidly and the simulation goes haywire. >>>> >>>> It seems like it is due to compounding of small numerical errors. >>>> Could there be a problem with the orthonormalization scheme that >>>> carries over from timestep to timestep but only manifests itself >>>> after sufficient time, for instance? >>>> >>>> Thanks, >>>> Brandon > -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu From ongphuongvu at yahoo.com Fri Oct 20 07:21:49 2006 From: ongphuongvu at yahoo.com (vu ongphuong) Date: Thu, 19 Oct 2006 22:21:49 -0700 (PDT) Subject: [Pw_forum] why total energy vs. lattice parameter is not a parabola? Message-ID: <20061020052149.15352.qmail@web50607.mail.yahoo.com> Dear all users, I calculate the lattice canstant of SrTiO3. When I plot the total energy vs. lattice parameter, the shape is not a parabola even the lattice parameter is taken in very near-minimum-region. The the lattice canstant obtained is 3.851Angstrong (the experimental value 3.905) When I calculate with VASP package the result is 3.86 and the total energy vs. lattice parameter is really a parabola. Anybody can explain me? --------------------------------- Do you Yahoo!? Everyone is raving about the all-new Yahoo! Mail. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061019/7e5a25f7/attachment.htm From baroni at sissa.it Fri Oct 20 10:57:12 2006 From: baroni at sissa.it (Stefano Baroni) Date: Fri, 20 Oct 2006 10:57:12 +0200 Subject: [Pw_forum] why total energy vs. lattice parameter is not a parabola? In-Reply-To: <20061020052149.15352.qmail@web50607.mail.yahoo.com> References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> Message-ID: <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> Dear Vu Ongphuong, the explanation goes back to very early days of plane-wave total energy calculations. The reason is, in order to have a consistent accuracy at different values of the lattice parameter, a0, a same kinetic-energy cutoff has to be used for all of them. The consequence is that the number of plane waves will depend on the lattice parameter. Being the number of plane waves an integer number, it will display jumps in correspondence to some values of a0. If the cutoff is large enough, the jumps will be small and the energy-vs-volume curve will look like a parabola. If the cutoff is not large enough _for the given set of pseudopotentials that you are using_, then the curve will rather look like a sawtooth ... Two fixes are possible: 1) use a larger cutoff (or a softer pseudopotential), or 2) use some phenomenological equation of state (such as Birtch's or Murnaghan's) to fit the sawtooth ... Stefano On Oct 20, 2006, at 7:21 AM, vu ongphuong wrote: > Dear all users, > > I calculate the lattice canstant of SrTiO3. When I plot the total > energy vs. lattice parameter, the shape is not a parabola even the > lattice parameter is taken in very near-minimum-region. The the > lattice canstant obtained is 3.851Angstrong (the experimental > value 3.905) > > When I calculate with VASP package the result is 3.86 and the total > energy vs. lattice parameter is really a parabola. > > Anybody can explain me? > > Do you Yahoo!? > Everyone is raving about the all-new Yahoo! Mail. --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061020/e6c4743e/attachment.htm From hsd22 at hermes.cam.ac.uk Fri Oct 20 14:53:30 2006 From: hsd22 at hermes.cam.ac.uk (H.S.Domingos) Date: Fri, 20 Oct 2006 13:53:30 +0100 (BST) Subject: [Pw_forum] why total energy vs. lattice parameter is not a parabola? In-Reply-To: <20061020052149.15352.qmail@web50607.mail.yahoo.com> References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> Message-ID: I would suggest that you test all the potentials that you are using individually. By, for example, making small bulks with just one element. After you have made certain that the convergence of the cutoffs is addequate and that the k-point density also converges every calculation, then you consider a strontium titanate cell with all well tested. I have had a similar problem with ZnO and it is essential to get potentials, cutoffs, k-points and geometries right. I would also warn you that strontium titanate has displacive transitions and that for very large bulks your results should be different since there are coordination polyhedra that will be rotated. Helder ======================================================================= | Dr. Helder S. Domingos | | | | INESC Microsyst & Nanotechnol, Lisbon, P-1000 Portugal | | and | | R&D unit for Molecular Chemical Physics | | Chemistry Department, University of Coimbra | ======================================================================= On Thu, 19 Oct 2006, vu ongphuong wrote: > Dear all users, > > I calculate the lattice canstant of SrTiO3. When I plot the total energy vs. lattice parameter, the shape is not a parabola even the lattice parameter is taken in very near-minimum-region. The the lattice canstant obtained is 3.851Angstrong (the experimental value 3.905) > > When I calculate with VASP package the result is 3.86 and the total energy vs. lattice parameter is really a parabola. > > Anybody can explain me? > > > --------------------------------- > Do you Yahoo!? > Everyone is raving about the all-new Yahoo! Mail. From baroni at sissa.it Fri Oct 20 15:43:06 2006 From: baroni at sissa.it (Stefano Baroni) Date: Fri, 20 Oct 2006 15:43:06 +0200 Subject: [Pw_forum] why total energy vs. lattice parameter is not a parabola? In-Reply-To: References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> Message-ID: On Oct 20, 2006, at 2:53 PM, H.S.Domingos wrote: > I would suggest that you test all the potentials that you are using > individually. By, for example, making small bulks with just one > element. > After you have made certain that the convergence of the cutoffs is > addequate and that the k-point density also converges every > calculation, then you consider a strontium titanate cell with all > well tested. Hi Helder: I think your suggestion is a sensible one, particularly in what concerns the choice of the cutoff (which is an atomic property, hence independent of the structure) vs. potential. As for k-point sampling, instead, it depends on the specific band structure, and so a k-point set that is adequate for a material may prove to be inadequate (or even meaningless, if the structure is totally different) for another ... S. > I have had a similar problem with ZnO and it is essential to get > potentials, cutoffs, k-points and geometries right. > > I would also warn you that strontium titanate has displacive > transitions and that for very large bulks your results should be > different since there are coordination polyhedra that will be rotated. > > > Helder > > > ====================================================================== > = > | Dr. Helder S. > Domingos | > | > | > | INESC Microsyst & Nanotechnol, Lisbon, P-1000 > Portugal | > | > and | > | R&D unit for Molecular Chemical > Physics | > | Chemistry Department, University of > Coimbra | > ====================================================================== > = > > > On Thu, 19 Oct 2006, vu ongphuong wrote: > >> Dear all users, >> >> I calculate the lattice canstant of SrTiO3. When I plot the >> total energy vs. lattice parameter, the shape is not a parabola >> even the lattice parameter is taken in very near-minimum-region. >> The the lattice canstant obtained is 3.851Angstrong (the >> experimental value 3.905) >> >> When I calculate with VASP package the result is 3.86 and the >> total energy vs. lattice parameter is really a parabola. >> >> Anybody can explain me? >> >> >> --------------------------------- >> Do you Yahoo!? >> Everyone is raving about the all-new Yahoo! Mail. > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061020/1aceb484/attachment.htm From wmbmacam at lg.ehu.es Fri Oct 20 16:06:52 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Fri, 20 Oct 2006 16:06:52 +0200 Subject: [Pw_forum] More pseudopotential troubles In-Reply-To: <45A5B021-EC4C-4115-9D55-D6658ED5AA0A@nest.sns.it> References: <4523D7B5.9050609@lg.ehu.es> <452A1602.3080805@lg.ehu.es> <452CC459.3020408@lg.ehu.es> <56AA0834-AA06-443A-A895-950D75F4E186@nest.sns.it> <4533884C.7040508@lg.ehu.es> <45351567.4030809@lg.ehu.es> <45A5B021-EC4C-4115-9D55-D6658ED5AA0A@nest.sns.it> Message-ID: <4538D7FC.8040506@lg.ehu.es> Hi again, Paolo Giannozzi wrote: > On Oct 17, 2006, at 19:39 , Miguel Mart?nez Canales wrote: > >> ns= 3 l= 1 sum= 0.000034911 r(ikk) 1.899836304 >> from gener_pseudo : info # -1 >> chi too large beyond r_c > > this occasionally happens for no apparent good reason. I think > it basically means that the PP generation failed. Try different > electronic configuration / matching radii / pseudization algorithm / > pseudization energies / phase of the moon OK, if I change the reference configuration to [Ar] 3d0 4s2 I will no longer get the "chi too large message"... but the transferability of the pseudopotential will be *really* awful. If I leave the config at 3d0 4s0 but add the "unscreening test" 3d0 4s2 it will stil fail. Let's say however that the p channel will only have one projector, pseudized using the RRKJ method. Now, with lloc = -1 and rcloc =1.5 the input will work. However, if I add an additional projector to the d channel, the "chi message" will reappear. Even if it is only an additional projector with rcutus = rcut. My questions are, then: * Do you have those auxiliary chi functions when you are not generating an USPP pseudopotential? * The chi message will disappear for the d channel at rcut~1.00. The RRKJ estimated cutoff is around 225 Ry. However, the estimated cutoff for rcutus = 1.70 is 45 Ry in the worst case. Will the high cutoff for rcut affect the pseudopotential and/or its convergence? Thanks very much in advance, Miguel PS: No matter how I change the pseudization energy, rcutus or lloc (as long as it isn't 1), the p channel will still give the "chi too large" message, at least for rcut > 0.80. If rcut > 1.4, the RRKJ pseudization will fail. And I find 90 Ry a bit too high for a future phonon calculation. From jiaanyan at gmail.com Fri Oct 20 16:45:30 2006 From: jiaanyan at gmail.com (Jia-An Yan) Date: Fri, 20 Oct 2006 10:45:30 -0400 Subject: [Pw_forum] questions about symmetry and the elphon output Message-ID: <6581de880610200745q5170899foac3ec8f5791d1fc3@mail.gmail.com> Hi everyone, I am calculating the el-ph interactions of graphene with Pwscf 3.1.1. There are two questions here: 1) First thing is about the symmetry obtained with pw.x. For the same graphene system, when I use ibrav=0 with cell parameters, I got only 8 symmetry opertions( with inversion). There are six modes in the phonon spectra. However, since the graphene system is hexagonal, I rotate the coordinate system and use ibrav=4 to define the system, then I got 24 sym operations (with inversion). There are also six modes (with 2 degenerate modes) obtained. Why does the pw.x obtained different symmetry for the same system? Can I adjust it manually? Here is my input: C.scf.fit.in: &control calculation='scf' restart_mode='from_scratch', prefix='C', pseudo_dir = '/dsgpfs/jayan/celph', outdir='/dsgpfs/jayan/scratch', wf_collect=.true. / &system ibrav= 0, celldm(1) =4.628,nat=2,ntyp= 1, ecutwfc =30.0,nr1=24,nr2=24, occupations='smearing', smearing='methfessel-paxton', degauss=0.02, la2F=.true., / &electrons conv_thr = 1.0d-10 mixing_beta = 0.7 / ATOMIC_SPECIES C 12.0107 C_nc ATOMIC_POSITIONS (alat) C 0.0000000000 0.0000000000 0.0000000000 C 0.5773502693 0.0000000000 0.0000000000 K_POINTS {automatic} 20 20 1 0 0 0 CELL_PARAMETERS 0.866025404 -0.500000000 0.000000000 0.000000000 1.000000000 0.000000000 0.000000000 0.000000000 6.000000000 2) The second question is about the output of the ph.x. I notice that the C.elph.out output following lines (2 2 1 1): There are 4 irreducible representations Representation 1 2 modes - To be done Representation 2 2 modes - To be done Representation 3 1 modes - To be done Representation 4 1 modes - To be done PHONON : 16.94s CPU time but at the end of the calculation, I got the omega as follows: omega( 1) = 16.131001 [THz] = 538.075842 [cm-1] omega( 2) = 16.131001 [THz] = 538.075842 [cm-1] omega( 3) = 30.116408 [THz] = 1004.581911 [cm-1] omega( 4) = 36.922309 [THz] = 1231.603836 [cm-1] omega( 5) = 36.922309 [THz] = 1231.603836 [cm-1] omega( 6) = 39.058597 [THz] = 1302.863199 [cm-1] Note that the order of the omega is different from the calculation order (2 1 2 1). Then I am quite confused at this, why are the output omega not the same order as the real calculation? Did the progam sort them after completion? I am interested in some data for each representation, so I need to get the omega for the corresponding representation. Thanks for your help in advance! Jia-An Yan -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061020/06224274/attachment.htm From gryko at jsu.edu Fri Oct 20 15:50:24 2006 From: gryko at jsu.edu (gryko at jsu.edu) Date: Fri, 20 Oct 2006 08:50:24 -0500 Subject: [Pw_forum] Compiling espresso-3.1.1 on AMD Athlon 4200/Linux-64 In-Reply-To: <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> Message-ID: <1161352224.4538d420da100@gem.jsu.edu> Dear PW users: I am trying to cmpile espresso-3.1.1 on AMD Athlon 4200 dual core, running Fedora-5 (64 bit version) using g95. The compilation stops with the following error: g95 -O3 -cpp -D__LINUX64 -D__G95 -D__FFTW -D__USE_INTERNAL_FFTW -I../include -I. -I../Modules -I../PW -I../PH -I../iotk/src -I../CPV -c iotk_scan.f90 In file iotk_scan.spp:412 control = modulo(header,iotk_ncontrol+1) 1 Error: 'p' argument of 'modulo' intrinsic at (1) must be the same type and kind as 'a' There are several posts on g95, but I still cannot solve this problem. Thank you for any hints or comments. Jan From ferretti.andrea at unimore.it Fri Oct 20 18:24:44 2006 From: ferretti.andrea at unimore.it (Andrea Ferretti) Date: Fri, 20 Oct 2006 18:24:44 +0200 (CEST) Subject: [Pw_forum] Compiling espresso-3.1.1 on AMD Athlon 4200/Linux-64 In-Reply-To: <1161352224.4538d420da100@gem.jsu.edu> References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> <1161352224.4538d420da100@gem.jsu.edu> Message-ID: this is a kind inconsistency in call to intrinsic functions; such problems have been fixed in the cvs version of espresso substituting lines like control = modulo(header,iotk_ncontrol+1) with control = modulo(header,int(iotk_ncontrol+1,kind=kind(header))) you can try to repeat the trick for all the lines where the compiler complains, otherwise try to you replace the espresso-3.1.1/iotk/ with espresso_CVS/iotk, it may probably work without any further problem.. andrea On Fri, 20 Oct 2006 gryko at jsu.edu wrote: > > Dear PW users: > > I am trying to cmpile espresso-3.1.1 on AMD Athlon 4200 dual core, > running Fedora-5 (64 bit version) using g95. The compilation stops > with the following error: > > g95 -O3 -cpp -D__LINUX64 -D__G95 -D__FFTW -D__USE_INTERNAL_FFTW -I../include > -I. -I../Modules -I../PW -I../PH -I../iotk/src -I../CPV -c iotk_scan.f90 > In file iotk_scan.spp:412 > > control = modulo(header,iotk_ncontrol+1) > 1 > Error: 'p' argument of 'modulo' intrinsic at (1) must be the same type and kind > as 'a' > > There are several posts on g95, but I still cannot solve this problem. > Thank you for any hints or comments. > > Jan > > _______________________________________________ > 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 20 18:37:36 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 20 Oct 2006 18:37:36 +0200 Subject: [Pw_forum] Compiling espresso-3.1.1 on AMD Athlon 4200/Linux-64 In-Reply-To: References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> <1161352224.4538d420da100@gem.jsu.edu> Message-ID: <88A82158-5D85-44E0-8140-DF8BB27848B9@nest.sns.it> On Oct 20, 2006, at 18:24 , Andrea Ferretti wrote: > such problems have been fixed in the cvs version of espresso > substituting lines like > > control = modulo(header,iotk_ncontrol+1) > > with > > control = modulo(header,int(iotk_ncontrol+1,kind=kind(header))) > > you can try to repeat the trick for all the lines where the compiler > complains there should be just two, in iotk_scan.spp. Then one has to regenerate *.f90 files: "make update" in iotk/ if I remember correctly Paolo From giannozz at nest.sns.it Fri Oct 20 18:52:13 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 20 Oct 2006 18:52:13 +0200 Subject: [Pw_forum] questions about symmetry and the elphon output In-Reply-To: <6581de880610200745q5170899foac3ec8f5791d1fc3@mail.gmail.com> References: <6581de880610200745q5170899foac3ec8f5791d1fc3@mail.gmail.com> Message-ID: <1DA3696F-4331-4614-BA1B-107E976B441F@nest.sns.it> On Oct 20, 2006, at 16:45 , Jia-An Yan wrote > For the same graphene system, when I use ibrav=0 with cell > parameters, > I got only 8 symmetry operations( with inversion) [...] However, > since the graphene system is hexagonal, I rotate the coordinate > system and use ibrav=4 to define the system, then I got 24 sym > operations (with inversion). There are also six modes (with 2 > degenerate modes) obtained. Why does the pw.x obtained different > symmetry for the same system? because you did not specify option "hexagonal" to card CELL_PARAMETERS, I guess > Note that the order of the omega is different from the calculation > order > (2 1 2 1) irreps used for the calculation are randomly generated; normal modes are ordered according to their frequency. There is no reason for them to follow the same ordering. Paolo From jiaanyan at gmail.com Fri Oct 20 21:58:55 2006 From: jiaanyan at gmail.com (Jia-An Yan) Date: Fri, 20 Oct 2006 15:58:55 -0400 Subject: [Pw_forum] questions about symmetry and the elphon output Message-ID: <6581de880610201258v7b95ca57o21719e8367277535@mail.gmail.com> Thanks, Paolo, I get it. Jia-An >On Oct 20, 2006, at 16:45 , Jia-An Yan wrote > >> For the same graphene system, when I use ibrav=0 with cell >> parameters, >> I got only 8 symmetry operations( with inversion) [...] However, >> since the graphene system is hexagonal, I rotate the coordinate >> system and use ibrav=4 to define the system, then I got 24 sym >> operations (with inversion). There are also six modes (with 2 >> degenerate modes) obtained. Why does the pw.x obtained different >> symmetry for the same system? > >because you did not specify option "hexagonal" to card >CELL_PARAMETERS, I guess > >> Note that the order of the omega is different from the calculation >> order >> (2 1 2 1) > >irreps used for the calculation are randomly generated; >normal modes are ordered according to their frequency. >There is no reason for them to follow the same ordering. > >Paolo > >_____________________________ __________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061020/feffceb3/attachment.htm From cesards at msi.umn.edu Fri Oct 20 23:29:28 2006 From: cesards at msi.umn.edu (cesards at msi.umn.edu) Date: Fri, 20 Oct 2006 16:29:28 -0500 (CDT) Subject: [Pw_forum] constraint variable cell relaxation Message-ID: <53470.128.101.191.226.1161379768.squirrel@www.msi.umn.edu> Dear Jess, To the extent of my knowledge, constrained cell dynamics was never developed, for either PR or Wentzcovitch lagrangians. Even if somebody did so, Pwscf does not implement it. With some programming you can do that in a non-formal way. At the end you will have to find a way to check for the validity of your results. Depending on the constraint you will have also to allow symmetries to break. Your, Cesar R.S. da Silva. > --__--__-- > > Message: 6 > Date: Thu, 19 Oct 2006 16:39:24 -0500 > From: "Jess Kondor" > To: pw_forum at pwscf.org > Subject: [Pw_forum] constraint variable cell relaxation > Reply-To: pw_forum at pwscf.org > > ------=_Part_83102_17565948.1161293964985 > Content-Type: text/plain; charset=ISO-8859-1; format=flowed > Content-Transfer-Encoding: 7bit > Content-Disposition: inline > > Dear All, > > Sorry, if this question was asked, but I did not find in the 'archive' > the > similar one. > > Is it possible to perform constraint variable cell relaxation? For > example, > to fix "a" and "b" parameters and relax ''c''? > > Thanks, > JK > From jess.kondor at gmail.com Sat Oct 21 00:11:18 2006 From: jess.kondor at gmail.com (Jess Kondor) Date: Fri, 20 Oct 2006 17:11:18 -0500 Subject: [Pw_forum] constraint variable cell relaxation In-Reply-To: References: <9d7293e10610200657t3615a69cy9d61baf0e491036e@mail.gmail.com> Message-ID: <9d7293e10610201511g1ddc4a10l3b9b12cf7701b97b@mail.gmail.com> Dear Cesar, Thanks for your point. In fact, I see the 'cp.x' can do it (keyword 'cell_dofree'). Might be a good point to extend functionality of PWscf. Sincerely, JK -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061020/d1ae8451/attachment.htm From proffess at yandex.ru Sat Oct 21 01:36:27 2006 From: proffess at yandex.ru (Sergey Lisenkov) Date: Sat, 21 Oct 2006 03:36:27 +0400 (MSD) Subject: [Pw_forum] error in para_zcholdc Message-ID: <45395D7B.000007.28532@webmail12.yandex.ru> Dear developers, I run the job using CVS version (19 October) of pwscf and got the following error: exit write_ns negative rho (up, down): 0.290E-02 0.290E-02 total cpu time spent up to now is 457.36 secs total energy = -3931.72473920 Ry Harris-Foulkes estimate = -3937.90523612 Ry estimated scf accuracy < 0.00004312 Ry total magnetization = 0.09 Bohr mag/cell absolute magnetization = 33.08 Bohr mag/cell iteration # 4 ecut= 30.00 Ry beta=0.10 Davidson diagonalization with overlap %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from para_zcholdc : error # 1385 a is not positive definite %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... Any hints? Best wishes, Sergey From nguyenhalvt at yahoo.com Sat Oct 21 12:28:28 2006 From: nguyenhalvt at yahoo.com (Nguyen Ngoc Ha) Date: Sat, 21 Oct 2006 03:28:28 -0700 (PDT) Subject: [Pw_forum] Error on calculation GS cp.x? Message-ID: <20061021102828.63617.qmail@web31005.mail.mud.yahoo.com> Dear PWscf Users! I used cp.x to calculation ground state firstly, here is my input: &control calculation='cp', restart_mode='from_scratch', nstep=100, iprint=20, isave=20, dt=10.0, pseudo_dir = '/home/hthp/Desktop/PWscf/PP_MgO/' , outdir='/home/hthp/tmp/', / &system ibrav = 14, A = 5.9555 , B = 5.9555 , C = 14.1056 , cosAB = 0.0 , cosAC = 0.0 , cosBC = 0.0 , nat = 18, ntyp = 3, ecutwfc = 25. , nr1b = 16, nr2b = 16, nr3b = 16 / &electrons orthogonalization = 'ortho', electron_dynamics = 'sd', / &ions ion_dynamics = 'none', ion_temperature = 'not_controlled', / ATOMIC_SPECIES Mg 24.30500 Mg.pw91-np-van.UPF O 15.99400 O.pw91-van_ak.UPF C 12.01 C.pw91-van_ak.UPF ATOMIC_POSITIONS alat Mg 0.00000 -0.00000 0.12309 0 0 0 Mg 0.00000 0.50000 0.12309 0 0 0 Mg 0.50000 0.00000 0.12309 0 0 0 Mg 0.50000 0.50000 0.12309 0 0 0 Mg 0.25000 0.25000 -0.00000 0 0 0 Mg 0.25000 0.75000 -0.00000 0 0 0 Mg 0.75000 0.25000 -0.00000 0 0 0 Mg 0.75000 0.75000 -0.00000 0 0 0 O 0.00000 -0.00000 -0.00000 0 0 0 O 0.00000 0.50000 -0.00000 0 0 0 O 0.50000 -0.00000 -0.00000 0 0 0 O 0.50000 0.50000 -0.00000 0 0 0 O 0.25000 0.25000 0.12309 0 0 0 O 0.25000 0.75000 0.12309 0 0 0 O 0.75000 0.25000 0.12309 0 0 0 O 0.75000 0.75000 0.12309 0 0 0 C -0.39221 -0.52212 0.23145 1 1 1 O -0.52206 -0.52200 0.28047 1 1 1 But I received error output: from rhoofr: total integrated electronic density in g-space = 122.000000 in r-space = 122.000000 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from dspev_drv : error # 60 diagonalization failed %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... I don't know what's problem! Please help me! Thanks very much for your help! Nguyen, Ngoc Ha Hanoi University of Education Faculty of Chemistry Department of Physical Chemistry Tel: Office: 04/8330842 Home: 04/7681083 Mobile: 0912129517 __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From giannozz at nest.sns.it Sat Oct 21 14:00:26 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sat, 21 Oct 2006 14:00:26 +0200 Subject: [Pw_forum] Error on calculation GS cp.x? In-Reply-To: <20061021102828.63617.qmail@web31005.mail.mud.yahoo.com> References: <20061021102828.63617.qmail@web31005.mail.mud.yahoo.com> Message-ID: On Oct 21, 2006, at 12:28 , Nguyen Ngoc Ha wrote: > I used cp.x to calculation ground state firstly [...] > > But I received error output: > [...] > from dspev_drv : error # 60 > diagonalization failed for the first steps you should either use a smaller time step or orthogonalization = 'Gram-Schmidt' Paolo From lanhaiping at gmail.com Sat Oct 21 20:31:50 2006 From: lanhaiping at gmail.com (lan haiping) Date: Sun, 22 Oct 2006 02:31:50 +0800 Subject: [Pw_forum] question about bands calculation Message-ID: Hi! i have a question about bands calculation . Just after finished a '*scf*' calculation for tio2, i performed a series (about 5) '*nscf*' calculations with different k-grid paths. Except k-grid path, all settings are the same for these series ' nscf ' calculation. In some nscf-output, the values of "Highest occupied level and lowest unoccuped level" are found. But these values are not found in the other nscf-output . I donot know what the reasons are related to . Morever, i also came to a calculation failed with complains "Too many eigenvalues are not converged". Regards, hai-ping -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061022/a506dc4c/attachment.htm From gryko at jsu.edu Sat Oct 21 23:33:12 2006 From: gryko at jsu.edu (gryko at jsu.edu) Date: Sat, 21 Oct 2006 16:33:12 -0500 Subject: [Pw_forum] Compiling espresso-3.1.1 on AMD Athlon 4200/Linux-64 In-Reply-To: References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> <1161352224.4538d420da100@gem.jsu.edu> Message-ID: <1161466392.453a9218c3d84@gem.jsu.edu> Hello: Thank you very much for your answers. Here what I did: 1. I've got the CVS version. The CVS version does not compile -it misses a number of LAPACK subroutines for some reason 2. I deleted iotk directory form the tar version of espresso-3.1.1 and substituted iotk directory from the CVS version 3. running configure and then make all produces executables. There were a lot of errors due to incompatible types, like Real*8 and complex*8, but no fatal errors 4. When trying to run examples, the execution stops at the second iteration: iteration # 1 ecut= 18.00 ryd beta=0.70 Davidson diagonalization with overlap ethr = 1.00E-02, avg # of iterations = 2.0 Threshold (ethr) on eigenvalues was too large: Diagonalizing with lowered threshold Davidson diagonalization with overlap ethr = 7.75E-03, avg # of iterations = 1.0 total cpu time spent up to now is 0.80 secs total energy = -15.84097415 ryd estimated scf accuracy < 0.06141563 ryd iteration # 2 ecut= 18.00 ryd beta=0.70 Davidson diagonalization with overlap ethr = 7.68E-04, avg # of iterations = 1.0 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from create_directory : error #4294967295 unable to create directory /home/gryko/tmp/silicon.save But directory /home/gryko/tmp/silicon.save is actually made in the first iteration! It looks that program attempts to make directory that already exists. Any suggestions? By the way, everything works on 32 bit version, Intel PC. Quoting Andrea Ferretti : > > this is a kind inconsistency in call to intrinsic functions; > > such problems have been fixed in the cvs version of espresso > substituting lines like > > control = modulo(header,iotk_ncontrol+1) > > with > > control = modulo(header,int(iotk_ncontrol+1,kind=kind(header))) > > > you can try to repeat the trick for all the lines where the compiler > complains, otherwise try to you replace the espresso-3.1.1/iotk/ with > espresso_CVS/iotk, it may probably work without any further problem.. > > andrea > From lukethulin at netscape.net Sun Oct 22 16:28:20 2006 From: lukethulin at netscape.net (lukethulin at netscape.net) Date: Sun, 22 Oct 2006 10:28:20 -0400 Subject: [Pw_forum] question about bands calculation References: Message-ID: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> Hai-ping, I don't think you're going about a band structure calculation correctly. The scf calculation should only use automatic k-point selection of (4,4,4) or something like that. The following nscf calculation needs you to input k-paths along high symmetry lines that vary by crystal structure. What form of titania are you modeling, anatase, rutile? Luke ----- Original Message ----- From: lan haiping To: pw_forum Sent: Saturday, October 21, 2006 2:31 PM Subject: [Pw_forum] question about bands calculation Hi! i have a question about bands calculation . Just after finished a 'scf' calculation for tio2, i performed a series (about 5) 'nscf' calculations with different k-grid paths. Except k-grid path, all settings are the same for these series ' nscf ' calculation. In some nscf-output, the values of "Highest occupied level and lowest unoccuped level" are found. But these values are not found in the other nscf-output . I donot know what the reasons are related to . Morever, i also came to a calculation failed with complains "Too many eigenvalues are not converged". Regards, hai-ping -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061022/d90a03f6/attachment.htm From giannozz at nest.sns.it Sun Oct 22 20:09:03 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 22 Oct 2006 20:09:03 +0200 Subject: [Pw_forum] question about bands calculation In-Reply-To: References: Message-ID: On Oct 21, 2006, at 20:31 , lan haiping wrote: > Just after finished a 'scf' calculation for tio2, i performed a > series (about 5) 'nscf' calculations with different k-grid paths. > Except k-grid path, all settings are the same for these series ' > nscf ' calculation. In some nscf-output, the values of "Highest > occupied level and lowest unoccuped level" are found. But these > values are not found in the other nscf-output . you should specify a number of bands larger than the number of of occupied states Paolo From giannozz at nest.sns.it Sun Oct 22 20:17:37 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 22 Oct 2006 20:17:37 +0200 Subject: [Pw_forum] Compiling espresso-3.1.1 on AMD Athlon 4200/Linux-64 In-Reply-To: <1161466392.453a9218c3d84@gem.jsu.edu> References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> <1161352224.4538d420da100@gem.jsu.edu> <1161466392.453a9218c3d84@gem.jsu.edu> Message-ID: <5CEF1C90-885E-4236-BEEA-A2371343E32D@nest.sns.it> On Oct 21, 2006, at 23:33 , gryko at jsu.edu wrote: > 1. I've got the CVS version. The CVS version does not compile - > it misses > a number of LAPACK subroutines for some reason "some reason" is not a sufficient reason. Which routines are missing? did you re-run "configure" ? you cannot re-use a make.sys file from a previous version, it is slightly different > 2. I deleted iotk directory form the tar version of espresso-3.1.1 > and substituted iotk directory from the CVS version it would have been easier to replace two lines, in my opinion > 3. running configure and then make all produces executables. > There were a lot of errors due to incompatible types, like > Real*8 and > complex*8 these should be irrelevant > from create_directory : error #4294967295 > unable to create directory /home/gryko/tmp/silicon.save unless it is a compiler bug, it is something going out of bounds > Any suggestions? use 32-bit integers. There is no advantage in using 64-bit integers. Paolo From giannozz at nest.sns.it Sun Oct 22 20:23:23 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 22 Oct 2006 20:23:23 +0200 Subject: [Pw_forum] error in para_zcholdc In-Reply-To: <45395D7B.000007.28532@webmail12.yandex.ru> References: <45395D7B.000007.28532@webmail12.yandex.ru> Message-ID: On Oct 21, 2006, at 1:36 , Sergey Lisenkov wrote: > from para_zcholdc : error # 1385 > a is not positive definite if a is REALLY not positive definite, there is little that can be done. If it is a weirdness of the new parallel algorithm for subspace diagonalization: disable it by commenting the call to routine check_para_diag_efficiency in PW/setup.f90 Paolo From lanhaiping at gmail.com Mon Oct 23 06:38:18 2006 From: lanhaiping at gmail.com (lan haiping) Date: Mon, 23 Oct 2006 12:38:18 +0800 Subject: [Pw_forum] question about bands calculation In-Reply-To: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> References: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> Message-ID: Dear Luke, thanks for your reply. my structure is anatase titania. Actually, i set the k-point paths according to high symmetry lines such as \Gamma[0,0,0] to Z [0,0,0.5] , Z to Q[0,0.5,0.5] etc. Is it right ? Regards, hai-ping On 10/22/06, lukethulin at netscape.net wrote: > > Hai-ping, > > I don't think you're going about a band structure calculation correctly. > The scf calculation should only use automatic k-point selection of (4,4,4) > or something like that. The following nscf calculation needs you to input > k-paths along high symmetry lines that vary by crystal structure. What form > of titania are you modeling, anatase, rutile? > > Luke > > ----- Original Message ----- > *From:* lan haiping > *To:* pw_forum > *Sent:* Saturday, October 21, 2006 2:31 PM > *Subject:* [Pw_forum] question about bands calculation > > > Hi! > > i have a question about bands calculation . > Just after finished a '*scf*' calculation for tio2, i performed a > series (about 5) '*nscf*' calculations with different k-grid paths. > Except k-grid path, all settings are the same for these series ' nscf ' > calculation. > In some nscf-output, the values of "Highest occupied level and lowest > unoccuped level" are found. But these values are not found in the other > nscf-output . I donot know what the reasons are related to . Morever, i also > came to a calculation failed with complains "Too many eigenvalues are not > converged". > > Regards, > hai-ping > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/733351ff/attachment.htm From lanhaiping at gmail.com Mon Oct 23 06:39:39 2006 From: lanhaiping at gmail.com (lan haiping) Date: Mon, 23 Oct 2006 12:39:39 +0800 Subject: [Pw_forum] question about bands calculation In-Reply-To: References: Message-ID: Dear Paolo, for anatase titinia i calculated, there are 48 electrons ,i.e 24 bands. i then specified 30 bands both in 'scf' and 'nscf' calculations. regards, hai-ping On 10/23/06, Paolo Giannozzi wrote: > > > On Oct 21, 2006, at 20:31 , lan haiping wrote: > > > Just after finished a 'scf' calculation for tio2, i performed a > > series (about 5) 'nscf' calculations with different k-grid paths. > > Except k-grid path, all settings are the same for these series ' > > nscf ' calculation. In some nscf-output, the values of "Highest > > occupied level and lowest unoccuped level" are found. But these > > values are not found in the other nscf-output . > > you should specify a number of bands larger than the number of > of occupied states > > Paolo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/50ce4116/attachment.htm From srijan.india at gmail.com Mon Oct 23 08:28:54 2006 From: srijan.india at gmail.com (Srijan Kumar Saha) Date: Mon, 23 Oct 2006 11:58:54 +0530 Subject: [Pw_forum] Re: from Srijan In-Reply-To: <3fcc7e670610222238v3d67a66dr2b6247a8d13bd053@mail.gmail.com> References: <57039.172.16.1.1.1161270463.squirrel@172.16.1.1> <39074.172.16.1.1.1161325253.squirrel@172.16.1.1> <3fcc7e670610202254p4444d95ep1ab03cd0863207c3@mail.gmail.com> <59699.172.16.1.1.1161441285.squirrel@172.16.1.1> <3fcc7e670610211140h1ed5c0t914804f08b476b6f@mail.gmail.com> <56262.172.16.1.1.1161535201.squirrel@172.16.1.1> <3fcc7e670610221013g60cd24b3o9e56c626b3bb1cf0@mail.gmail.com> <39848.172.16.1.1.1161538170.squirrel@172.16.1.1> <3fcc7e670610222238v3d67a66dr2b6247a8d13bd053@mail.gmail.com> Message-ID: <3fcc7e670610222328u2e2ab39clf795a632087dcfbf@mail.gmail.com> > My Dear Friends, > > LaMnO3 has almost cubic structure with. lattice parameter a = 0.54724nm > > In most of the place I have found the lattice parameter of LaMnO3 is > about 5.5 angstrom ===> around 10 Bohr. > > > I can bypass the problem and use smaller value of celldm (say 3 > angstrom) but I honestly > want to know why my program is converging for < 4.1 angstrom but not > for > 4.1 angstrom (even after 100 iterations)????????? Why 4.1angstrom is so sensitive??? Have fun > Srijan > > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/448a73a9/attachment.htm From lukethulin at netscape.net Mon Oct 23 14:03:57 2006 From: lukethulin at netscape.net (lukethulin at netscape.net) Date: Mon, 23 Oct 2006 08:03:57 -0400 Subject: [Pw_forum] question about bands calculation References: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> Message-ID: <003a01c6f69d$fb2c7150$6601a8c0@LukePC> hai-ping, I've successfully done band structure modeling for anatase, I'll go back and find what I chose for high symmetry lines, or find the reference I used. I guess I just thought it was strange that the nscf calculations were being done separately, when you could just pick several paths and do them all in one calculation. Luke ----- Original Message ----- From: lan haiping To: pw_forum at pwscf.org Sent: Monday, October 23, 2006 12:38 AM Subject: Re: [Pw_forum] question about bands calculation Dear Luke, thanks for your reply. my structure is anatase titania. Actually, i set the k-point paths according to high symmetry lines such as \Gamma[0,0,0] to Z [0,0,0.5] , Z to Q[0,0.5,0.5] etc. Is it right ? Regards, hai-ping On 10/22/06, lukethulin at netscape.net wrote: Hai-ping, I don't think you're going about a band structure calculation correctly. The scf calculation should only use automatic k-point selection of (4,4,4) or something like that. The following nscf calculation needs you to input k-paths along high symmetry lines that vary by crystal structure. What form of titania are you modeling, anatase, rutile? Luke ----- Original Message ----- From: lan haiping To: pw_forum Sent: Saturday, October 21, 2006 2:31 PM Subject: [Pw_forum] question about bands calculation Hi! i have a question about bands calculation . Just after finished a 'scf' calculation for tio2, i performed a series (about 5) 'nscf' calculations with different k-grid paths. Except k-grid path, all settings are the same for these series ' nscf ' calculation. In some nscf-output, the values of "Highest occupied level and lowest unoccuped level" are found. But these values are not found in the other nscf-output . I donot know what the reasons are related to . Morever, i also came to a calculation failed with complains "Too many eigenvalues are not converged". Regards, hai-ping -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/0df9cb06/attachment.htm From lukethulin at netscape.net Mon Oct 23 14:17:16 2006 From: lukethulin at netscape.net (lukethulin at netscape.net) Date: Mon, 23 Oct 2006 08:17:16 -0400 Subject: [Pw_forum] question about bands calculation References: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> Message-ID: <003b01c6f69d$fc5b5460$6601a8c0@LukePC> hai-ping, From lanhaiping at gmail.com Mon Oct 23 16:14:19 2006 From: lanhaiping at gmail.com (lan haiping) Date: Mon, 23 Oct 2006 22:14:19 +0800 Subject: [Pw_forum] question about bands calculation In-Reply-To: <003a01c6f69d$fb2c7150$6601a8c0@LukePC> References: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> <003a01c6f69d$fb2c7150$6601a8c0@LukePC> Message-ID: Dear Luke, thanks for you pacient explaination and suggest. i recalculated this anatase tio2 in succession again with input below, and found again there were no output about highest occupied and lowest unoccupied level in the 'bands' output. i just think it is so strange to understand, though i could calculate correctly-like bands structure. kpoints for 'scf' is using mp grid 8x8x8 without offsets.and kpoint path for 'bands' calculations is \Gamma[0,0,0] to X [0,0,0.5], from x[0,0,0.5] to [0.0.5,0.5], from [0.0, 0.5,0.5] to [0.5,0.5,0.5] ,from[0.5,0.5,0.5] to[0.0 ,0.0 ,0.5] Do you think any fault related to my kpoint path consideration,or any other parameters in the input file? &CONTROL title = 'TiO2' , calculation = 'bands', verbosity = 'default', tprnfor = .true., outdir = '/home/haiping/tmp/', prefix = 'tio2', disk_io = 'default', nstep = 200, pseudo_dir = '/home/haiping/espresso-3.0/pseudo/', etot_conv_thr= 1.0D-4 forc_conv_thr= 1.0D-4 / &SYSTEM ibrav= 7, celldm(1)=7.1526162323759346,celldm(3)=2.5136063408190221, nat = 6 , ntyp = 2 , nbnd = 30 , ecutwfc = 30.0000000000, ecutrho = 240.00 occupations = 'fixed' , / &ELECTRONS electron_maxstep = 100, conv_thr = 1.0e-12, mixing_mode = 'plain' , mixing_beta = 0.300000000, mixing_ndim = 8, diagonalization = 'cg' , / ATOMIC_SPECIES O 15.999 O.pbe-rrkjus.UPF Ti 47.867 Ti.pbe-sp-van_ak.UPF ATOMIC_POSITIONS angstrom Ti 0.000000000 -0.946250000 -1.189250000 Ti 0.000000000 0.946250000 1.189250000 O 0.000000000 0.946250000 -0.796426921 O 0.000000000 -0.946250000 0.796426921 O 1.892500000 0.946250000 1.582073079 O -1.892500000 -0.946250000 -1.582073079 K_POINTS crystal 41 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 0.05 0.5 1 0.0 0.10 0.5 1 0.0 0.15 0.5 1 0.0 0.20 0.5 1 0.0 0.25 0.5 1 0.0 0.30 0.5 1 0.0 0.35 0.5 1 0.0 0.40 0.5 1 0.0 0.45 0.5 1 0.0 0.50 0.5 1 0.05 0.50 0.5 1 0.10 0.50 0.5 1 0.15 0.50 0.5 1 0.20 0.50 0.50 1 0.25 0.50 0.50 1 0.30 0.50 0.50 1 0.35 0.50 0.50 1 0.40 0.50 0.50 1 0.45 0.50 0.50 1 0.50 0.50 0.50 1 0.467 0.467 0.50 1 0.433 0.433 0.50 1 0.400 0.400 0.50 1 0.367 0.367 0.50 1 0.333 0.333 0.50 1 0.30 0.30 0.50 1 0.267 0.267 0.50 1 0.233 0.233 0.50 1 0.20 0.20 0.50 1 0.167 0.167 0.50 1 0.133 0.133 0.50 1 0.100 0.100 0.50 1 0.067 0.067 0.50 1 0.033 0.033 0.50 1 0.00 0.00 0.50 1 Yours, hai-ping On 10/23/06, lukethulin at netscape.net wrote: > > hai-ping, > > I've successfully done band structure modeling for anatase, I'll go back > and find what I chose for high symmetry lines, or find the reference I > used. I guess I just thought it was strange that the nscf calculations were > being done separately, when you could just pick several paths and do them > all in one calculation. > > Luke > > ----- Original Message ----- > *From:* lan haiping > *To:* pw_forum at pwscf.org > *Sent:* Monday, October 23, 2006 12:38 AM > *Subject:* Re: [Pw_forum] question about bands calculation > > > Dear Luke, > thanks for your reply. my structure is anatase titania. > Actually, i set the k-point paths according to high symmetry lines such as > \Gamma[0,0,0] to Z [0,0,0.5] , Z to Q[0,0.5,0.5] etc. > Is it right ? > > Regards, > hai-ping > > > On 10/22/06, lukethulin at netscape.net wrote: > > > > Hai-ping, > > > > I don't think you're going about a band structure calculation > > correctly. The scf calculation should only use automatic k-point selection > > of (4,4,4) or something like that. The following nscf calculation needs you > > to input k-paths along high symmetry lines that vary by crystal structure. > > What form of titania are you modeling, anatase, rutile? > > > > Luke > > > > ----- Original Message ----- > > *From:* lan haiping > > *To:* pw_forum > > *Sent:* Saturday, October 21, 2006 2:31 PM > > *Subject:* [Pw_forum] question about bands calculation > > > > > > Hi! > > > > i have a question about bands calculation . > > Just after finished a '*scf*' calculation for tio2, i performed a > > series (about 5) '*nscf*' calculations with different k-grid paths. > > Except k-grid path, all settings are the same for these series ' nscf ' > > calculation. > > In some nscf-output, the values of "Highest occupied level and lowest > > unoccuped level" are found. But these values are not found in the other > > nscf-output . I donot know what the reasons are related to . Morever, i also > > came to a calculation failed with complains "Too many eigenvalues are not > > converged". > > > > Regards, > > hai-ping > > > > > > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/87c4ac06/attachment.htm From jiaanyan at gmail.com Mon Oct 23 16:32:53 2006 From: jiaanyan at gmail.com (Jia-An Yan) Date: Mon, 23 Oct 2006 10:32:53 -0400 Subject: [Pw_forum] problems in the el-ph/phonon calculations Message-ID: <6581de880610230732u49f98d18ne1342f254402b30b@mail.gmail.com> Hi everyone, I am performing an el-phonon/phonon calculation on a 2d hexagonal lattice. I use ibrav=0 to specify the structure. However, it seems that there are something wrong with the calculations. In the el-phonn/phonon output file, the q point list always misss some points. Although it is ok with the calculations, the next q2r part complains the following error message: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from init : error # 1 missing q-point(s)! %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% the two input files C.scf.in: -------------------------------------------------------------------- &control calculation='scf' restart_mode='from_scratch', prefix='c_ph', pseudo_dir = '/dsgpfs/jayan/celph', outdir='/dsgpfs/jayan/scratch', wf_collect=.true. / &system ibrav= 0, celldm(1) =4.628,nat=2,ntyp= 1, ecutwfc =30.0, occupations='smearing', smearing='methfessel-paxton', degauss=0.02, / &electrons conv_thr = 1.0d-10 mixing_beta = 0.7 / ATOMIC_SPECIES C 12.0107 C_nc ATOMIC_POSITIONS (alat) C 0.0000000000 0.0000000000 0.0000000000 C 0.5773502693 0.0000000000 0.0000000000 K_POINTS {automatic} 12 12 1 0 0 0 CELL_PARAMETERS {hexagonal} 0.866025404 -0.500000000 0.000000000 0.000000000 1.000000000 0.000000000 0.000000000 0.000000000 6.000000000 C.elph.in: -------------------------------------------------------------------- &inputph tr2_ph=1.0d-14, prefix='c_ph', amass(1)=12.0107, outdir='/dsgpfs/jayan/scratch', fildyn='c_ph.dyn', ldisp=.true., nq1=4,nq2=4,nq3=1 / 0.000 0.0000 0.00000 -------------------------------------------------------------------- The code is ok with example07. In addition, I also run nscf calculation separately for each q-point and generate the needed dyn files, when running the q2r.x, I got the following error: -------------------------------------------------------------------------------------------------------------- DCFT3 : 2538-2030 The transform length (ARG NO. 9) is not an allowed value. The next higher allowed value is (2). -------------------------------------------------------------------------------------------------------------- It seems that there are something wrong with the fft transform. Can anybody here give me some help? Your help is greatly appreciated! Jia-An Yan -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/bc288ef0/attachment.htm From yuhailin_79 at msn.com Mon Oct 23 16:57:41 2006 From: yuhailin_79 at msn.com (hailin yu) Date: Mon, 23 Oct 2006 22:57:41 +0800 Subject: [Pw_forum] help: phonon calculation of Ag Message-ID: Dear all: I have calculated the phonon dispersion of Ag and our results( see the attach) is well greement with a previous work( PRB 8,5532(1973)) except a lit difference: along the [110] direction, there is a cross point (up two curves) in the literature, while our results have no such a point ( alougth it's seem too). So we want to know our results is reliable? If not, why is the reason? Thanks in advance! yuhailin _________________________________________________________________ FREE pop-up blocking with the new MSN Toolbar - get it now! http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/ -------------- next part -------------- A non-text attachment was scrubbed... Name: Agphonondisp.BMP Type: image/bmp Size: 0 bytes Desc: not available Url : /pipermail/attachments/20061023/09bce83a/attachment.bin From yuhailin_79 at msn.com Mon Oct 23 16:58:08 2006 From: yuhailin_79 at msn.com (hailin yu) Date: Mon, 23 Oct 2006 22:58:08 +0800 Subject: [Pw_forum] help: phonon calculation of Ag Message-ID: Dear all: I have calculated the phonon dispersion of Ag and our results( see the attach) is well greement with a previous work( PRB 8,5532(1973)) except a lit difference: along the [110] direction, there is a cross point (up two curves) in the literature, while our results have no such a point ( alougth it's seem too). So we want to know our results is reliable? If not, why is the reason? Thanks in advance! yuhailin _________________________________________________________________ FREE pop-up blocking with the new MSN Toolbar - get it now! http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/ -------------- next part -------------- A non-text attachment was scrubbed... Name: Agphonondisp.BMP Type: image/bmp Size: 0 bytes Desc: not available Url : /pipermail/attachments/20061023/006bbc61/attachment.bin From yuhailin_79 at msn.com Mon Oct 23 16:58:29 2006 From: yuhailin_79 at msn.com (hailin yu) Date: Mon, 23 Oct 2006 22:58:29 +0800 Subject: [Pw_forum] help: phonon calculation of Ag Message-ID: Dear all: I have calculated the phonon dispersion of Ag and our results( see the attach) is well greement with a previous work( PRB 8,5532(1973)) except a lit difference: along the [110] direction, there is a cross point (up two curves) in the literature, while our results have no such a point ( alougth it's seem too). So we want to know our results is reliable? If not, why is the reason? Thanks in advance! yuhailin _________________________________________________________________ Don't just search. Find. Check out the new MSN Search! http://search.msn.com/ -------------- next part -------------- A non-text attachment was scrubbed... Name: Agphonondisp.BMP Type: image/bmp Size: 0 bytes Desc: not available Url : /pipermail/attachments/20061023/4759a793/attachment.bin From gryko at jsu.edu Mon Oct 23 16:41:14 2006 From: gryko at jsu.edu (gryko at jsu.edu) Date: Mon, 23 Oct 2006 09:41:14 -0500 Subject: [Pw_forum] Summary: Compiling espresso-3.1.1 on AMD Athlon 4200/Linux-64 In-Reply-To: <88A82158-5D85-44E0-8140-DF8BB27848B9@nest.sns.it> References: <20061020052149.15352.qmail@web50607.mail.yahoo.com> <33476023-AD91-4AA7-92AF-F5465F599D3C@sissa.it> <1161352224.4538d420da100@gem.jsu.edu> <88A82158-5D85-44E0-8140-DF8BB27848B9@nest.sns.it> Message-ID: <1161614474.453cd48ab33fd@gem.jsu.edu> Thanks to Paolo and Andrea - indeed, using g95 32 bit integer compiler helps and espresso-3.1.1 compiles correctly. In brief: 1. use Linux x86_64/EMT64 (32bit D.I) g95- compiler from g95 web site 2. correct one line in iotk/src/iotk_scan.spp as indicated by Paolo and Andrea (or just copy iotk directory from the CVS distribution...) 3. add ARFLAGS_DYNAMIC = ruv in make.sys From lukethulin at netscape.net Mon Oct 23 17:12:20 2006 From: lukethulin at netscape.net (lukethulin at netscape.net) Date: Mon, 23 Oct 2006 11:12:20 -0400 Subject: [Pw_forum] question about bands calculation References: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> <003a01c6f69d$fb2c7150$6601a8c0@LukePC> Message-ID: <006301c6f6b5$a2b78f10$6601a8c0@LukePC> hai-ping, I'm not sure, your input file looks pretty good to me. You can try this path that I've used in past. K_POINTS tpiba 23 0.000000000 0.000000000 0.400000000 1.000000000 0.000000000 0.000000000 0.350000000 1.000000000 0.000000000 0.000000000 0.300000000 1.000000000 0.000000000 0.000000000 0.250000000 1.000000000 0.000000000 0.000000000 0.200000000 1.000000000 0.000000000 0.000000000 0.150000000 1.000000000 0.000000000 0.000000000 0.100000000 1.000000000 0.000000000 0.000000000 0.050000000 1.000000000 0.000000000 0.000000000 0.000000000 1.000000000 0.050000000 0.050000000 0.000000000 1.000000000 0.100000000 0.100000000 0.000000000 1.000000000 0.150000000 0.150000000 0.000000000 1.000000000 0.200000000 0.200000000 0.000000000 1.000000000 0.250000000 0.250000000 0.000000000 1.000000000 0.300000000 0.300000000 0.000000000 1.000000000 0.350000000 0.350000000 0.000000000 1.000000000 0.400000000 0.400000000 0.000000000 1.000000000 0.450000000 0.450000000 0.000000000 1.000000000 0.500000000 0.500000000 0.000000000 1.000000000 0.500000000 0.500000000 0.050000000 1.000000000 0.500000000 0.500000000 0.100000000 1.000000000 0.500000000 0.500000000 0.150000000 1.000000000 0.500000000 0.500000000 0.200000000 1.000000000 My bands output file gave: Highest occupied, lowest unoccupied level (ev): 7.2352 9.3408 Luke ----- Original Message ----- From: lan haiping To: pw_forum at pwscf.org Sent: Monday, October 23, 2006 10:14 AM Subject: Re: [Pw_forum] question about bands calculation Dear Luke, thanks for you pacient explaination and suggest. i recalculated this anatase tio2 in succession again with input below, and found again there were no output about highest occupied and lowest unoccupied level in the 'bands' output. i just think it is so strange to understand, though i could calculate correctly-like bands structure. kpoints for 'scf' is using mp grid 8x8x8 without offsets.and kpoint path for 'bands' calculations is \Gamma[0,0,0] to X [0,0,0.5], from x[0,0,0.5] to [0.0.5,0.5], from [0.0, 0.5,0.5] to [0.5,0.5,0.5] ,from[0.5,0.5 ,0.5] to[0.0,0.0 ,0.5] Do you think any fault related to my kpoint path consideration,or any other parameters in the input file? &CONTROL title = 'TiO2' , calculation = 'bands', verbosity = 'default', tprnfor = .true., outdir = '/home/haiping/tmp/', prefix = 'tio2', disk_io = 'default', nstep = 200, pseudo_dir = '/home/haiping/espresso-3.0/pseudo/', etot_conv_thr= 1.0D-4 forc_conv_thr= 1.0D-4 / &SYSTEM ibrav= 7, celldm(1)=7.1526162323759346,celldm(3)=2.5136063408190221, nat = 6 , ntyp = 2 , nbnd = 30 , ecutwfc = 30.0000000000, ecutrho = 240.00 occupations = 'fixed' , / &ELECTRONS electron_maxstep = 100, conv_thr = 1.0e-12, mixing_mode = 'plain' , mixing_beta = 0.300000000, mixing_ndim = 8, diagonalization = 'cg' , / ATOMIC_SPECIES O 15.999 O.pbe-rrkjus.UPF Ti 47.867 Ti.pbe-sp-van_ak.UPF ATOMIC_POSITIONS angstrom Ti 0.000000000 -0.946250000 -1.189250000 Ti 0.000000000 0.946250000 1.189250000 O 0.000000000 0.946250000 -0.796426921 O 0.000000000 -0.946250000 0.796426921 O 1.892500000 0.946250000 1.582073079 O -1.892500000 -0.946250000 -1.582073079 K_POINTS crystal 41 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 0.05 0.5 1 0.0 0.10 0.5 1 0.0 0.15 0.5 1 0.0 0.20 0.5 1 0.0 0.25 0.5 1 0.0 0.30 0.5 1 0.0 0.35 0.5 1 0.0 0.40 0.5 1 0.0 0.45 0.5 1 0.0 0.50 0.5 1 0.05 0.50 0.5 1 0.10 0.50 0.5 1 0.15 0.50 0.5 1 0.20 0.50 0.50 1 0.25 0.50 0.50 1 0.30 0.50 0.50 1 0.35 0.50 0.50 1 0.40 0.50 0.50 1 0.45 0.50 0.50 1 0.50 0.50 0.50 1 0.467 0.467 0.50 1 0.433 0.433 0.50 1 0.400 0.400 0.50 1 0.367 0.367 0.50 1 0.333 0.333 0.50 1 0.30 0.30 0.50 1 0.267 0.267 0.50 1 0.233 0.233 0.50 1 0.20 0.20 0.50 1 0.167 0.167 0.50 1 0.133 0.133 0.50 1 0.100 0.100 0.50 1 0.067 0.067 0.50 1 0.033 0.033 0.50 1 0.00 0.00 0.50 1 Yours, hai-ping On 10/23/06, lukethulin at netscape.net wrote: hai-ping, I've successfully done band structure modeling for anatase, I'll go back and find what I chose for high symmetry lines, or find the reference I used. I guess I just thought it was strange that the nscf calculations were being done separately, when you could just pick several paths and do them all in one calculation. Luke ----- Original Message ----- From: lan haiping To: pw_forum at pwscf.org Sent: Monday, October 23, 2006 12:38 AM Subject: Re: [Pw_forum] question about bands calculation Dear Luke, thanks for your reply. my structure is anatase titania. Actually, i set the k-point paths according to high symmetry lines such as \Gamma[0,0,0] to Z [0,0,0.5] , Z to Q[0,0.5,0.5] etc. Is it right ? Regards, hai-ping On 10/22/06, lukethulin at netscape.net < lukethulin at netscape.net> wrote: Hai-ping, I don't think you're going about a band structure calculation correctly. The scf calculation should only use automatic k-point selection of (4,4,4) or something like that. The following nscf calculation needs you to input k-paths along high symmetry lines that vary by crystal structure. What form of titania are you modeling, anatase, rutile? Luke ----- Original Message ----- From: lan haiping To: pw_forum Sent: Saturday, October 21, 2006 2:31 PM Subject: [Pw_forum] question about bands calculation Hi! i have a question about bands calculation . Just after finished a 'scf' calculation for tio2, i performed a series (about 5) 'nscf' calculations with different k-grid paths. Except k-grid path, all settings are the same for these series ' nscf ' calculation. In some nscf-output, the values of "Highest occupied level and lowest unoccuped level" are found. But these values are not found in the other nscf-output . I donot know what the reasons are related to . Morever, i also came to a calculation failed with complains "Too many eigenvalues are not converged". Regards, hai-ping -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/698ccfa0/attachment.htm From sodovar at gmail.com Mon Oct 23 17:20:08 2006 From: sodovar at gmail.com (Ruslan Minibaev) Date: Mon, 23 Oct 2006 19:20:08 +0400 Subject: [Pw_forum] Sn USPP Message-ID: Dear all I need PBE US pseudopotential for Sn. I've tried to generate it with the code made by Vanderbilt but obtained bad results. Can anybody help me? Best regards Ruslan Minibaev -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/c666748f/attachment.htm From baroni at sissa.it Mon Oct 23 19:41:00 2006 From: baroni at sissa.it (Stefano Baroni) Date: Mon, 23 Oct 2006 19:41:00 +0200 Subject: [Pw_forum] Sn USPP In-Reply-To: References: Message-ID: Sn norm conserving PP's are soft enough not to need to be made "ultrasoft", I would say ... S. On Oct 23, 2006, at 5:20 PM, Ruslan Minibaev wrote: > Dear all > I need PBE US pseudopotential for Sn. > I've tried to generate it with the code made by Vanderbilt but > obtained bad results. > Can anybody help me? > > Best regards > Ruslan Minibaev --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061023/aa3306d1/attachment.htm From giannozz at nest.sns.it Tue Oct 24 10:05:12 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 24 Oct 2006 10:05:12 +0200 Subject: [Pw_forum] help: phonon calculation of Ag In-Reply-To: References: Message-ID: <47613941-6D35-47BC-95A5-EA2DA0BB1EFC@nest.sns.it> On Oct 23, 2006, at 16:57 , hailin yu wrote: > I have calculated the phonon dispersion of Ag and our results > ( see the attach) the attachment is empty. Please provide a URL (a pointer to a file on a web site) if you can, instead of sending attachments. > along the [110] direction, there is a cross point (up two curves) > in the literature, > while our results have no such a point ( alougth it's seem too). symmetry tells you whether two branches may cross or not. Otherwise, you have to look close to the crossing point with a dense q-point grid. There is no easy way to get a plot with correct crossing Paolo From giannozz at nest.sns.it Tue Oct 24 10:18:59 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 24 Oct 2006 10:18:59 +0200 Subject: [Pw_forum] problems in the el-ph/phonon calculations In-Reply-To: <6581de880610230732u49f98d18ne1342f254402b30b@mail.gmail.com> References: <6581de880610230732u49f98d18ne1342f254402b30b@mail.gmail.com> Message-ID: On Oct 23, 2006, at 16:32 , Jia-An Yan wrote: > The transform length (ARG NO. 9) is not an allowed value. > The next higher allowed value is (2). the code performs a FFT to bring force constants from reciprocal to real space (this is why it is called "q2r"). The FFTs in IBM ESSL libraries do not like to perform FFT with dimension 1 (or 3, or 5,...). The simplest soultion is to copy your dynamical matrices to another machine whose FFT's do not complain. Paolo From lanhaiping at gmail.com Tue Oct 24 15:13:03 2006 From: lanhaiping at gmail.com (lan haiping) Date: Tue, 24 Oct 2006 21:13:03 +0800 Subject: [Pw_forum] question about bands calculation In-Reply-To: <006301c6f6b5$a2b78f10$6601a8c0@LukePC> References: <000c01c6f5e6$534f7fc0$6501a8c0@Luke> <003a01c6f69d$fb2c7150$6601a8c0@LukePC> <006301c6f6b5$a2b78f10$6601a8c0@LukePC> Message-ID: thanks ,Luke. On 10/23/06, lukethulin at netscape.net wrote: > > hai-ping, > > I'm not sure, your input file looks pretty good to me. You can try this > path that I've used in past. > > K_POINTS tpiba > 23 > 0.000000000 0.000000000 0.400000000 1.000000000 > 0.000000000 0.000000000 0.350000000 1.000000000 > 0.000000000 0.000000000 0.300000000 1.000000000 > 0.000000000 0.000000000 0.250000000 1.000000000 > 0.000000000 0.000000000 0.200000000 1.000000000 > 0.000000000 0.000000000 0.150000000 1.000000000 > 0.000000000 0.000000000 0.100000000 1.000000000 > 0.000000000 0.000000000 0.050000000 1.000000000 > 0.000000000 0.000000000 0.000000000 1.000000000 > 0.050000000 0.050000000 0.000000000 1.000000000 > 0.100000000 0.100000000 0.000000000 1.000000000 > 0.150000000 0.150000000 0.000000000 1.000000000 > 0.200000000 0.200000000 0.000000000 1.000000000 > 0.250000000 0.250000000 0.000000000 1.000000000 > 0.300000000 0.300000000 0.000000000 1.000000000 > 0.350000000 0.350000000 0.000000000 1.000000000 > 0.400000000 0.400000000 0.000000000 1.000000000 > 0.450000000 0.450000000 0.000000000 1.000000000 > 0.500000000 0.500000000 0.000000000 1.000000000 > 0.500000000 0.500000000 0.050000000 1.000000000 > 0.500000000 0.500000000 0.100000000 1.000000000 > 0.500000000 0.500000000 0.150000000 1.000000000 > 0.500000000 0.500000000 0.200000000 1.000000000 > > My bands output file gave: > > > > Highest occupied, lowest unoccupied level (ev): 7.2352 9.3408 > > > > Luke > > ----- Original Message ----- > *From:* lan haiping > *To:* pw_forum at pwscf.org > *Sent:* Monday, October 23, 2006 10:14 AM > *Subject:* Re: [Pw_forum] question about bands calculation > > > Dear Luke, > thanks for you pacient explaination and suggest. > > i recalculated this anatase tio2 in succession again with input below, and > found again there were no output about highest occupied and lowest > unoccupied level in the 'bands' output. i just think it is so strange to > understand, though i could calculate correctly-like bands structure. > > > kpoints for 'scf' is using mp grid 8x8x8 without offsets.and kpoint path > for 'bands' calculations is \Gamma[0,0,0] to X [0,0,0.5], from x[0,0,0.5] > to [0.0.5,0.5], from [0.0, 0.5,0.5] to [0.5,0.5,0.5] ,from[0.5,0.5 ,0.5] > to[0.0,0.0 ,0.5] > Do you think any fault related to my kpoint path consideration,or any > other parameters in > the input file? > > > > &CONTROL > title = 'TiO2' , > calculation = 'bands', > verbosity = 'default', > tprnfor = .true., > outdir = '/home/haiping/tmp/', > prefix = 'tio2', > disk_io = 'default', > nstep = 200, > pseudo_dir = '/home/haiping/espresso-3.0/pseudo/', > etot_conv_thr= 1.0D-4 > forc_conv_thr= 1.0D-4 > / > > &SYSTEM > ibrav= 7, > celldm(1)=7.1526162323759346,celldm(3)=2.5136063408190221, > nat = 6 , > ntyp = 2 , > nbnd = 30 , > ecutwfc = 30.0000000000, > ecutrho = 240.00 > occupations = 'fixed' , > / > > &ELECTRONS > electron_maxstep = 100, > conv_thr = 1.0e-12, > mixing_mode = 'plain' , > mixing_beta = 0.300000000, > mixing_ndim = 8, > diagonalization = 'cg' , > / > ATOMIC_SPECIES > O 15.999 O.pbe-rrkjus.UPF > Ti 47.867 Ti.pbe-sp-van_ak.UPF > ATOMIC_POSITIONS angstrom > Ti 0.000000000 -0.946250000 -1.189250000 > Ti 0.000000000 0.946250000 1.189250000 > O 0.000000000 0.946250000 -0.796426921 > O 0.000000000 -0.946250000 0.796426921 > O 1.892500000 0.946250000 1.582073079 > O -1.892500000 -0.946250000 -1.582073079 > K_POINTS crystal > 41 > 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 0.05 0.5 1 > 0.0 0.10 0.5 1 > 0.0 0.15 0.5 1 > 0.0 0.20 0.5 1 > 0.0 0.25 0.5 1 > 0.0 0.30 0.5 1 > 0.0 0.35 0.5 1 > 0.0 0.40 0.5 1 > 0.0 0.45 0.5 1 > 0.0 0.50 0.5 1 > 0.05 0.50 0.5 1 > 0.10 0.50 0.5 1 > 0.15 0.50 0.5 1 > 0.20 0.50 0.50 1 > 0.25 0.50 0.50 1 > 0.30 0.50 0.50 1 > 0.35 0.50 0.50 1 > 0.40 0.50 0.50 1 > 0.45 0.50 0.50 1 > 0.50 0.50 0.50 1 > 0.467 0.467 0.50 1 > 0.433 0.433 0.50 1 > 0.400 0.400 0.50 1 > 0.367 0.367 0.50 1 > 0.333 0.333 0.50 1 > 0.30 0.30 0.50 1 > 0.267 0.267 0.50 1 > 0.233 0.233 0.50 1 > 0.20 0.20 0.50 1 > 0.167 0.167 0.50 1 > 0.133 0.133 0.50 1 > 0.100 0.100 0.50 1 > 0.067 0.067 0.50 1 > 0.033 0.033 0.50 1 > 0.00 0.00 0.50 1 > > > > Yours, > > hai-ping > > > > > On 10/23/06, lukethulin at netscape.net wrote: > > > > hai-ping, > > > > I've successfully done band structure modeling for anatase, I'll go back > > and find what I chose for high symmetry lines, or find the reference I > > used. I guess I just thought it was strange that the nscf calculations were > > being done separately, when you could just pick several paths and do them > > all in one calculation. > > > > Luke > > > > ----- Original Message ----- > > *From:* lan haiping > > *To:* pw_forum at pwscf.org > > *Sent:* Monday, October 23, 2006 12:38 AM > > *Subject:* Re: [Pw_forum] question about bands calculation > > > > > > Dear Luke, > > thanks for your reply. my structure is anatase titania. > > Actually, i set the k-point paths according to high symmetry lines such > > as \Gamma[0,0,0] to Z [0,0,0.5] , Z to Q[0,0.5,0.5] etc. > > Is it right ? > > > > Regards, > > hai-ping > > > > > > On 10/22/06, lukethulin at netscape.net < lukethulin at netscape.net> wrote: > > > > > > Hai-ping, > > > > > > I don't think you're going about a band structure calculation > > > correctly. The scf calculation should only use automatic k-point selection > > > of (4,4,4) or something like that. The following nscf calculation needs you > > > to input k-paths along high symmetry lines that vary by crystal structure. > > > What form of titania are you modeling, anatase, rutile? > > > > > > Luke > > > > > > ----- Original Message ----- > > > *From:* lan haiping > > > *To:* pw_forum > > > *Sent:* Saturday, October 21, 2006 2:31 PM > > > *Subject:* [Pw_forum] question about bands calculation > > > > > > > > > Hi! > > > > > > i have a question about bands calculation . > > > Just after finished a '*scf*' calculation for tio2, i performed a > > > series (about 5) '*nscf*' calculations with different k-grid paths. > > > Except k-grid path, all settings are the same for these series ' nscf ' > > > calculation. > > > In some nscf-output, the values of "Highest occupied level and lowest > > > unoccuped level" are found. But these values are not found in the other > > > nscf-output . I donot know what the reasons are related to . Morever, i also > > > came to a calculation failed with complains "Too many eigenvalues are not > > > converged". > > > > > > Regards, > > > hai-ping > > > > > > > > > > > > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061024/d56e3950/attachment.htm From murda at sissa.it Tue Oct 24 17:09:00 2006 From: murda at sissa.it (Garold Murdachaew) Date: Tue, 24 Oct 2006 17:09:00 +0200 (CEST) Subject: [Pw_forum] cube file from pp.x with orthogonal unit cell Message-ID: Dear pw_forum members, I am examining charge densities obtained from pw.x. I use the pp.x program to obtain densities in various formats. Although I am happy with the xsf format as provided by pp.x (I can use xcrysden for displaying the densities), a colleague wants to use gaussview for which the gaussian cube file is ideal. I notice that when I use xsf format, pp.x prints some coordinates with strange values (see below). Are these atoms in the next unit cell? There is another problem with use of cube files: pp.x does not allow one to change the axes so as to have the density within an orthogonal unit cell (my unit cell is hexagonal). I tried to use an input file of the form: &inputpp / &plot nfile = 1 filepp(1) = 'slab.rho' weight(1) = 1.0 iflag = 3 output_format = 6 fileout = 'slab_orthog.cube' e1(1)=0.75, e1(2)=0.00, e1(3)=0.00, e2(1)=0.00, e2(2)=0.75, e2(3)=0.00, e3(1)=0.00, e3(2)=0.00, e3(3)=6.00, x0(1)=0.00, x0(2)=0.00, x0(3)=0.00, nx=48, ny=48, nz=320 / but I gather that e_i cell vectors (and x0, n_i) are not allowed for the cube output_format=6. Since gaussview can only read in cube files with orthogonal axes, do I need to write a converter program to convert xsf (after orthogonalization) to cube format? Is there a better way (perhaps fitting and interpolating the density in some way)? Thank you for your time and assistance. Best, Garold Murdachaew -- Garold Murdachaew Postdoctoral Fellow - Condensed Matter Sector SISSA-ISAS (International School for Advanced Studies) New Building Room 208 Via Beirut 2-4, 34014 Trieste, Italy tel : +39 040 3787 477 fax : +39 040 3787 528 email : murda at sissa.it ps: Here are top parts of each density file for a clean Cu(111) slab, with three layers of atoms and 3 atoms per layer: native pw.x "rho" format, slab.rho: clean_slab.pw 48 48 320 48 48 320 9 1 0 9.50068671 0.00000000 0.00000000 0.00000000 0.00000000 0.00000000 0.750000000000000 0.433012701890000 0.000000000000000E+000 0.000000000000000E+000 0.866025403780000 0.000000000000000E+000 0.000000000000000E+000 0.000000000000000E+000 6.00000000000000 685.9169146890 10.0000000000 30.0000000000 0 1 Cu 11.00 1 0.500000000 0.577350269 0.408248290 1 2 0.000000000 0.577350269 0.408248290 1 3 0.250000000 0.144337567 0.408248290 1 4 0.250000000 0.721687836 0.000000000 1 5 0.500000000 0.288675135 0.000000000 1 6 0.000000000 0.288675135 0.000000000 1 7 0.250000000 0.433012702 -0.408248290 1 8 0.500000000 0.000000000 -0.408248290 1 9 0.000000000 0.000000000 -0.408248290 1 3.388479695E-02 3.429044967E-02 3.532584725E-02 3.674544602E-02 3.830175591E-02 3.979813932E-02 4.103542694E-02 4.186003670E-02 4.214093017E-02 4.186003772E-02 4.103542896E-02 3.979814230E-02 3.830175977E-02 3.674545064E-02 3.532585250E-02 . . . xsf format, slab.xsf: CRYSTAL PRIMVEC 3.770658668 2.176990797 0.000000000 0.000000000 4.353981594 0.000000000 0.000000000 0.000000000 30.165269347 PRIMCOORD 9 1 Cu 2.513772446 2.902654395 2.052486605 Cu 0.000000000 2.902654395 2.052486605 Cu 1.256886223 0.725663598 2.052486605 Cu 1.256886223 3.628317993 0.000000000 Cu 2.513772446 1.451327200 0.000000000 Cu 0.000000000 1.451327200 0.000000000 Cu 1.256886223 2.176990798 -2.052486605 Cu 2.513772446 0.000000000 -2.052486605 Cu 0.000000000 0.000000000 -2.052486605 BEGIN_BLOCK_DATAGRID_3D 3D_PWSCF DATAGRID_3D_UNKNOWN 49 49 321 0.000000 0.000000 0.000000 3.770659 2.176991 0.000000 0.000000 4.353982 0.000000 0.000000 0.000000 30.165269 0.33885E-01 0.34290E-01 0.35326E-01 0.36745E-01 0.38302E-01 0.39798E-01 0.41035E-01 0.41860E-01 0.42141E-01 0.41860E-01 0.41035E-01 0.39798E-01 0.38302E-01 0.36745E-01 0.35326E-01 0.34290E-01 0.33885E-01 0.34398E-01 . . . cube format with problem in coordinates, slabe.cube: Cubfile created from PWScf calculation Total SCF Density 9 0.000000 0.000000 0.000000 48 0.148448 0.085707 0.000000 48 0.000000 0.171413 0.000000 320 0.000000 0.000000 0.178138 29 29.000000 4.750343 5.485224 3.878639 29 29.000000 7.125515 9.599142 3.878639 29 29.000000 2.375172 9.599142 3.878639 29 29.000000 2.375172 6.856530 57.004120 29 29.000000 4.750343 2.742612 57.004120 29 29.000000 7.125515 6.856530 57.004120 29 29.000000 2.375172 4.113918 53.125481 29 29.000000 4.750343 8.227836 53.125481 29 29.000000 7.125515 12.341754 53.125481 0.33885E-01 0.33291E-01 0.32321E-01 0.31048E-01 0.29553E-01 0.27983E-01 0.26472E-01 0.25117E-01 0.24038E-01 0.23277E-01 0.22832E-01 0.22745E-01 0.23008E-01 0.23607E-01 0.24524E-01 0.25740E-01 0.27185E-01 0.28739E-01 . . . From giannozz at nest.sns.it Tue Oct 24 18:54:55 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 24 Oct 2006 18:54:55 +0200 Subject: [Pw_forum] cube file from pp.x with orthogonal unit cell In-Reply-To: References: Message-ID: On Oct 24, 2006, at 17:09 , Garold Murdachaew wrote: > I notice that when I use xsf format, pp.x prints some coordinates > with > strange values (see below). Are these atoms in the next unit cell? coordinates are given - in a0 (lattice parameter) units, in the internal pp.x format - in Angstrom in the xsf format - in Bohr radii in the cube format Paolo From vranjan at ncsu.edu Tue Oct 24 23:58:33 2006 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Tue, 24 Oct 2006 17:58:33 -0400 (EDT) Subject: [Pw_forum] bfgs and trust radius In-Reply-To: References: Message-ID: <49340.152.14.74.117.1161727113.squirrel@webmail.ncsu.edu> Hi ! I getting the following message : --------------------------------------------------------------------------- number of scf cycles = 14 number of bfgs steps = 1 energy old = -1402.7124539122 ryd energy new = -1402.7077099529 ryd CASE: energy_new > energy_old new trust radius = 0.0000480606 bohr trust_radius < trust_radius_min resetting bfgs history %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from bfgs : error # 1 bfgs history already reset at previous step %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... --------------------------------------------------------------------------------- Based on earlier discussions I feel the following are important parameters, which I have set as : etot_conv_thr = 1.0d-4 forc_conv_thr = 1.0d-5 &system ibrav= 0, nat= 31, ntyp= 6, celldm(1) = 7.2225240402, ecutwfc = 35.0, ecutrho = 420.0 occupations='smearing', smearing='mv', degauss=0.05 nosym=.true. / &electrons mixing_mode = 'plain' mixing_beta = 0.5 conv_thr = 1.0d-8 diagonalization = 'david' , / &ions upscale=100 / ------------------------------------------------------------------------- I have faced the above problem several times. But changing one thing or the another may result in convergence. The above generally happens when the system is already close to convergence. I would like to know if it is a good idea to reduce the trust_radius_min below its present default value set in pwscf ? Thank you, Vivek From vranjan at ncsu.edu Wed Oct 25 00:22:43 2006 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Tue, 24 Oct 2006 18:22:43 -0400 (EDT) Subject: [Pw_forum] problems with integrated charge In-Reply-To: References: Message-ID: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> Hi ! I am trying to relax a structure. Within 3 scf steps the program aborts with the following : ---------------------------------------------------------------------------- WARNING: integrated charge= 207.40294106, expected= 208.00000000 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from electrons : error # 1 charge is wrong %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% ------------------------------------------------------------------------------ I have tried the following : occupations='smearing', smearing='mv', degauss=0.5 and changed the degauss value from 0.01 initially to 0.05, then 0.1, and now 0.5. The integrated charge remains exactly the same at 207.40294106,when it expects it to be 208. The program aborts always in the 3rd scf iteration. What could be the reason ? Thank you, Vivek From wyanchao at gmail.com Wed Oct 25 06:02:03 2006 From: wyanchao at gmail.com (wang yanchao) Date: Tue, 24 Oct 2006 21:02:03 -0700 Subject: [Pw_forum] wang yanchao wants to chat Message-ID: <55639d30610242102w1feecdd6l@mail.gmail.com> ----------------------------------------------------------------------- wang yanchao wants to stay in better touch using some of Google's coolest new products. If you already have Gmail or Google Talk, visit: http://mail.google.com/mail/b-40feaf067b-e61a66ba1f-3b44e0bb0862e065 You'll need to click this link to be able to chat with wang yanchao. To get Gmail - a free email account from Google with over 2,600 megabytes of storage - and chat with wang yanchao, visit: http://mail.google.com/mail/a-40feaf067b-e61a66ba1f-f079d4d409 Gmail offers: - Powerful spam protection - Built-in search for finding your messages and a helpful way of organizing emails into "conversations" - No pop-up ads or untargeted banners - just text ads and related information that are relevant to the content of your messages - Instant messaging capabilities right inside Gmail All this, and its yours for free. But wait, there's more! You can also get Google Talk: http://www.google.com/talk/ Its a small Windows* download that lets you make free calls to your friends through your computer. It's simple and clutter-free, and it works with any computer speaker and microphone. Gmail and Google Talk are still in beta. We're working hard to add new features and make improvements, so we might also ask for your comments and suggestions periodically. We appreciate your help in making our products even better! Thanks, The Google Team To learn more about Gmail and Google Talk, visit: http://mail.google.com/mail/help/about.html http://www.google.com/talk/about.html (If clicking the URLs in this message does not work, copy and paste them into the address bar of your browser). * Not a Windows user? No problem. You can also connect to the Google Talk service from any platform using third-party clients (http://www.google.com/talk/otherclients.html). From wyanchao at gmail.com Wed Oct 25 13:44:39 2006 From: wyanchao at gmail.com (wang yanchao) Date: Wed, 25 Oct 2006 19:44:39 +0800 Subject: [Pw_forum] question Message-ID: <55639d30610250444p7c71146eq8ae0faa309f92c1a@mail.gmail.com> Dear all: I have a question when I read the source program of PWscf.The route of the source program is PWscf/PW/kpoint_grid.f90.In it I find "if(t_rev(ns).eq.1) xkr=-xkr".I think "t_rev" is used to decide time revers.But I write "t_rev" when I caculate example01 namely Cu`s scf.Theresult is "0"rather than "1".I feel very confused. if it is "0",then "(t_rev.eq.1)"is false .This indicate that Cu`structure has no time revers. But in fact, Cu`structure is fcc,it should have time revers.So I do not know it clearly.Could you please tell me why ?Thank you! sincerly Yanchao Wang -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061025/266bea8d/attachment.htm From dalcorso at sissa.it Wed Oct 25 14:00:36 2006 From: dalcorso at sissa.it (Andrea Dal Corso) Date: Wed, 25 Oct 2006 14:00:36 +0200 Subject: [Pw_forum] question In-Reply-To: <55639d30610250444p7c71146eq8ae0faa309f92c1a@mail.gmail.com> References: <55639d30610250444p7c71146eq8ae0faa309f92c1a@mail.gmail.com> Message-ID: <1161777636.3995.13.camel@dhpc-5-32.sissa.it> On Wed, 2006-10-25 at 19:44 +0800, wang yanchao wrote: > Dear all: > I have a question when I read the source program of PWscf.The > route of the source program is PWscf/PW/kpoint_grid.f90.In it I find > "if(t_rev(ns).eq.1) xkr=-xkr".I think "t_rev" is used to decide time > revers.But I write "t_rev" when I caculate example01 namely Cu`s > scf.The result is "0"rather than "1".I feel very confused. if it is > "0",then "(t_rev.eq.1)"is false .This indicate that Cu`structure has > no time revers. But in fact, Cu`structure is fcc,it should have time > revers.So I do not know it clearly.Could you please tell me why ?Thank > you! > > The reason is that t_rev is set and used only with magnetic spin-orbit calculations. In this case there might be symmetry operations which require multiplication by the time reversal operator. In all other cases t_rev is set to zero, but time reversal is implicitly used because the code takes only k or -k, not both. Andrea > > > > sincerly > > Yanchao Wang -- Andrea Dal Corso Tel. 0039-040-3787428 SISSA, Via Beirut 2/4 Fax. 0039-040-3787528 34014 Trieste (Italy) e-mail: dalcorso at sissa.it From baroni at sissa.it Wed Oct 25 14:12:13 2006 From: baroni at sissa.it (Stefano Baroni) Date: Wed, 25 Oct 2006 14:12:13 +0200 Subject: [Pw_forum] question In-Reply-To: <1161777636.3995.13.camel@dhpc-5-32.sissa.it> References: <55639d30610250444p7c71146eq8ae0faa309f92c1a@mail.gmail.com> <1161777636.3995.13.camel@dhpc-5-32.sissa.it> Message-ID: <4B7044B9-7DF9-4943-B854-1256BF837684@sissa.it> Hi! I would like first of all to congratutate Wang Yanchao on his digging into the code. This is a very good practice, which I would like to encourage. If I may just add a comment to the already exhaustive answer of Andrea: >> "0",then "(t_rev.eq.1)"is false .This indicate that Cu`structure has >> no time revers. But in fact, Cu`structure is fcc,it should have time >> revers. the same holds not only for fcc, but for ANY finite as well as infinite system in the absence of magnetic fields ... cheers - Stefano B. On Oct 25, 2006, at 2:00 PM, Andrea Dal Corso wrote: > On Wed, 2006-10-25 at 19:44 +0800, wang yanchao wrote: >> Dear all: >> I have a question when I read the source program of PWscf.The >> route of the source program is PWscf/PW/kpoint_grid.f90.In it I find >> "if(t_rev(ns).eq.1) xkr=-xkr".I think "t_rev" is used to decide time >> revers.But I write "t_rev" when I caculate example01 namely Cu`s >> scf.The result is "0"rather than "1".I feel very confused. if it is >> "0",then "(t_rev.eq.1)"is false .This indicate that Cu`structure has >> no time revers. But in fact, Cu`structure is fcc,it should have time >> revers.So I do not know it clearly.Could you please tell me why ? >> Thank >> you! >> >> > > The reason is that t_rev is set and used only with magnetic spin-orbit > calculations. In this case there might be symmetry operations which > require multiplication by the time reversal operator. > In all other cases t_rev is set to zero, but time reversal is > implicitly used because the code takes only k or -k, not both. > > Andrea > > > >> >> >> >> sincerly >> >> Yanchao Wang > -- > Andrea Dal Corso Tel. 0039-040-3787428 > SISSA, Via Beirut 2/4 Fax. 0039-040-3787528 > 34014 Trieste (Italy) e-mail: dalcorso at sissa.it > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061025/844dff29/attachment.htm From wierzbom at ts.infn.it Wed Oct 25 15:39:56 2006 From: wierzbom at ts.infn.it (Malgorzata Wierzbowska) Date: Wed, 25 Oct 2006 15:39:56 +0200 (CEST) Subject: [Pw_forum] problems with integrated charge In-Reply-To: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> References: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> Message-ID: On Tue, 24 Oct 2006, Vivek Ranjan wrote: > Hi ! > > I am trying to relax a structure. Within 3 scf steps the program aborts > with the following : > ---------------------------------------------------------------------------- > WARNING: integrated charge= 207.40294106, expected= 208.00000000 > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from electrons : error # 1 > charge is wrong > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > ------------------------------------------------------------------------------ > > I have tried the following : > > occupations='smearing', smearing='mv', degauss=0.5 > > and changed the degauss value from 0.01 initially to 0.05, then 0.1, and > now 0.5. The integrated charge remains exactly the same at > 207.40294106,when it expects it to be 208. The program aborts always in > the 3rd scf iteration. > > What could be the reason ? > Maybe, ecut, rhocut, choice of PP, or the structure is far from "physical/chemical"..... - for me it happened when I made a mistake in the geometry. Gosia > Thank you, > Vivek > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From giannozz at nest.sns.it Wed Oct 25 16:06:05 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 25 Oct 2006 16:06:05 +0200 Subject: [Pw_forum] problems with integrated charge In-Reply-To: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> References: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> Message-ID: On Oct 25, 2006, at 24:22 , Vivek Ranjan wrote: > I am trying to relax a structure. Within 3 scf steps the program > aborts > with the following : > > WARNING: integrated charge= 207.40294106, expected= > 208.00000000 > > from electrons : error # 1 > charge is wrong > [...] > What could be the reason ? incorrect data. The only reason related to the code I am aware of, is a failure of the parallel algorithm for subspace diagonalization. You can try to disable it, following the instructions I posted a few days ago. If you have evidence that your data is good and the code still doesn't work, please provide a test Paolo From giannozz at nest.sns.it Wed Oct 25 16:07:44 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 25 Oct 2006 16:07:44 +0200 Subject: [Pw_forum] Re: from Srijan In-Reply-To: <3fcc7e670610222328u2e2ab39clf795a632087dcfbf@mail.gmail.com> References: <57039.172.16.1.1.1161270463.squirrel@172.16.1.1> <39074.172.16.1.1.1161325253.squirrel@172.16.1.1> <3fcc7e670610202254p4444d95ep1ab03cd0863207c3@mail.gmail.com> <59699.172.16.1.1.1161441285.squirrel@172.16.1.1> <3fcc7e670610211140h1ed5c0t914804f08b476b6f@mail.gmail.com> <56262.172.16.1.1.1161535201.squirrel@172.16.1.1> <3fcc7e670610221013g60cd24b3o9e56c626b3bb1cf0@mail.gmail.com> <39848.172.16.1.1.1161538170.squirrel@172.16.1.1> <3fcc7e670610222238v3d67a66dr2b6247a8d13bd053@mail.gmail.com> <3fcc7e670610222328u2e2ab39clf795a632087dcfbf@mail.gmail.com> Message-ID: On Oct 23, 2006, at 8:28 , Srijan Kumar Saha wrote: > > I can bypass the problem and use smaller value of celldm (say 3 > angstrom) but I honestly want to know why my program is converging > for < 4.1 angstrom but not > for > 4.1 angstrom (even after 100 iterations)????????? Why 4.1 > angstrom is so sensitive??? bad data, bad pseudotentials, ... Paolo From khkreddy at ipc.iisc.ernet.in Wed Oct 25 18:00:05 2006 From: khkreddy at ipc.iisc.ernet.in (K.Hari Krishna Reddy) Date: Wed, 25 Oct 2006 21:30:05 +0530 (IST) Subject: [Pw_forum] NMR properties Message-ID: <1225.10.50.18.13.1161792005.squirrel@ipc.iisc.ernet.in> Hi all, Is there any way to calculate the NMR properties using the PWscf code ? -- K.Hari Krishna Reddy (Int.Ph.D) Prof. E.D.Jemmis' group (080-22933348) Inorganic and Physical Chemistry Indian Institute of Science From lyu7 at ncsu.edu Wed Oct 25 18:00:40 2006 From: lyu7 at ncsu.edu (Liping YU) Date: Wed, 25 Oct 2006 12:00:40 -0400 Subject: [Pw_forum] problems with integrated charge In-Reply-To: References: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> Message-ID: <453F8A28.40900@ncsu.edu> Paolo Giannozzi wrote: > > On Oct 25, 2006, at 24:22 , Vivek Ranjan wrote: > >> I am trying to relax a structure. Within 3 scf steps the program aborts >> with the following : >> >> WARNING: integrated charge= 207.40294106, expected= 208.00000000 >> >> from electrons : error # 1 >> charge is wrong >> [...] >> What could be the reason ? > > incorrect data. The only reason related to the code I am aware of, is > a failure of the parallel algorithm for subspace diagonalization. You > can try to disable it, following the instructions I posted a few days > ago. I often get this error message when parallel algorithm for subspace diagonalization is used for large supercell ( especially when celldm(3) is big for tetragonal cell ) calculations. For the same system with smaller super cell, there is no such problem when a serial algorithm is used. So I think this is due to parallel algorithm in the code not the bad data. To let program continue running without this error message, I changed the line in PW/electron.90 " ABS( charge - nelec ) / charge > *1.0D-3* " to " ABS( charge - nelec ) / charge > *1.d0* ". When scf calculation is close to convergence, such WARNING message will disappear. I don't know if this is a good way to do. If yes, then criterion for relative charge difference is too strict and we need to change it. If not, as Paolo suggested, maybe we need to disable the parallel algorithm for subspace diagonalization. Liping > If you have evidence that your data is good and the code still doesn't > work, please provide a test > > Paolo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum From vranjan at ncsu.edu Wed Oct 25 19:57:50 2006 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Wed, 25 Oct 2006 13:57:50 -0400 (EDT) Subject: [Pw_forum] problems with integrated charge In-Reply-To: <453F8A28.40900@ncsu.edu> References: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> <453F8A28.40900@ncsu.edu> Message-ID: <58371.152.14.74.117.1161799070.squirrel@webmail.ncsu.edu> thnx for your suggestions. i have yet to work it out. i agree that the problem occurs when the system consists of large number of atoms. i do not think its because of the bad data (if bad data means wrong initial geometry). i have checked geometry in xcrysden. i have also encountered these problems several times and often changing some parameters may make the programme work, but i have not been able to come to a systematic analysis of why it happens. its mostly trial and error. thnx anyways, vivek > Paolo Giannozzi wrote: >> >> On Oct 25, 2006, at 24:22 , Vivek Ranjan wrote: >> >>> I am trying to relax a structure. Within 3 scf steps the program aborts >>> with the following : >>> >>> WARNING: integrated charge= 207.40294106, expected= 208.00000000 >>> >>> from electrons : error # 1 >>> charge is wrong >>> [...] >>> What could be the reason ? >> >> incorrect data. The only reason related to the code I am aware of, is >> a failure of the parallel algorithm for subspace diagonalization. You >> can try to disable it, following the instructions I posted a few days >> ago. > I often get this error message when parallel algorithm for subspace > diagonalization > is used for large supercell ( especially when celldm(3) is big for > tetragonal cell ) > calculations. For the same system with smaller super cell, there is no > such problem > when a serial algorithm is used. So I think this is due to parallel > algorithm in the > code not the bad data. > > To let program continue running without this error message, I changed the > line in PW/electron.90 " ABS( charge - nelec ) / charge > *1.0D-3* " to > " ABS( charge - nelec ) / charge > *1.d0* ". When scf calculation is > close to > convergence, such WARNING message will disappear. I don't know if this is > a good way to do. If yes, then criterion for relative charge difference > is too > strict and we need to change it. If not, as Paolo suggested, maybe we need > to disable the parallel algorithm for subspace diagonalization. > > Liping > >> If you have evidence that your data is good and the code still doesn't >> work, please provide a test >> >> Paolo >> >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________________________________ web : http://www.geocities.com/vivekranjanweb __________________________________________________________________________ Vivek Ranjan Email : vranjan at chips.ncsu.edu Center for High Performance Simulation % From zxf8881 at hotmail.com Thu Oct 26 05:23:25 2006 From: zxf8881 at hotmail.com (zhou xiangfeng) Date: Thu, 26 Oct 2006 03:23:25 +0000 Subject: [Pw_forum] problem about calculate the phonon frequencies Message-ID: An HTML attachment was scrubbed... URL: /pipermail/attachments/20061026/1d379717/attachment.htm From srijan.india at gmail.com Thu Oct 26 07:37:22 2006 From: srijan.india at gmail.com (Srijan Kumar Saha) Date: Thu, 26 Oct 2006 11:07:22 +0530 Subject: [Pw_forum] Re: from Srijan In-Reply-To: References: <57039.172.16.1.1.1161270463.squirrel@172.16.1.1> <3fcc7e670610202254p4444d95ep1ab03cd0863207c3@mail.gmail.com> <59699.172.16.1.1.1161441285.squirrel@172.16.1.1> <3fcc7e670610211140h1ed5c0t914804f08b476b6f@mail.gmail.com> <56262.172.16.1.1.1161535201.squirrel@172.16.1.1> <3fcc7e670610221013g60cd24b3o9e56c626b3bb1cf0@mail.gmail.com> <39848.172.16.1.1.1161538170.squirrel@172.16.1.1> <3fcc7e670610222238v3d67a66dr2b6247a8d13bd053@mail.gmail.com> <3fcc7e670610222328u2e2ab39clf795a632087dcfbf@mail.gmail.com> Message-ID: <3fcc7e670610252237g1ded982bp159710fa344073d8@mail.gmail.com> Respected Sir, Thank you for your kind reply. But I'm unable to find out what should I change to get a good convergence. I have used the pseudopotential given in pwscf website only. Here is my input file: &control calculation='scf' restart_mode='from_scratch', prefix='LaMnO3', pseudo_dir = '/home/srijan/pseudo/', outdir='/home/srijan/srijan/lamno3/tmp/' tstress = .true. tprnfor = .true. / &system ibrav= 1, celldm(1) =14.8, nat= 5, ntyp= 3, ecutwfc =30.0, ecutrho=150.0, occupations='smearing', smearing='methfessel-paxton', degauss=0.03 nbnd=55, / &electrons conv_thr=1.0d-6, mixing_beta=0.4, / ATOMIC_SPECIES La 138.91 La.pbe-nsp-van.UPF Mn 54.93 Mn.pbe-sp-van.UPF O 16.00 O.pbe-van_bm.UPF ATOMIC_POSITIONS Mn 0.000 0.000 0.000 La 0.500 0.500 0.500 O 0.000 0.500 0.500 O 0.500 0.500 0.000 O 0.500 0.000 0.500 K_POINTS {automatic} 4 4 4 0 0 0 I want to know why my program is converging> for < 4.1 angstrom but not> for > 4.1 angstrom (even after 100 iterations)????????? Why 4.1> angstrom is so sensitive??? The problem is only with the new verson 3.1.1 but the program is converging at verson 2.0. The another problem I'm facing with the el ph calculation. running the scf calculation with dense k-point grid...2 done running the scf calculation...2 done running the el-ph calculation...2 done running q2r...forrtl: severe (24): end-of-file during read, unit 5, file stdin Image PC Routine Line Source q2r.x 083220B2 Unknown Unknown Unknown q2r.x 08320605 Unknown Unknown Unknown q2r.x 08300D3D Unknown Unknown Unknown q2r.x 082CAA2E Unknown Unknown Unknown q2r.x 082CA75E Unknown Unknown Unknown q2r.x 082DEFC8 Unknown Unknown Unknown q2r.x 0804D811 Unknown Unknown Unknown q2r.x 0804D3D5 Unknown Unknown Unknown libc.so.6 00BB0E23 Unknown Unknown Unknown q2r.x 0804D311 Unknown Unknown Unknown done running matdyn for frequency calculation...forrtl: No such file or directory forrtl: severe (29): file not found, Looking forward to your soonest reply. With best regards, Srijan -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061026/3c626a53/attachment.htm From wyanchao at gmail.com Thu Oct 26 16:18:51 2006 From: wyanchao at gmail.com (wang yanchao) Date: Thu, 26 Oct 2006 22:18:51 +0800 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1203 - 3 msgs In-Reply-To: <20061025121301.13404.20139.Mailman@democritos.sissa.it> References: <20061025121301.13404.20139.Mailman@democritos.sissa.it> Message-ID: <55639d30610260718r7c91a2b7v1fe0a39da39de07f@mail.gmail.com> Dear Andrea: Thank you for your help.When caculating el-ph interaction of the doped material, I used the smearing .I set the degauss=0.05.The total energy does not converge,so I set the different value to test.but it does not converge still.So I set the different value of K-point to test again,the result is not very good. I want to know how to deal with this program.In addtion, I want to know what the relationship between the value of degauss and K-point.Thank you! sincerely Wang yanchao -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061026/c47d8104/attachment.htm From giannozz at nest.sns.it Thu Oct 26 16:25:37 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 26 Oct 2006 16:25:37 +0200 Subject: [Pw_forum] problems with integrated charge In-Reply-To: <453F8A28.40900@ncsu.edu> References: <42577.152.14.74.117.1161728563.squirrel@webmail.ncsu.edu> <453F8A28.40900@ncsu.edu> Message-ID: <28B096EC-C939-449F-898D-412DDCFC3D59@nest.sns.it> On Oct 25, 2006, at 18:00 , Liping YU wrote: > So I think this is due to parallel algorithm in the code not the > bad data. so please try if such problem is solved in the cvs version. It should be, since the algorithm used in parallel subspace diagonalization has been rewritten > To let program continue running without this error message, I > changed the > line in PW/electron.90 " ABS( charge - nelec ) / charge > *1.0D-3* > " to > " ABS( charge - nelec ) / charge > *1.d0* ". When scf calculation is > close to convergence, such WARNING message will disappear. > I don't know if this is a good way to do. it may work: if the error on the charge is due to some insufficent convergence in the subspace diagonalization, it may disappear (but it might also NOT disappear) with the progress of self-consistency. In any event, an error of 1 electron is something to be worried about P. From wyanchao at gmail.com Thu Oct 26 16:23:27 2006 From: wyanchao at gmail.com (wang yanchao) Date: Thu, 26 Oct 2006 22:23:27 +0800 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1203 - 3 msgs In-Reply-To: <20061025121301.13404.20139.Mailman@democritos.sissa.it> References: <20061025121301.13404.20139.Mailman@democritos.sissa.it> Message-ID: <55639d30610260723l4e88293ctad655f12a36651f4@mail.gmail.com> Dear Stefano: Thank you for your encouraging! I will try my best to do it well. best wishes sincerely Yanchao Wang -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061026/684413ec/attachment.htm From ceresoli at sissa.it Thu Oct 26 16:38:50 2006 From: ceresoli at sissa.it (Davide Ceresoli) Date: Thu, 26 Oct 2006 16:38:50 +0200 Subject: [Pw_forum] NMR properties In-Reply-To: <1225.10.50.18.13.1161792005.squirrel@ipc.iisc.ernet.in> References: <1225.10.50.18.13.1161792005.squirrel@ipc.iisc.ernet.in> Message-ID: <4540C87A.8000301@sissa.it> K.Hari Krishna Reddy wrote: > Hi all, > Is there any way to calculate the NMR properties using the PWscf code ? > > > http://www.democritos.it/pipermail/pw_forum/2006-September/004883.html From giannozz at nest.sns.it Thu Oct 26 18:38:27 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 26 Oct 2006 18:38:27 +0200 Subject: [Pw_forum] Re: from Srijan In-Reply-To: <3fcc7e670610252237g1ded982bp159710fa344073d8@mail.gmail.com> References: <57039.172.16.1.1.1161270463.squirrel@172.16.1.1> <3fcc7e670610202254p4444d95ep1ab03cd0863207c3@mail.gmail.com> <59699.172.16.1.1.1161441285.squirrel@172.16.1.1> <3fcc7e670610211140h1ed5c0t914804f08b476b6f@mail.gmail.com> <56262.172.16.1.1.1161535201.squirrel@172.16.1.1> <3fcc7e670610221013g60cd24b3o9e56c626b3bb1cf0@mail.gmail.com> <39848.172.16.1.1.1161538170.squirrel@172.16.1.1> <3fcc7e670610222238v3d67a66dr2b6247a8d13bd053@mail.gmail.com> <3fcc7e670610222328u2e2ab39clf795a632087dcfbf@mail.gmail.com> <3fcc7e670610252237g1ded982bp159710fa344073d8@mail.gmail.com> Message-ID: <8CF8AA02-5864-4B2C-ADEB-C0E6C9C5948A@nest.sns.it> On Oct 26, 2006, at 7:37 , Srijan K > I'm unable to find out what should I change to get a good > convergence. it is explained in the manual. Your material turns out to be metallic or almost metallic so its convergence is typically less straightforward than for insulators. You have to experiment a bit with k-points, broadening, with parameter "mixing_beta" (reduce it), even with cutoff (you might need higher cutoffs, especially for charge density). Check if you get sensible results for the expected electronic structure: if not, there might be something wrong with pseudopotentials P. From srijan.india at gmail.com Thu Oct 26 19:21:37 2006 From: srijan.india at gmail.com (Srijan Kumar Saha) Date: Thu, 26 Oct 2006 22:51:37 +0530 Subject: [Pw_forum] Re: from Srijan In-Reply-To: <3fcc7e670610261001n465c6cfcma0665e052b4ce82a@mail.gmail.com> References: <57039.172.16.1.1.1161270463.squirrel@172.16.1.1> <56262.172.16.1.1.1161535201.squirrel@172.16.1.1> <3fcc7e670610221013g60cd24b3o9e56c626b3bb1cf0@mail.gmail.com> <39848.172.16.1.1.1161538170.squirrel@172.16.1.1> <3fcc7e670610222238v3d67a66dr2b6247a8d13bd053@mail.gmail.com> <3fcc7e670610222328u2e2ab39clf795a632087dcfbf@mail.gmail.com> <3fcc7e670610252237g1ded982bp159710fa344073d8@mail.gmail.com> <8CF8AA02-5864-4B2C-ADEB-C0E6C9C5948A@nest.sns.it> <3fcc7e670610261001n465c6cfcma0665e052b4ce82a@mail.gmail.com> Message-ID: <3fcc7e670610261021w57a4c28vc1b20ae2b0699973@mail.gmail.com> Respected Sir, Plz ignore the last mail. Before finishing, that had been mistakingly sent. Thank you very much for your kind reply. I'm little confused. Why are you saying my material is metallic????? I have heard that LaMnO3 is AFM insulator. If I don't know about my compound much then how will I make a better input file???? Is it only by trial and error method???? Or some specific rules are there??????????? You are right. It was a problem of pseudo potentials. I have changed those and getting better convergence. But I took last pseudo pot from pwscf website. Why there is only one PP for Mn????? Is there any other source where from I can get ready made PP????????? Or I have to generate pseudo potential myself???????? With best regards, Srijan -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061026/4105ad20/attachment.htm From srijan.india at gmail.com Thu Oct 26 19:01:54 2006 From: srijan.india at gmail.com (Srijan Kumar Saha) Date: Thu, 26 Oct 2006 22:31:54 +0530 Subject: [Pw_forum] Re: from Srijan In-Reply-To: <8CF8AA02-5864-4B2C-ADEB-C0E6C9C5948A@nest.sns.it> References: <57039.172.16.1.1.1161270463.squirrel@172.16.1.1> <3fcc7e670610211140h1ed5c0t914804f08b476b6f@mail.gmail.com> <56262.172.16.1.1.1161535201.squirrel@172.16.1.1> <3fcc7e670610221013g60cd24b3o9e56c626b3bb1cf0@mail.gmail.com> <39848.172.16.1.1.1161538170.squirrel@172.16.1.1> <3fcc7e670610222238v3d67a66dr2b6247a8d13bd053@mail.gmail.com> <3fcc7e670610222328u2e2ab39clf795a632087dcfbf@mail.gmail.com> <3fcc7e670610252237g1ded982bp159710fa344073d8@mail.gmail.com> <8CF8AA02-5864-4B2C-ADEB-C0E6C9C5948A@nest.sns.it> Message-ID: <3fcc7e670610261001n465c6cfcma0665e052b4ce82a@mail.gmail.com> Respected Sir, Thank you very much for your kind reply. I'm little confused. Why are you saying my material is metallic????? I have heard that LaMnO3 is AFM insulator. If you don't know about your compound much then how will you make a better input file???? Is it only by trial and error method???? Or some specific rules are there??????????? You are right. It was a problem of pseudo potentials. I have changed those and getting better convergence. But I took last pseudo pot from pwscf website. Why there is only one PP for Mn????? Is there any other source where from I can get ready made PP. On 10/26/06, Paolo Giannozzi wrote: > > > On Oct 26, 2006, at 7:37 , Srijan K > > > I'm unable to find out what should I change to get a good > > convergence. > > it is explained in the manual. Your material turns out to be metallic or > almost metallic so its convergence is typically less straightforward > than > for insulators. You have to experiment a bit with k-points, broadening, > with parameter "mixing_beta" (reduce it), even with cutoff (you might > need higher cutoffs, especially for charge density). Check if you get > sensible results for the expected electronic structure: if not, there > might be something wrong with pseudopotentials > > P. > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061026/68511057/attachment.htm From sbraccia at sissa.it Thu Oct 26 21:38:19 2006 From: sbraccia at sissa.it (Carlo Sbraccia) Date: Thu, 26 Oct 2006 21:38:19 +0200 (CEST) Subject: [Pw_forum] bfgs and trust radius In-Reply-To: <49340.152.14.74.117.1161727113.squirrel@webmail.ncsu.edu> References: <49340.152.14.74.117.1161727113.squirrel@webmail.ncsu.edu> Message-ID: My advise is that you should not reduce trust_radius_min. Instead, increase the accuracy of your forces by reducing conv_thr even more (or by increasing upscale). You can monitor the error on the forces due to lack of self-consistency by checking the size of total scf correction (it must be much smaller than the total force). However this is not the only source of numerical errors. Since some terms are computed in real space on a equispaced grid, the energy of the system is not invariant for arbitrary translations. This can easily confuse BFGS algorithm when the convergence criteria are too tight. To reduce this error you should increase ecutrho. However I don't see the reason to waste cpu-time by using such a tight criteria. Also, consider that the self-consistency itself is going to fail if you ask conv_thr < 1.D-12 because the iterative diagonalization schemes we use can suffer numerical instabilities when the required accuracy for eigenvalues/eigenvectors is to high. carlo On Tue, 24 Oct 2006, Vivek Ranjan wrote: > Hi ! > > I getting the following message : > > --------------------------------------------------------------------------- > number of scf cycles = 14 > number of bfgs steps = 1 > > energy old = -1402.7124539122 ryd > energy new = -1402.7077099529 ryd > > CASE: energy_new > energy_old > > new trust radius = 0.0000480606 bohr > > trust_radius < trust_radius_min > > resetting bfgs history > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from bfgs : error # 1 > bfgs history already reset at previous step > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > stopping ... > --------------------------------------------------------------------------------- > > Based on earlier discussions I feel the following are important > parameters, which I have set as : > > etot_conv_thr = 1.0d-4 > forc_conv_thr = 1.0d-5 > &system > ibrav= 0, > nat= 31, > ntyp= 6, > celldm(1) = 7.2225240402, > ecutwfc = 35.0, > ecutrho = 420.0 > occupations='smearing', smearing='mv', degauss=0.05 > nosym=.true. > / > &electrons > mixing_mode = 'plain' > mixing_beta = 0.5 > conv_thr = 1.0d-8 > diagonalization = 'david' , > / > &ions > upscale=100 > / > ------------------------------------------------------------------------- > > I have faced the above problem several times. But changing one thing or > the another may result in convergence. The above generally happens when > the system is already close to convergence. > > I would like to know if it is a good idea to reduce the trust_radius_min > below its present default value set in pwscf ? > > Thank you, > Vivek > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From amit76.india at gmail.com Fri Oct 27 19:48:05 2006 From: amit76.india at gmail.com (Amit Kumar) Date: Fri, 27 Oct 2006 23:18:05 +0530 Subject: [Pw_forum] URGENT Message-ID: Dear users I could not understand how one can set atomic positions in any calculations. In perticular if I know the space group PNMA62 and fractional coordinate of sample LaMnO3 molecule (Everywhere position of La Mn O1 O2 are given How will I find out atomic position of 3rd Oxygen using symetry??? Any reference for example??? ). Is anybody aware of how coordinates in premittive cell can be generated for perticular space group and frictional coordinates. How will I make a supercell of 4 formula unit for Orthorhombic structure???? Is there any good reference on formation of supercell for non cubic crystal????? Why the manual is not as clear as ABINIT in PWSCF???? Thanks in anticipations. With best regards Amit -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061027/b9810c8f/attachment.htm From baroni at sissa.it Fri Oct 27 20:42:30 2006 From: baroni at sissa.it (Stefano Baroni) Date: Fri, 27 Oct 2006 20:42:30 +0200 Subject: [Pw_forum] URGENT In-Reply-To: References: Message-ID: <4728CA91-971C-4C74-821B-2BF33328A10E@sissa.it> On Oct 27, 2006, at 7:48 PM, Amit Kumar wrote: > Why the manual is not as clear as ABINIT in PWSCF???? because we made the choice to make the code better than the manual, rather the the other way around ;-) S. --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061027/14bb052d/attachment.htm From marzari at MIT.EDU Fri Oct 27 20:10:04 2006 From: marzari at MIT.EDU (Nicola Marzari) Date: Fri, 27 Oct 2006 14:10:04 -0400 Subject: [Pw_forum] URGENT In-Reply-To: References: Message-ID: <45424B7C.7070803@mit.edu> Dear Amit, the questions you ask very relevant, but part of the basic training in this field - not something this mailing list is supposed to answer. If you do not have a colleague or supervisor that can help, your best bet at this stage is to read some the excellent books on the topic (recent ones from Richard Martin, first and foremost, or Jorge Kohanoff, or Efthimios Kaxiras, would be great starting point). Equally well, you can take an entire class in modeling and simulation: http://ocw.mit.edu/OcwWeb/Materials-Science-and-Engineering/3-320Spring-2005/CourseHome/index.htm Best luck, nicola Amit Kumar wrote: > > > Dear users > I could not understand how one can set atomic positions in any calculations. > In perticular if I know the space group PNMA62 and fractional coordinate of > sample LaMnO3 > molecule (Everywhere position of La Mn O1 O2 are given How will I find out atomic position of > > 3rd Oxygen using symetry??? Any reference for example??? ). Is anybody aware of how coordinates in premittive cell can be > generated for perticular space group and frictional coordinates. How will I make a supercell of 4 formula unit for Orthorhombic structure???? > > Is there any good reference on formation of supercell for non cubic crystal????? > Why the manual is not as clear as ABINIT in PWSCF???? > Thanks in > anticipations. > With best regards > Amit > -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu From akohlmey at cmm.chem.upenn.edu Fri Oct 27 23:37:55 2006 From: akohlmey at cmm.chem.upenn.edu (Axel Kohlmeyer) Date: Fri, 27 Oct 2006 17:37:55 -0400 Subject: [Pw_forum] URGENT In-Reply-To: References: Message-ID: <7b6913e90610271437n739a2399p48fd094b04286fd@mail.gmail.com> dear amit, please note, that a subject line of 'URGENT' and using many question marks in your mail formatting, will only help to not been taken seriously. developers/hackers tend to have sometimes unusual ways of giving an email credibility (or not). please have a look at: http://www.catb.org/~esr/faqs/smart-questions.html which explains a lot of it and how you can and should try to deal with it (note that this page contains some irony!). On 10/27/06, Amit Kumar wrote: > Dear users > I could not understand how one can set atomic positions in any calculations. > In perticular if I know the space group PNMA62 and fractional coordinate of > sample LaMnO3 > molecule (Everywhere position of La Mn O1 O2 are given How will I find out > atomic position of > > 3rd Oxygen using symetry??? Any reference for example??? ). Is anybody aware > of how coordinates in premittive cell can be > generated for perticular space group and frictional coordinates. How will I > make a supercell of 4 formula unit for Orthorhombic structure???? > > Is there any good reference on formation of supercell for non cubic > crystal????? you may want to look at the material posted at: http://www.vlab.msi.umn.edu/events/lecture.shtml and especially. http://www.vlab.msi.umn.edu/events/download/tutorial_wyckoff.pdf the people at VLAB have done a great job in re-editing the material of the recent tutorial. > Why the manual is not as clear as ABINIT in PWSCF???? because there are a _lot_ of people who ask questions like you do, but very _few_ do submit a text to be included into the manual, once they have found out how it works and how it should be explained to a new/inexperienced/regular/clueless user. unlike the common perception, writing useful documentation actually does _not_ require a lot of experience, but simply some dedication and time investment. once done changes or additions can be mailed to pwscf at pwscf.org, or the mailing list here or added to the wiki on www.quantum-espresso.org and then experienced people will review it (which takes only fraction of their time compared to having to write it from scratch and thus gives them more time to help you with the more tricky problems and improve the codes). note, that like in all other community driven projects, _your_ contribution is the way of how you 'pay back' for getting all the nice code, material, and support for free to be used in your research. > Thanks in > anticipations. > With best regards > Amit thanks in anticipation of your contribution ;-), axel. -- ======================================================================= Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu Center for Molecular Modeling -- University of Pennsylvania Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323 tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425 ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From vranjan at ncsu.edu Mon Oct 30 00:15:07 2006 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Sun, 29 Oct 2006 18:15:07 -0500 (EST) Subject: [Pw_forum] fluorine psp In-Reply-To: <7b6913e90610271437n739a2399p48fd094b04286fd@mail.gmail.com> References: <7b6913e90610271437n739a2399p48fd094b04286fd@mail.gmail.com> Message-ID: <59378.75.178.188.8.1162163707.squirrel@webmail.ncsu.edu> Does anybody have a well tested ultrasoft fluorine psp with perdew-zunger exch-corr ? Thanks in advance, vivek From amit76.india at gmail.com Mon Oct 30 08:29:10 2006 From: amit76.india at gmail.com (Amit Kumar) Date: Mon, 30 Oct 2006 12:59:10 +0530 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1208 - 1 msg In-Reply-To: <20061030063751.24735.69565.Mailman@democritos.sissa.it> References: <20061030063751.24735.69565.Mailman@democritos.sissa.it> Message-ID: >Why the manual is not as clear as ABINIT in PWSCF???? because we made the choice to make the code better than the manual, rather the the other way around ;-) Respected Users, >>>>>>>>>> It's definitely Good Choice but not the best !!!!! It would be better if the manual and the code - both are equally good. Writing a good manual is not as difficult as writing a beutiful code. I think any project asistant can do it. I don't know why in Example directory there are so many simple example programs ( Al , Ni, Cu, C, Si, etc. etc.)???????? It would be better if there is some simple examples and at least one example of some complex material of some difficult structure (i.e. non cubic say monoclinic structure). You know there are so many tutorial on how to handle simple cubic system. But no (as far as I know) tutorial on how to construct a supercell for a monoclin ic structure????? I think my question is not a stupid one. Because I have checked almost all the letters of the PWFORUM and almost half of them are just trivial one (like What's the difference between alat and bohr??? Is it posiible to include spin in PWSCF???? How to calculate electron phonon coupling using PWSCF????? All are very very trivial questions.). What do you people think??????? If you think that PWFORUM is only for experts then you can creat another forum PW FORUM_for_novice. Because it's very difficult to find out useful informations in the sea of trivial questions. I really want to learn how to handle the complicated structures in pwscf???? More specifically, construct a supercell for non-cubic structures to introduce A type , G type and C type antiferromagnetism in the system. Please don't tell me to ask the Superviser or colleague. Because most of the colleagues just use all the input files given by their Supervisers blindly. That's why for a new structure again they have to look for their Guide. I prefer to learn the stuffs myself deriving them, reading books , searching nets etc. etc. I think it would be better if some experts give some useful hints for difficult problems instead of giving 1000 simple examples (spoon feeding!!!!!!!!!!). What do you think?????? I'll be highly obliged if you kindly give me some useful references for supercell construction. By the way, many many thanks for all who gave me a long reply and informations about references on the web. With best regards, Amit -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061030/b33ee95e/attachment.htm From cyrille.barreteau at cea.fr Mon Oct 30 09:06:01 2006 From: cyrille.barreteau at cea.fr (Cyrille Barreteau) Date: Mon, 30 Oct 2006 09:06:01 +0100 Subject: [Pw_forum] relaxation Message-ID: <4545B269.8000500@cea.fr> Dear pwscf users, I am trying to relax a cuboctahedron iron cluster. I am starting from a very rough estimate of the cluster geometry and let it relax but it seems that the energy is not going down and the code just stops after a few iteration (13) with the following message: CASE: energy_new > energy_old new trust radius = 0.0000499805 bohr trust_radius < trust_radius_min resetting bfgs history %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from bfgs : error # 1 bfgs history already reset at previous step %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... Any idea of what is done wrong in my input data file (attached to this email) thanks Cyrille -- ================================================================== Cyrille Barreteau CEA Saclay DSM/DRECAM/SPCSI Batiment 462 91191 Gif sur Yvette Cedex FRANCE ~~~~~~~~~~~~~~~~~~~~~~~~ phone : +33 (0)1 69 08 29 51 fax : +33 (0)1 69 08 84 46 email : cyrille.barreteau at cea.fr ~~~~~~~~~~~~~~~~~~~~~~~~ http://www-drecam.cea.fr/Images/Pisp/cbarreteau/cbarreteau_fr.html ================================================================== -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: cuboctahedron_relax.in Url: /pipermail/attachments/20061030/631459fb/attachment.txt From lanhaiping at gmail.com Mon Oct 30 10:03:34 2006 From: lanhaiping at gmail.com (lan haiping) Date: Mon, 30 Oct 2006 17:03:34 +0800 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1208 - 1 msg In-Reply-To: References: <20061030063751.24735.69565.Mailman@democritos.sissa.it> Message-ID: Hi, if you wanna dig out some useful information related to your question , you can try to search with google , refined keywords with 'pwscf' or others. Please read the instructions such as INPUT files in $espresso/Doc, or tutorials mentioned in Axel and nicola's last mails. hope help. regards, hai-ping On 10/30/06, Amit Kumar wrote: > > > > >Why the manual is not as clear as ABINIT in PWSCF???? > > because we made the choice to make the code better than the manual, > rather the > the other way around ;-) > > > Respected Users, > > >>>>>>>>>> It's definitely Good Choice > but not the best !!!!! It would be better if the manual and the code - > both are > equally good. > Writing a good manual is not as difficult as writing a beutiful code. > I think any project asistant can do it. > I don't know why in Example directory there are so many simple example > programs > ( > Al , Ni, Cu, C, > Si, etc. etc.)???????? It would be better if there is some simple > examples and at > least one example of > some complex material of some difficult structure (i.e. non cubic say > monoclinic > structure). > You know there are so many tutorial on how to handle simple cubic > system. > But no (as far as I know) tutorial on how to construct a supercell for a > monoclin > ic structure????? > I think my question is not a stupid one. Because I have checked almost > all the > letters of the PWFORUM > and almost half of them are just trivial one (like What's the > difference between > alat and bohr??? Is it posiible to include > spin in PWSCF???? How to calculate electron phonon coupling using > PWSCF????? All > are very very trivial > questions.). What do you people think??????? > If you think that PWFORUM is only for experts then you can creat > another forum PW > FORUM_for_novice. > Because it's very difficult to find out useful informations in the sea > of trivial questions. > I really want to learn how to handle the complicated structures in > pwscf???? > More specifically, construct a supercell for non-cubic structures to > introduce A type , G type and C type > antiferromagnetism in the system. > Please don't tell me to ask the Superviser or colleague. > Because most of the colleagues just use all the input files given by > their Supervisers blindly. > That's why for a new structure again they have to look for their Guide. > I prefer to learn the stuffs myself deriving them, reading books , > searching nets etc. etc. > I think it would be better if some experts give some useful hints for > difficult problems instead of > giving 1000 simple examples (spoon feeding!!!!!!!!!!). What do you > think?????? > I'll be highly obliged if you kindly give me some useful references for > supercell construction. > > By the way, many many thanks for all who gave me a long reply and > informations about references on the > web. > > With best regards, > Amit > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061030/1d74f970/attachment.htm From wmbmacam at lg.ehu.es Mon Oct 30 10:52:35 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Mon, 30 Oct 2006 10:52:35 +0100 Subject: [Pw_forum] Non cubic structures In-Reply-To: References: Message-ID: <4545CB63.5010409@lg.ehu.es> Amit Kumar wrote: > I could not understand how one can set atomic positions in any calculations. > In perticular if I know the space group PNMA62 and fractional coordinate of > sample LaMnO3 molecule (Everywhere position of La Mn O1 O2 are given How > will I find out atomic position of 3rd Oxygen using symetry??? Well, as far as I know, PWscf doesn't know that much crystallography. It isn't enough to specify the space group and then the Wyckoff positions of your atoms. In your case, if you are studiing the Cherepanov et.al. structure, you will have to specify in your input all the 4c positions of La, all the 4a positions of Mn, all the 4c positions of O1 and all the general 8d positions of O2. The other thing you have to specify is that your cell is a primitive orthorhombic one. A bit of spoon feeding would be (please check the values): ibrav= 8, celldm(1)=a, celldm(2)=b/a, celldm(3)=c/a, nat = 20 , ntyp = 3 ATOMIC_POSITIONS {crystal} La 0.5184 0.2500 0.0070 La 0.9816 0.7500 0.5070 La 0.4816 0.7500 0.9930 La 0.0184 0.2500 0.4930 Mn 0.0000 0.0000 0.0000 Mn 0.5000 0.0000 0.5000 Mn 0.0000 0.5000 0.0000 Mn 0.5000 0.5000 0.5000 O 0.9950 0.2500 0.0750 O 0.5050 0.7500 0.5750 O 0.0050 0.7500 0.9250 O 0.4950 0.2500 0.4250 O 0.2880 0.0960 0.2300 O 0.2120 0.9040 0.7300 O 0.7120 0.5960 0.7700 O 0.7880 0.4040 0.2700 O 0.7120 0.9040 0.7700 O 0.7880 0.0960 0.2700 O 0.2880 0.4040 0.2300 O 0.2120 0.5960 0.7300 Here, the keyword is {crystal}. Had nothing been defined, the coordinates would have been read in celldm(a) units. While this doesn't affect systems with a=b=c, it is important to note. What's more, the crystal basis changes with ibrav. It is documented at the end of (line 1144 onward) in Doc/INPUT_PW. If I am terribly wrong, I'd be glad if someone corrected me. Any reference for example??? ). Is anybody aware of how coordinates in premittive cell can be > generated for perticular space group and frictional coordinates. How will I make a supercell of 4 formula unit for Orthorhombic structure???? > > Is there any good reference on formation of supercell for non cubic crystal????? > Why the manual is not as clear as ABINIT in PWSCF???? > Thanks in > anticipations. > With best regards > Amit > -- ---------------------------------------- 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 ---------------------------------------- "If you have an apple and I have an apple and we exchange these apples then you and I will still each have one apple. But if you have an idea and I have an idea and we exchange these ideas, then each of us will have two ideas." George Bernard Shaw From eminekb at yahoo.com Mon Oct 30 11:44:53 2006 From: eminekb at yahoo.com (e kb) Date: Mon, 30 Oct 2006 02:44:53 -0800 (PST) Subject: [Pw_forum] In pseudopotential with 4d electrons as valance In-Reply-To: <59378.75.178.188.8.1162163707.squirrel@webmail.ncsu.edu> Message-ID: <20061030104454.18506.qmail@web34810.mail.mud.yahoo.com> Dear users, Finally I generated an ultrasoft pseudopotential for Indium wihch treats 4d electrons as valance by using atomic code. But due to 4d, the required energy cutoff for wavefunctions is high, at least for my computational sources, 60 Ryd. It is claimed that in VASP, there exists a potential with the same treatment, 4d as valance, and requires 16.2 Ryd. (J. Furthmuller et. al., Phys. Rev. B, 72, 205106 (2005)). I wonder if they are using a different method than ld1 and if it is so, is it possible to learn this method and implement it in ld1? If that is the case I would like to contribute. Thanks in advance, Emine Kucukbenli --------------------------------- We have the perfect Group for you. Check out the handy changes to Yahoo! Groups. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061030/7d869e3a/attachment.htm From srijan.india at gmail.com Mon Oct 30 13:45:00 2006 From: srijan.india at gmail.com (Srijan Kumar Saha) Date: Mon, 30 Oct 2006 18:15:00 +0530 Subject: [Pw_forum] Non cubic structures In-Reply-To: <4545CB63.5010409@lg.ehu.es> References: <4545CB63.5010409@lg.ehu.es> Message-ID: <3fcc7e670610300445l2a29ebecvc9d414627bc6599e@mail.gmail.com> Dear Users, I think there is one error in the letter of Miguel Mart?nez Canales. He wrote : ibrav= 8, celldm(1)=a, celldm(2)=b/a, celldm(3)=c/a, > nat = 20 , ntyp = 3 I think the corrected one should be : ibrav= 8, celldm(1)=a (here 'a' is in Bohr Unit (alat)), celldm(2)=b/a (but here 'a' should be in Angstrom unit) Am I right Miguel?????????? > Dear Users I'm also having a simple confusion. My question: > ATOMIC_POSITIONS {bohr} > > Y 0.0000000E+00 0.0000000E+00 0.0000000E+00 > > Y 7.1680498 7.1680498 7.1680498 > > Cr1 3.5840249 3.5840249 3.5840249 > > Cr2 10.7520752 10.7520752 10.7520752 > > O 3.5840249 0.0000000E+00 3.5840249 > > O 3.5840249 3.5840249 0.0000000E+00 > > O 0.0000000E+00 3.5840249 3.5840249 > > O 7.1680498 3.5840249 3.5840249 > > O 3.5840249 7.1680498 3.5840249 > > O 3.5840249 3.5840249 7.1680498 > > > > > > > > If I use 2 formula unit then it should be > > > > Y 0.0000000E+00 0.0000000E+00 0.0000000E+00 > > Y 7.1680498 7.1680498 7.1680498 > > > > > > OR just > > Y 0.0000000E+00 0.0000000E+00 0.0000000E+00 > > Y 7.1680498 0.0000000E+00 0.0000000E+00 > > > > > >shifted by 0.5 in X-direction only. > I'm also very much confused with the construction of correct super cell. My main confusion is if my system is non cubic (like ibrav=8) and I generate a 2 unit supercell then what I'll put for ibrav????? Will I give the same ibrav=8????? What'll be the value of celldm(1), celldm(2), celldm(3)?????? It's really strange that there are so many simple example in the pwscf example directory but not a single example discussing that type of confusion. > Looking forward to ur soonest reply. > Thanks in Advance. Srijan -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061030/b96af087/attachment.htm From marzari at MIT.EDU Mon Oct 30 13:54:11 2006 From: marzari at MIT.EDU (Nicola Marzari) Date: Mon, 30 Oct 2006 07:54:11 -0500 Subject: [Pw_forum] Non cubic structures In-Reply-To: <3fcc7e670610300445l2a29ebecvc9d414627bc6599e@mail.gmail.com> References: <4545CB63.5010409@lg.ehu.es> <3fcc7e670610300445l2a29ebecvc9d414627bc6599e@mail.gmail.com> Message-ID: <4545F5F3.8040208@mit.edu> b/a is a ratio between lengths - so units do not matter. Can you stop your annoying posts, now ? nicola Srijan Kumar Saha wrote: > > > > Dear Users, > > I think there is one error in the letter of Miguel Mart?nez Canales. > He wrote : > > ibrav= 8, celldm(1)=a, celldm(2)=b/a, celldm(3)=c/a, > > nat = 20 , ntyp = 3 > > > I think the corrected one should be : > > ibrav= 8, celldm(1)=a (here 'a' is in Bohr Unit (alat)), > > celldm(2)=b/a (but here 'a' should be in Angstrom unit) > > Am I right Miguel?????????? > -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu From srijan.india at gmail.com Mon Oct 30 14:47:09 2006 From: srijan.india at gmail.com (Srijan Kumar Saha) Date: Mon, 30 Oct 2006 19:17:09 +0530 Subject: [Pw_forum] Non cubic structures In-Reply-To: <4545F5F3.8040208@mit.edu> References: <4545CB63.5010409@lg.ehu.es> <3fcc7e670610300445l2a29ebecvc9d414627bc6599e@mail.gmail.com> <4545F5F3.8040208@mit.edu> Message-ID: <3fcc7e670610300547v2497f153s1ee99068bd8c890c@mail.gmail.com> Dear Users, I just wanted to tell Celldm(1) is in Bohr unit. Usually a, b and c are given in angstrom. So first transform a, b, c all in Bohr and divide by a (which is in Bohr). Otherwise keep everything in angstrom first and divide by a (which is in angstrom now) and then only change 'a' in Bohr to get celldm(1). Definitely, b/a is a ratio between length. But we can't devide Angstrom by Bohr. It was not clear from Miguel's letter. So I just pointed it out for other students. Sorry I did not understand my post might be annoying for others. I'll try to not post anything again. Regards, Srijan . On 10/30/06, Nicola Marzari wrote: > > > b/a is a ratio between lengths - so units do not matter. > > Can you stop your annoying posts, now ? > > nicola > > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061030/7824d2a8/attachment.htm From marzari at MIT.EDU Mon Oct 30 16:55:36 2006 From: marzari at MIT.EDU (Nicola Marzari) Date: Mon, 30 Oct 2006 10:55:36 -0500 Subject: [Pw_forum] Non cubic structures In-Reply-To: <3fcc7e670610300547v2497f153s1ee99068bd8c890c@mail.gmail.com> References: <4545CB63.5010409@lg.ehu.es> <3fcc7e670610300445l2a29ebecvc9d414627bc6599e@mail.gmail.com> <4545F5F3.8040208@mit.edu> <3fcc7e670610300547v2497f153s1ee99068bd8c890c@mail.gmail.com> Message-ID: <45462078.1040804@mit.edu> Sorry Srijan - didn't mean to be so rude. If anything, let's try to keep a high informational content for the postings - that includes thinking carefully before asking a question, and before providing an answer. To amend, here is some useful info (unrelated to this thread, but to inject some content), on a question often asked, and for which references are difficult to come by: http://en.wikipedia.org/wiki/Birch-Murnaghan_equation_of_state nicola Srijan Kumar Saha wrote: > > > Dear Users, > > > I just wanted to tell Celldm(1) is in Bohr unit. > Usually a, b and c are given in angstrom. So first transform a, b, c > all in Bohr and > divide by a (which is in Bohr). Otherwise keep everything in angstrom > first and divide > by a (which is in angstrom now) and then only change 'a' in Bohr to get > celldm(1). > Definitely, b/a is a ratio between length. > But we can't devide Angstrom by Bohr. > It was not clear from Miguel's letter. > So I just pointed it out for other students. > Sorry I did not understand my post might be annoying for others. > I'll try to not post anything again. > > Regards, > Srijan --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu From giannozz at nest.sns.it Mon Oct 30 17:32:10 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 30 Oct 2006 17:32:10 +0100 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1208 - 1 msg In-Reply-To: References: <20061030063751.24735.69565.Mailman@democritos.sissa.it> Message-ID: <0329DFAE-7F4B-4563-AEEB-C4045FF0305F@nest.sns.it> On Oct 30, 2006, at 8:29 , Amit Kumar wrote: > Writing a good manual is not as difficult as writing a beutiful > code. > I think any project assistant can do it. writing a good manual is not easy: it requires a good knowledge of what is going on, it takes a considerable amount of time, it is much less fun than writing code, and nobody reads manuals anyway ... Paolo --- Paolo Giannozzi, Democritos and University of Udine, Italy From vranjan at ncsu.edu Mon Oct 30 18:50:50 2006 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Mon, 30 Oct 2006 12:50:50 -0500 (EST) Subject: [Pw_forum] help on generating pseudopotential In-Reply-To: <20061030104454.18506.qmail@web34810.mail.mud.yahoo.com> References: <20061030104454.18506.qmail@web34810.mail.mud.yahoo.com> Message-ID: <51981.152.14.74.117.1162230650.squirrel@webmail.ncsu.edu> Hi, I am trying to generate some pseudopotentials. Currently, I am doing test calculations in the directory "atomic_data". In the following I have copied the input file for oxygen (pseudo-gen/o.in). In particular I would like your comments on why there are two entries for 2S, two entries for 2P, and how does one choose occupation, reference energy to pseudize, and cutoff radius for 3D. ************************************************************************* &input title='O', zed=8.0, rel=0, beta=0.5, rlderiv=1.5, eminld=-4.0, emaxld=4.0, deld=0.02d0, nld=3, iswitch=3, dft='PBE', / 4 1S 1 0 2.0 1 2S 2 0 2.0 1 2P 2 1 4.0 1 3D 3 2 -1.0 1 &inputp pseudotype=3, lloc=2, file_pseudopw='OPBE.RRKJ3', / 5 2S 1 0 2.00 0.00 1.40 1.70 2S 1 0 0.00 0.05 1.40 1.70 2P 2 1 4.00 0.00 1.40 1.70 2P 2 1 0.00 0.05 1.40 1.70 3D 3 2 -2.00 0.15 1.40 1.40 *********************************************************************** Thanks in advance, Vivek From proffess at yandex.ru Mon Oct 30 19:20:52 2006 From: proffess at yandex.ru (Sergey Lisenkov) Date: Mon, 30 Oct 2006 21:20:52 +0300 (MSK) Subject: [Pw_forum] timer on XT3 Message-ID: <45464284.000003.23349@webmail9.yandex.ru> Dear developers, I just wanted to say that in CVS version of pwscf timer is not working: ftn -Mpreprocess -fastsse -O2 --target=catamount -o pw.x \ pwscf.o ../Modules/atom.o ../Modules/autopilot.o ../Modules/basic_algebra_routines.o ../Modules/berry_phase.o ../Modules/bfgs_module.o ../Modules/cell_base.o ../Modules/check_stop.o ../Modules/clocks.o ../Modules/constants.o ../Modules/constraints_module.o ../Modules/control_flags.o ../Modules/electrons_base.o ../Modules/fft_base.o ../Modules/fft_scalar.o ../Modules/fft_types.o ../Modules/functionals.o ../Modules/input_parameters.o ../Modules/io_files.o ../Modules/io_global.o ../Modules/ions_base.o ../Modules/ions_nose.o ../Modules/kind.o ../Modules/mp_global.o ../Modules/mp_wave.o ../Modules/mp.o ../Modules/metadyn_base.o ../Modules/metadyn_io.o ../Modules/metadyn_vars.o ../Modules/path_base.o ../Modules/path_formats.o ../Modules/path_variables.o ../Modules/path_opt_routines.o ../Modules/path_io_routines.o ../Modules/path_reparametrisation.o ../Modules/parallel_include.o ../Modules/parameters.o ../Modules/parser.o ../Modules/printout_base.o ../Modules/pseudo_types.o ../Modules/ptoolkit.o ../Modules/random_numbers.o ../Modules/read_cards.o ../Modules/read_namelists.o ../Modules/read_upf.o ../Modules/read_uspp.o ../Modules/recvec.o ../Modules/shmem_include.o ../Modules/splinelib.o ../Modules/stick_base.o ../Modules/task_groups.o ../Modules/timestep.o ../Modules/uspp.o ../Modules/version.o ../Modules/wavefunctions.o ../Modules/wave_base.o ../Modules/xml_io_base.o libpw.a ../flib/ptools.a ../flib/flib.a ../clib/clib.a ../iotk/src/libiotk.a -lacml -L/u/sergey/lib/fftw2d/lib/ -lfftw ../clib/clib.a(cptimer.o)(.text+0x58): In function `scnds_': : warning: times is not implemented and will always fail ( cd ../bin; ln -fs ../PW/pw.x . ) as a result: .... looking for the optimal diagonalization algorithm ... dimension time para (sec) time serial (sec) 224 0.00000000 0.00000000 320 0.00000000 0.00000000 416 0.00000000 0.00000000 512 0.00000000 0.00000000 608 0.00000000 0.00000000 704 0.00000000 0.00000000 800 0.00000000 0.00000000 ... total cpu time spent up to now is 0.00 secs In 3.1.1 version there is no such problem. Thanks, Sergey From gn203 at hermes.cam.ac.uk Mon Oct 30 19:37:42 2006 From: gn203 at hermes.cam.ac.uk (Gana Natarajan) Date: Mon, 30 Oct 2006 18:37:42 +0000 (GMT) Subject: [Pw_forum] unoccupied states Message-ID: Hi, I would like to include some unoccupied states in the conduction band of a semiconductor so that I can obtain the electronic density of states in the conduction band and also estimate the band gap. How can I do this? Thanks, Gana ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Gana Natarajan PhD student Department of Chemistry Lensfield Road Cambridge CB2 1EW United Kingdom Tel: +44 -(01223) 336 532 Fax: +44- (01223) 336 362 Gana From lanhaiping at gmail.com Mon Oct 30 19:42:05 2006 From: lanhaiping at gmail.com (lan haiping) Date: Tue, 31 Oct 2006 02:42:05 +0800 Subject: [Pw_forum] unoccupied states In-Reply-To: References: Message-ID: Dear Gana, you can set number of bands for calculation with keyword 'nbnd'. In pwscf, for a semiconductor or insulator, it default sets nbnd=nelec/2. more detail information you can find in the file $espresso/Doc/INPUT_PW. regards, hai-ping On 10/31/06, Gana Natarajan wrote: > > > Hi, > > I would like to include some unoccupied states in the conduction band of a > semiconductor so that I can obtain the electronic density of states in the > conduction band and also estimate the band gap. > How can I do this? > > > Thanks, > Gana > > > ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ > Gana Natarajan > PhD student > Department of Chemistry > Lensfield Road > Cambridge CB2 1EW > United Kingdom > Tel: +44 -(01223) 336 532 > Fax: +44- (01223) 336 362 > Gana > > > > > > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061031/6e7f9369/attachment.htm From akohlmey at cmm.chem.upenn.edu Mon Oct 30 17:48:27 2006 From: akohlmey at cmm.chem.upenn.edu (Axel Kohlmeyer) Date: Mon, 30 Oct 2006 11:48:27 -0500 (EST) Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #1208 - 1 msg In-Reply-To: Message-ID: On Mon, 30 Oct 2006, Amit Kumar wrote: AK> >Why the manual is not as clear as ABINIT in PWSCF???? AK> AK> because we made the choice to make the code better than the manual, AK> rather the AK> the other way around ;-) AK> AK> AK> Respected Users, AK> AK> >>>>>>>>>> It's definitely Good Choice AK> but not the best !!!!! It would be better if the manual and the code - AK> both are AK> equally good. AK> Writing a good manual is not as difficult as writing a beutiful code. AK> I think any project asistant can do it. well, are you willing to assign somebody to do it? are you willing to help out with it? AK> I don't know why in Example directory there are so many simple example AK> programs AK> ( AK> Al , Ni, Cu, C, AK> Si, etc. etc.)???????? It would be better if there is some simple examples AK> and at AK> least one example of AK> some complex material of some difficult structure (i.e. non cubic say AK> monoclinic AK> structure). AK> You know there are so many tutorial on how to handle simple cubic system. AK> But no (as far as I know) tutorial on how to construct a supercell for a AK> monoclin AK> ic structure????? so start writing one, and ask people here to correct it. _this_ is the way things work in 'community projects'. AK> I think my question is not a stupid one. Because I have checked almost all AK> the AK> letters of the PWFORUM AK> and almost half of them are just trivial one (like What's the difference AK> between AK> alat and bohr??? Is it posiible to include AK> spin in PWSCF???? How to calculate electron phonon coupling using AK> PWSCF????? All AK> are very very trivial AK> questions.). What do you people think??????? AK> If you think that PWFORUM is only for experts then you can creat another AK> forum PW AK> FORUM_for_novice. i think you are missing the point of a forum. splitting it does not help you a tiny bit, since to give good answers, you have to have good questions. if people only ask trivial questions (and it depends a _lot_ on the personal perspective what is a trivial question!!) you will only see trivial answers. you have to realize, that nobody here has the time or gets paid to provide you with a custom tutorial for your personal needs. the 'contract' usually is that you do a step towards people providing answers, i.e. make it easy to answer. right now i only see you complaining a lot in a style this is not respecting the large amounts of work many, many people have put into the package to make it as good as it is now. AK> Because it's very difficult to find out useful informations in the sea of AK> trivial questions. AK> I really want to learn how to handle the complicated structures in AK> pwscf???? so what is wrong with starting to learn the code using simple structures and then slowly move towards more complicated ones, once you are confident with what you have learned? [...] AK> I'll be highly obliged if you kindly give me some useful references for AK> supercell construction. after all you remarks about trivial questions, what is so difficult for you. unless you let us know where you are stumbling, nobody will help you. the quality of information you can get, completely depends on the quality of information you provide. as far as i recollect, you have the space group already, you have the irreducible positions of your atoms, now you just need to apply the symmetry operations included (usually not a big problem, if you enter the positions in fractional coordinates). AK> By the way, many many thanks for all who gave me a long reply and AK> informations about references on the AK> web. yes, but have you _looked_ at the stuff? ...or even tried it. one of the most common mistakes that people do when trying to do calculations with package programs like quantum espresso, is that they start with too complicated problems, so they cannot tell, what part of the input they don't understand. it takes a little longer to do this in stages. ...and again, please don't forget to submit your tutorial once you figured out what you needed in your case. ;-) axel. AK> AK> With best regards, AK> Amit AK> -- ======================================================================= Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu Center for Molecular Modeling -- University of Pennsylvania Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323 tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425 ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From degironc at sissa.it Tue Oct 31 00:33:34 2006 From: degironc at sissa.it (Stefano de Gironcoli) Date: Tue, 31 Oct 2006 00:33:34 +0100 (CET) Subject: [Pw_forum] In pseudopotential with 4d electrons as valance In-Reply-To: <20061030104454.18506.qmail@web34810.mail.mud.yahoo.com> References: <20061030104454.18506.qmail@web34810.mail.mud.yahoo.com> Message-ID: I think that for 4d electrons in In you could take a rather large ultrasoft radius as the nearest neighbhor distance will be set by s and p valence orbitals. So I'm surprised you need 60 Ry for them. what are your norm conserving and US radii ? stefano On Mon, 30 Oct 2006, e kb wrote: > Dear users, > Finally I generated an ultrasoft pseudopotential for Indium wihch treats 4d electrons as valance by using atomic code. > But due to 4d, the required energy cutoff for wavefunctions is high, at least for my computational sources, 60 Ryd. > It is claimed that in VASP, there exists a potential with the same treatment, 4d as valance, and requires 16.2 Ryd. > (J. Furthmuller et. al., Phys. Rev. B, 72, 205106 (2005)). > > I wonder if they are using a different method than ld1 and if it is so, is it possible to learn this method and implement it in ld1? > If that is the case I would like to contribute. > Thanks in advance, > Emine Kucukbenli > > > > --------------------------------- > We have the perfect Group for you. Check out the handy changes to Yahoo! Groups. From ezadshojaee at hotmail.com Tue Oct 31 09:41:15 2006 From: ezadshojaee at hotmail.com (Ezad Shojaee) Date: Tue, 31 Oct 2006 08:41:15 +0000 Subject: [Pw_forum] phonon calculation Message-ID: hi performing phonon calculation at a special point, the code gives us for example 12 frequencies. i want to know that how should i know the value of a special mode ,for example A {sub g}? thanx _________________________________________________________________ FREE pop-up blocking with the new MSN Toolbar - get it now! http://toolbar.msn.click-url.com/go/onm00200415ave/direct/01/ From wmbmacam at lg.ehu.es Tue Oct 31 10:05:15 2006 From: wmbmacam at lg.ehu.es (=?ISO-8859-1?Q?Miguel_Mart=EDnez_Canales?=) Date: Tue, 31 Oct 2006 10:05:15 +0100 Subject: [Pw_forum] help on generating pseudopotential In-Reply-To: <51981.152.14.74.117.1162230650.squirrel@webmail.ncsu.edu> References: <20061030104454.18506.qmail@web34810.mail.mud.yahoo.com> <51981.152.14.74.117.1162230650.squirrel@webmail.ncsu.edu> Message-ID: <454711CB.80108@lg.ehu.es> Hi Vivek, I would recommend you to first read Paolo Gianozzi's tutorial on pseudopotentials. It's the TeX file espresso/atomic_doc/pseudo-gen.tex (I'd generate the dvi file first with pdflatex). This should solve the question: > how does one choose occupation, reference energy to pseudize, and > cutoff radius for 3D. Back to: Vivek Ranjan wrote: > In particular I would like your comments on why there are two entries for 2S, > two entries for 2P, There are two entries for 2s and 2p because it is an ultrasoft pseudopotential (USPP), note that pseudotype=3. US PP's need more than one reference energies, which is why you are adding a second one for 2S and 2P with the pseudization energy =! 0. All these input parmeters are explained in the file espresso/atomic_doc/INPUT_LD1. Oh! And generating (good) pseudopotentials is not easy, and closer to cooking recipes than you'd want to. -- ---------------------------------------- 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 ---------------------------------------- "If you have an apple and I have an apple and we exchange these apples then you and I will still each have one apple. But if you have an idea and I have an idea and we exchange these ideas, then each of us will have two ideas." George Bernard Shaw From giannozz at nest.sns.it Tue Oct 31 11:08:10 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 31 Oct 2006 11:08:10 +0100 Subject: [Pw_forum] timer on XT3 In-Reply-To: <45464284.000003.23349@webmail9.yandex.ru> References: <45464284.000003.23349@webmail9.yandex.ru> Message-ID: On Oct 30, 2006, at 19:20 , Sergey Lisenkov wrote: > Dear developers > > I just wanted to say that in CVS version of pwscf timer is not working > [...] > ../clib/clib.a(cptimer.o)(.text+0x58): In function `scnds_': > : warning: times is not implemented and will always fail dear Sergey, thanks for bringing this to our atention, but if XT3 does not implement a standard library, it is not our fault... Paolo --- Paolo Giannozzi, Democritos and University of Udine, Italy From marzari at MIT.EDU Tue Oct 31 12:50:38 2006 From: marzari at MIT.EDU (Nicola Marzari) Date: Tue, 31 Oct 2006 06:50:38 -0500 Subject: [Pw_forum] Announcement - Trieste Total Energy Workshop, 11-13 January 2007 Message-ID: <4547388E.8080406@mit.edu> Dear Colleague, the 13th International Workshop on "Computational Physics and Materials Science: Total Energy and Force Methods" will take place in Trieste (Italy), 11-13 January 2007. More informations and application forms can be found at http://cdsagenda5.ictp.trieste.it/full_display.php?ida=a06178 http://www.democritos.it/events/totalenergy-XIII.php The registration deadline has been postponed until 15 November 2006. There is no registration charge to attend the workshop, but accommodation is provided on a first-registered basis. The conference will last for three full days - invited speakers are listed below. In addition, there will be three plenary lectures from Federico Capasso (Harvard University), Marvin Cohen (University of California, Berkeley), and Joshua Zak (Technion, Haifa), to celebrate the 60th birthdays of Alfonso Baldereschi, Roberto Car, and Raffaele Resta (Happy birthday(s) !). The conference is supported by ICTP, Sissa, Democritos, and Psi-k. US-based participants can apply for travel funding through the Materials Computation Center (http://www.mcc.uiuc.edu/travel/). We look forward to seeing you in Trieste ! Nicola Marzari, Pablo Ordejon, and Sandro Scandolo [Invited Speakers - Trieste Total Energy 2007] Igor A. Abrikosov (Linkopings University) Hadi Akbarzadeh (Isfahan University of Technology) Stefano Baroni (SISSA/Democritos) Aitor Bergara (Universidad del Pais Vasco) Stefan Bluegel (IFF Julich) Fabien Bruneval (Ecole Polytechnique, Paris) Kieron Burke (UC Irvine) Gerbrand Ceder (MIT) David Ceperley (University of Illinois, Urbana-Champaign) David Drabold (Ohio University) Stefano de Gironcoli (SISSA/Democritos) Alessandro De Vita (King's College, London) Dario Estrin (*) (University of Buenos Aires) Pablo Garcia Gonzales (UNED, Madrid) Antoine Georges (*) (Ecole Polytechnique, Paris) Feliciano Giustino (University of California, Berkeley) Xingao Gong (Fudan University) Thomas Kuehne (ETH Zurich/USI Lugano) Nicolas Lorente (Universite Paul Sabatier) Bengt Lundqvist (Chalmers University) Igor Mazin (Naval Research Laboratory) Francesco Mauri (Universite Pierre et Marie Curie) Richard Needs (University of Cambridge) Eva Pavarini (IFF Julich) Patrick Rinke (Fritz-Haber Institute, Berlin) Annabella Selloni (Princeton University) Massimiliano Stengel (University of California, Santa Barbara) Axel Svane (University of Aarhus) Timo Thonhauser (Rutgers University/MIT) Ilya Tokatly (*) (Erlangen/Nurnburg) Paolo Umari (Democritos) Umesh Waghmare (JNCASR Bangalore) (*) to be confirmed -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 2586534 marzari at mit.edu http://quasiamore.mit.edu From giannozz at nest.sns.it Tue Oct 31 15:16:21 2006 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 31 Oct 2006 15:16:21 +0100 Subject: [Pw_forum] Problems with yahoo accounts Message-ID: The system administrator of the pw_forum mailing list informs us that there is currently some trouble with yahoo accounts. Due to ill-conceived changes to the yahoo mail servers, many postings were not delivered to yahoo accounts. Right now the situation is that yahoo accounts receive postings with a delay and flagged as spam. Users with a yahoo accounts are advised to enter the SPAM folder, click on the "it's not spam button, and ad the sender to the list of contacts (or adressbook, or whatever it is called) so that at least postings from that sender will no longer be directed to the SPAM folder Paolo --- Paolo Giannozzi, Democritos and University of Udine, Italy From manij at hotmail.com Tue Oct 31 19:15:16 2006 From: manij at hotmail.com (Manish Jain) Date: Tue, 31 Oct 2006 12:15:16 -0600 Subject: [Pw_forum] InfraRed intensities Message-ID: Hello I am interested in calculating IR spectra for an insulator. Is the following procedure for calculating IR spectra (intensities and position of peaks) correct? a) Calculate the scf charge density and wavefunctions at a minimized energy structure. b) Do a phonon calculation on a grid of q points with epsil set to true. c) Use dynmat.x to compute the cross sections of individual modes. d) Then combine them all using some means to get an IR spectra??? Any help is appreciated. Thanks, Manish _________________________________________________________________ Try Search Survival Kits: Fix up your home and better handle your cash with Live Search! http://imagine-windowslive.com/search/kits/default.aspx?kit=improve&locale=en-US&source=hmtagline From baroni at sissa.it Tue Oct 31 21:59:57 2006 From: baroni at sissa.it (Stefano Baroni) Date: Tue, 31 Oct 2006 21:59:57 +0100 Subject: [Pw_forum] InfraRed intensities In-Reply-To: References: Message-ID: On Oct 31, 2006, at 7:15 PM, Manish Jain wrote: > Hello > > I am interested in calculating IR spectra for an insulator. Is the > following procedure for calculating IR spectra (intensities and > position of peaks) correct? > > a) Calculate the scf charge density and wavefunctions at a > minimized energy structure. > b) Do a phonon calculation on a grid of q points with epsil set to > true. > c) Use dynmat.x to compute the cross sections of individual modes. > d) Then combine them all using some means to get an IR spectra??? > > Any help is appreciated. > > Thanks, > Manish > > _________________________________________________________________ > Try Search Survival Kits: Fix up your home and better handle your > cash with Live Search! http://imagine-windowslive.com/search/kits/ > default.aspx?kit=improve&locale=en-US&source=hmtagline > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061031/d713f73b/attachment.htm From baroni at sissa.it Tue Oct 31 22:06:21 2006 From: baroni at sissa.it (Stefano Baroni) Date: Tue, 31 Oct 2006 22:06:21 +0100 Subject: [Pw_forum] InfraRed intensities In-Reply-To: References: Message-ID: <46479CEB-4182-42A4-900C-CE9E0D2F199F@sissa.it> On Oct 31, 2006, at 7:15 PM, Manish Jain wrote: > Hello > > I am interested in calculating IR spectra for an insulator. Is the > following procedure for calculating IR spectra (intensities and > position of peaks) correct? > > a) Calculate the scf charge density and wavefunctions at a > minimized energy structure. yep! > b) Do a phonon calculation on a grid of q points with epsil set to > true. wait! phonon absorption is a process in which the initial state is a crystal in the ground state (q=0) and a photon with a momentum q'<<2 \pi /a (let's say, q'\approx 0), and the final state is the crystal with a phonon with wavevector q" and no photon. Given momentum conservation, can you tell which values of the phonon wavevectors (q") are allowed? > c) Use dynmat.x to compute the cross sections of individual modes. yeap, but only for those wavevectors (q") that are allowed by conservation laws (see point b) > d) Then combine them all using some means to get an IR spectra??? OK, but what means are "some means"? > Any help is appreciated. Hope this helps a bit. Stefano B. > > Thanks, > Manish > > _________________________________________________________________ > Try Search Survival Kits: Fix up your home and better handle your > cash with Live Search! http://imagine-windowslive.com/search/kits/ > default.aspx?kit=improve&locale=en-US&source=hmtagline > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20061031/74847ad4/attachment.htm From manij at hotmail.com Tue Oct 31 23:27:43 2006 From: manij at hotmail.com (Manish Jain) Date: Tue, 31 Oct 2006 16:27:43 -0600 Subject: [Pw_forum] InfraRed intensities In-Reply-To: <46479CEB-4182-42A4-900C-CE9E0D2F199F@sissa.it> Message-ID: Dear Prof. Baroni, Thanks for your quick response. >>b) Do a phonon calculation on a grid of q points with epsil set to true. > >wait! phonon absorption is a process in which the initial state is a >crystal in the ground state (q=0) and a photon with a momentum q'<<2 \pi /a >(let's say, q'\approx 0), and the final state is the crystal with a phonon >with wavevector q" and no photon. Given momentum conservation, can you >tell which values of the phonon wavevectors (q") are allowed? I misstated. I was thinking about calculating the IR spectra AND the phonon dispersions.. (the sampling was not meant for the purpose of IR intensities). I understand that for IR spectra only the gamma point calculation is needed. >>d) Then combine them all using some means to get an IR spectra??? > >OK, but what means are "some means"? By "some means" I was referring to their calculation of the absorption from each IR active mode (calculated at the gamma point) and then just adding the spectra up to get the complete spectra. I was not sure if the code - dynmat.x - did this automatically or if I needed to do this separately. > >>Any help is appreciated. > >Hope this helps a bit. This clarifies a great deal. Thank you once again. Regards, Manish _________________________________________________________________ Get today's hot entertainment gossip http://movies.msn.com/movies/hotgossip?icid=T002MSN03A07001