From ans341 at mail.usask.ca Wed Sep 10 22:34:05 2008 From: ans341 at mail.usask.ca (ans341 at mail.usask.ca) Date: Wed, 10 Sep 2008 14:34:05 -0600 Subject: [Wannier] problems with installation Message-ID: <1221078845.48c82f3d1e914@webmail.usask.ca> Dear Wannier90 users, I have been trying to install Wannier90 according to instructions but without any success. The problem is in libraries (or my limited experience in program installation). I have tried the Intel's mkl library, ATLAS and the library included in quantum-espresso (4.0cvs). None of them works, the only result is a very long list of undefined references like: /home/.../flib/lapack.a(lapack_all.o)(.text+0x69d3): In function `zpotrf_': : undefined reference to `zgemm_' /home/.../flib/lapack.a(lapack_all.o)(.text+0x6a85): In function `zpotrf_': : undefined reference to `ztrsm_' /home/.../flib/lapack.a(lapack_all.o)(.text+0x6bb2): In function `zpotrf_': : undefined reference to `zherk_' /home/.../flib/lapack.a(lapack_all.o)(.text+0x6cb1): In function `zpotrf_': : undefined reference to `zgemm_' /home/.../flib/lapack.a(lapack_all.o)(.text+0x6d5b): In function `zpotrf_': : undefined reference to `ztrsm_' /home/.../flib/lapack.a(lapack_all.o)(.text+0x701b): In function `zpotf2_': : undefined reference to `zdotc_' Most probably I have forgotten something or done something wrong. In any case I would appreciate any help with this. Thank you. Anna Sankari From jry20 at cam.ac.uk Fri Sep 12 12:45:50 2008 From: jry20 at cam.ac.uk (Jonathan Yates) Date: Fri, 12 Sep 2008 11:45:50 +0100 (BST) Subject: [Wannier] problems with installation In-Reply-To: <1221078845.48c82f3d1e914@webmail.usask.ca> References: <1221078845.48c82f3d1e914@webmail.usask.ca> Message-ID: On Wed, 10 Sep 2008, ans341 at mail.usask.ca wrote: > I have been trying to install Wannier90 according to instructions but without > any success. The problem is in libraries (or my limited experience in program > installation). I have tried the Intel's mkl library, ATLAS and the library > included in quantum-espresso (4.0cvs). None of them works, the only result is a > very long list of undefined references like: > > /home/.../flib/lapack.a(lapack_all.o)(.text+0x69d3): In function `zpotrf_': > : undefined reference to `zgemm_' Anna, It's not really possible to give an exact diagnosis from the information you gave. As you say, something is going wrong at the linking stage (lapack and blas). The error message you posted is from the lapack library saying it can't find any BLAS routines. That can happen with some compilers if you get the BLAS and LAPACK libraries in the wrong order on the link line (lapack then blas) But I doubt that is the real problem or at least not in all the cases?. Focus on mkl (and I assume the intel compiler) - I think that might be the easier one to solve first. Can you post the following information: 1) Which version of Wannier90 2) which version of ifort (and the platform) 3) The link line and the first few error messages, ie scroll back up to find a line like: ifort wannier_prog.F90 -O2 -Vaxlib constants.o io.o utility.o parameters.o hamiltonian.o overlap.o kmesh.o disentangle.o wannierise.o plot.o transport.o -lmkl_lapack -lmkl -lguide -lpthread -o ../wannier90.x make[1]: Leaving directory `/scratch/jry20/WANNIER90/WAN1.1/src' 4) Have you managed to compile any other program using lapack and blas on your machine? Jonathan -- Dr Jonathan Yates | Theory of Condensed Matter Group Corpus Christi College | Cavendish Laboratory Cambridge, CB2 1RH, UK | Cambridge, CB3 OHE, UK email jry20 at cam.ac.uk | Tel +44 (0)1223 337461 From lanhaiping at gmail.com Tue Sep 16 15:53:00 2008 From: lanhaiping at gmail.com (lan haiping) Date: Tue, 16 Sep 2008 21:53:00 +0800 Subject: [Wannier] Question about wannier center ! Message-ID: Dear All, I read several lines of wannier routines ,and try to understand several key points formulated in the paper and programme. My question is about wannier center formulated in wannier90. In the MV paper, \bar{r}_n = -\frac{1}{N} \Sum_{k,b} w_b b Im lnM_{n,n}^{k,b}, while wan_omega subroutine formulates this equation with 2 quantities: sheet and csheet. I donot catch this point ,and donot know the meaning of this 2 quantities. Would you please explain some details in this point ? Thank you! -- Hai-Ping Lan Department of Electronics , Peking University , Bejing, 100871 lanhaiping at gmail.com, hplan at pku.edu.cn -------------- next part -------------- An HTML attachment was scrubbed... URL: From f_mirjani at ph.iut.ac.ir Wed Sep 17 13:49:36 2008 From: f_mirjani at ph.iut.ac.ir (fatemeh.mirjani) Date: Wed, 17 Sep 2008 15:19:36 +0330 (IRST) Subject: [Wannier] Wien2k2Wannier90 In-Reply-To: <22465509.73251221651307969.JavaMail.root@mta.iut.ac.ir> Message-ID: <17357862.73541221652176233.JavaMail.root@mta.iut.ac.ir> Dear all ; I have written an interface code between Wien2k & Wannier90. I tested my code for Silicon (4valance bands), but unfortunately the bandstructure of Wien2k2Wannier90 doesn't coincide to the bandstructure of PW2Wannier90. The total shape of band-structure from Wien2k to Wannier90 matches with the shape of band-structure from PWscf to Wannier90, but it oscillates a lot! Generally speaking, at the kpoints with higu symmetry and the first and the end of bandstructure there is agreement with PWscf but in other points there are a lot of oscillations. (the amount of spread in seedname.wout is very large and it cannot be improved by changing the centers and increasing the number of kpoints.) Please let me know if there is anyone that can guide me on this issue. Best, Fatemeh -- +------------------------------------------------------------------------------+ |Fatemeh Mirjani | Email: f_mirjani at ph.iut.ac.ir | |Computational Condensed Matter Research Lab. | Tel/Fax Office: +98311-3913746 | |ICTP Affiliated Center in Isfahan, | Tel Lab.: +98311-3913731 | +----------Physics Department, Isfahan University of Technology, Iran----------+ From jry20 at cam.ac.uk Wed Sep 17 17:23:50 2008 From: jry20 at cam.ac.uk (Jonathan Yates) Date: Wed, 17 Sep 2008 16:23:50 +0100 (BST) Subject: [Wannier] Question about wannier center ! In-Reply-To: References: Message-ID: On Tue, 16 Sep 2008, lan haiping wrote: > I read several lines of wannier routines ,and try to understand several > key > points formulated in the paper and programme. > My question is about wannier center formulated in wannier90. In the MV > paper, > \bar{r}_n = -\frac{1}{N} \Sum_{k,b} w_b b Im lnM_{n,n}^{k,b}, while > wan_omega > subroutine formulates this equation with 2 quantities: sheet and csheet. > I donot catch this point ,and donot know the meaning of this 2 quantities. > Would you please explain some details in this point ? Unless you use the guiding centres option, csheet=1.0 and sheet=0.0 and the formula in wann_omega reduces to the equation you reproduce above. In the above formula (and that for Omega_D) there appears the log of M_mn, ie the log of a complex number. This introduces a possible ambiguity (ie different branch cuts - see section IV.D.2 of MV). In many cases this causes no problems, but in certain systems it does cause problems in the minimisation - typically one gets a few or many WF with very large spread, and the value of Omega_D is large. One solution is to use the guiding_centres option. You will find more details in the wann_phases subroutine - here csheet and sheet are computed. These are then used control the choice of branch cut in wann_omega. Jonathan -- Dr Jonathan Yates | Theory of Condensed Matter Group Corpus Christi College | Cavendish Laboratory Cambridge, CB2 1RH, UK | Cambridge, CB3 OHE, UK email jry20 at cam.ac.uk | Tel +44 (0)1223 337461 From jry20 at cam.ac.uk Wed Sep 17 18:42:47 2008 From: jry20 at cam.ac.uk (Jonathan Yates) Date: Wed, 17 Sep 2008 17:42:47 +0100 (BST) Subject: [Wannier] Wien2k2Wannier90 In-Reply-To: <17357862.73541221652176233.JavaMail.root@mta.iut.ac.ir> References: <17357862.73541221652176233.JavaMail.root@mta.iut.ac.ir> Message-ID: On Wed, 17 Sep 2008, fatemeh.mirjani wrote: > I have written an interface code between Wien2k & Wannier90. I tested my > code for Silicon (4valance bands), but unfortunately the bandstructure > of Wien2k2Wannier90 doesn't coincide to the bandstructure of > PW2Wannier90. The total shape of band-structure from Wien2k to Wannier90 > matches with the shape of band-structure from PWscf to Wannier90, but it > oscillates a lot! Generally speaking, at the kpoints with higu symmetry > and the first and the end of bandstructure there is agreement with PWscf > but in other points there are a lot of oscillations. Remember that in plotting the bandstructure with wannier90 you are doing an interpolation. The original points (ie those you computed using the ab-initio code) will be reproduced exactly - I suspect these happen to coincide with the high-symmetry points you mention. But all other points are interpolated. The fact that you get lots of oscillations suggests to me that the Hamiltonian does not decay rapidly in your wannier basis - eg because your wannier functions are not well localised. Indeed you say: > (the amount of spread in seedname.wout is very large and it cannot be > improved by changing the centers and increasing the number of kpoints.) That's not a good sign. For silicon you should be able to get four well localised, bond centred wannier functions. The spread should be very close to the value in the original MV paper (or as given by pwscf+w90). You need to solve this problem before you can expect to get a good band structure. Jonathan -- Dr Jonathan Yates | Theory of Condensed Matter Group Corpus Christi College | Cavendish Laboratory Cambridge, CB2 1RH, UK | Cambridge, CB3 OHE, UK email jry20 at cam.ac.uk | Tel +44 (0)1223 337461 From f_mirjani at ph.iut.ac.ir Thu Sep 18 21:56:36 2008 From: f_mirjani at ph.iut.ac.ir (fatemeh.mirjani) Date: Thu, 18 Sep 2008 23:26:36 +0330 (IRST) Subject: [Wannier] Wien2k2Wannier90 Message-ID: <2027875.85211221767796977.JavaMail.root@mta.iut.ac.ir> Dear Jonathan ; Is it possible to guess that the bug in my code (which causes to bad localization) is because of A_mn(k) or because of M_mn(k,b)? Thanks in advance for your guidelines. With my best regards, Fatemeh -- +------------------------------------------------------------------------------+ |Fatemeh Mirjani | Email: f_mirjani at ph.iut.ac.ir | |Computational Condensed Matter Research Lab. | Tel/Fax Office: +98311-3913746 | |ICTP Affiliated Center in Isfahan, | Tel Lab.: +98311-3913731 | +----------Physics Department, Isfahan University of Technology, Iran----------+ From lanhaiping at gmail.com Thu Sep 18 21:15:20 2008 From: lanhaiping at gmail.com (lan haiping) Date: Fri, 19 Sep 2008 03:15:20 +0800 Subject: [Wannier] Question about wannier center ! In-Reply-To: References: Message-ID: Thank you ,Dr Yates . I read wann_phases, and didnot notice the control tag on these two quantities. Best, Hai-Ping On Wed, Sep 17, 2008 at 11:23 PM, Jonathan Yates wrote: > On Tue, 16 Sep 2008, lan haiping wrote: > > > I read several lines of wannier routines ,and try to understand several > > key > > points formulated in the paper and programme. > > My question is about wannier center formulated in wannier90. In the MV > > paper, > > \bar{r}_n = -\frac{1}{N} \Sum_{k,b} w_b b Im lnM_{n,n}^{k,b}, while > > wan_omega > > subroutine formulates this equation with 2 quantities: sheet and csheet. > > I donot catch this point ,and donot know the meaning of this 2 > quantities. > > Would you please explain some details in this point ? > > Unless you use the guiding centres option, csheet=1.0 and sheet=0.0 and > the formula in wann_omega reduces to the equation you reproduce above. > > In the above formula (and that for Omega_D) there appears the log of > M_mn, ie the log of a complex number. This introduces a possible ambiguity > (ie different branch cuts - see section IV.D.2 of MV). In many cases this > causes no problems, but in certain systems it does cause problems in the > minimisation - typically one gets a few or many WF with very large spread, > and the value of Omega_D is large. One solution is to use the > guiding_centres option. > You will find more details in the wann_phases subroutine - here csheet > and sheet are computed. These are then used control the choice of branch > cut in wann_omega. > > Jonathan > > > -- > Dr Jonathan Yates | Theory of Condensed Matter Group > Corpus Christi College | Cavendish Laboratory > Cambridge, CB2 1RH, UK | Cambridge, CB3 OHE, UK > email jry20 at cam.ac.uk | Tel +44 (0)1223 337461 > > _______________________________________________ > Wannier mailing list > Wannier at quantum-espresso.org > http://www.democritos.it/mailman/listinfo/wannier > -- Hai-Ping Lan Department of Electronics , Peking University , Bejing, 100871 lanhaiping at gmail.com, hplan at pku.edu.cn -------------- next part -------------- An HTML attachment was scrubbed... URL: From f_mirjani at ph.iut.ac.ir Fri Sep 19 21:05:35 2008 From: f_mirjani at ph.iut.ac.ir (fatemeh.mirjani) Date: Fri, 19 Sep 2008 22:35:35 +0330 (IRST) Subject: [Wannier] Wien2k2Wannier90 Message-ID: <18851821.92631221851135514.JavaMail.root@mta.iut.ac.ir> Dear all; if the spread becomes negative what does it mean? I found some mistakes in my code so I edited it . but I haven't succeed to get correct WFs yet. the flactuations in the bandstructure decreased but it is still wrong ! the spread becomes negative: Final State WF centre and spread 1 ( -1.171555, 0.170504, -0.083959 ) -380.61012680 WF centre and spread 2 ( 1.200799, -1.606712, -0.120351 ) -20.08399486 WF centre and spread 3 ( 0.126184, 0.772680, 2.072708 ) -384.87801891 WF centre and spread 4 ( 0.879826, 1.219855, -0.570418 ) -26.25860467 Sum of centres and spreads ( 1.035253, 0.556326, 1.297980 ) -811.83074525 Spreads (Bohr^2) Omega I = -1298.918567808 ================ Omega D = 213.863027239 Omega OD = 273.224795323 Final Spread (Bohr^2) Omega Total = -811.830745246 I know the omega & omega I & D should be positive. In this calculation only omega I is negative (Also the amoun of other omega aren't reasonable). With regard to the fact that omega I is invarient does it mean the projection is wrong? Thanks in advance. Fatemeh -- +------------------------------------------------------------------------------+ |Fatemeh Mirjani | Email: f_mirjani at ph.iut.ac.ir | |Computational Condensed Matter Research Lab. | Tel/Fax Office: +98311-3913746 | |ICTP Affiliated Center in Isfahan, | Tel Lab.: +98311-3913731 | +----------Physics Department, Isfahan University of Technology, Iran----------+ From jry20 at cam.ac.uk Sat Sep 20 15:03:03 2008 From: jry20 at cam.ac.uk (Jonathan Yates) Date: Sat, 20 Sep 2008 14:03:03 +0100 (BST) Subject: [Wannier] Wien2k2Wannier90 In-Reply-To: <18851821.92631221851135514.JavaMail.root@mta.iut.ac.ir> References: <18851821.92631221851135514.JavaMail.root@mta.iut.ac.ir> Message-ID: On Fri, 19 Sep 2008, fatemeh.mirjani wrote: > if the spread becomes negative what does it mean? > I found some mistakes in my code so I edited it . but I haven't succeed to get correct WFs yet. > the flactuations in the bandstructure decreased but it is still wrong ! > the spread becomes negative: > Final State > WF centre and spread 1 ( -1.171555, 0.170504, -0.083959 ) -380.61012680 > WF centre and spread 2 ( 1.200799, -1.606712, -0.120351 ) -20.08399486 > WF centre and spread 3 ( 0.126184, 0.772680, 2.072708 ) -384.87801891 > WF centre and spread 4 ( 0.879826, 1.219855, -0.570418 ) -26.25860467 > Sum of centres and spreads ( 1.035253, 0.556326, 1.297980 ) -811.83074525 > > Spreads (Bohr^2) Omega I = -1298.918567808 > ================ Omega D = 213.863027239 > Omega OD = 273.224795323 > Final Spread (Bohr^2) Omega Total = -811.830745246 > > I know the omega & omega I & D should be positive. In this calculation > only omega I is negative (Also the amoun of other omega aren't > reasonable). With regard to the fact that omega I is invarient does it > mean the projection is wrong? For an isolated group of bands Omega_I is determined only from the data contained in the *.mmn file - so you can ignore the projections and focus on getting the Mmn matrix computed correctly. A simpler test case, which I've found useful in the past, is to consider just the lower valence band of GaAs. This gives an s-like WF centred on the As atom (see Table III of MV for representative values of the spread) Jonathan -- Dr Jonathan Yates | Theory of Condensed Matter Group Corpus Christi College | Cavendish Laboratory Cambridge, CB2 1RH, UK | Cambridge, CB3 OHE, UK email jry20 at cam.ac.uk | Tel +44 (0)1223 337461