From lucre at ms1.hinet.net Wed Oct 1 06:10:15 2003 From: lucre at ms1.hinet.net (Lu Frederic) Date: Wed, 1 Oct 2003 12:10:15 +0800 Subject: [Pw_forum] Re: About PWscf 1.3 version Message-ID: <003201c387d1$ec7aeee0$ef3310ac@phys> Dear Igor and PWscf users, First of all, I am not reply the issue about PWscf 1.3 version, because there are something are general, like as band structure of HCP structure using PWscf. If you want to do the calculation of band structure using PWscf for MgB2 (hexagonal structure), you must be careful in the setting of K_POINTS, because they read k-points in 2pi/a units, no matter what c =! a. So you have two choice: one is KPOINTS { crystal } that is my favorite and I always use it. the other is KPOINTS in 2pi/a units, you must transform 2pi/c into 2pi/a in c-axis if 2pi/c = 0.5, c/a=1.140255, then 2pi/a is 0.5/1.140255 = 0.4384984 I post your band.in but I made some modified, you will get the some result as me(see the enclosure: figure format). Best regards, Lu Fu-Fa, 10/01/2003 -------------- next part -------------- A non-text attachment was scrubbed... Name: hexMgB2_pwscf.gif Type: image/gif Size: 9616 bytes Desc: not available Url : /pipermail/attachments/20031001/4154bf94/attachment.gif -------------- next part -------------- A non-text attachment was scrubbed... Name: MgB2.band_shein.in Type: application/octet-stream Size: 3521 bytes Desc: not available Url : /pipermail/attachments/20031001/4154bf94/attachment.obj From ferretti.andrea at unimore.it Wed Oct 1 11:16:01 2003 From: ferretti.andrea at unimore.it (Andrea Ferretti) Date: Wed, 1 Oct 2003 11:16:01 +0200 (CEST) Subject: [Pw_forum] pseudopotential generation In-Reply-To: <003201c387d1$ec7aeee0$ef3310ac@phys> Message-ID: dear all, I'm trying to generate a pseudopotential for Pt using Paolo Giannozzi's code... after some work (thanks Paolo) I succeded in generating something that can at least be tested... I did many checks on cutoff radii and atomic configurations but when I used ldb code (in the package) to check KB form I always found that a ghost appears if I use lloc=2, desappears with lloc=1 and also in this better case I'm not able to reproduce 7s and 6d atomic eigenvalues (Pt has a 6s1 6p0 5d9 ground state configuration).... I report the output of ldb code: ----------------------------------------------------------------- l_loc is read from PP file, l_loc read on input is ignored!!! Warning: PPs for 2 Dear pwscf community! I have tried to create a Si pseudopotenial with Paolo Giannozzi's programs ld1/trou (accepting the suggested cutoff radii) and converted it using ps2cp and cpmd2upf (pwscf v1.3; I've attached shell scripts with input parameters). When I run pw.x for an scf calculation, I get band energies of about 250 eV instead of a few eV (comparing to the results of pw_examples/example5). What's going wrong? Thank you for any help! With kind regards, Danilo Neuber -- Dipl.-Ing. Danilo Neuber Technische Universit?t Graz Telefon: +43 / 316 / 873 - 8188 Institut f?r Theoretische Physik Telefax: +43 / 316 / 873 - 8678 Petersgasse 16, A-8010 Graz -------------- next part -------------- A non-text attachment was scrubbed... Name: create_si_pp.sh Type: application/x-sh Size: 809 bytes Desc: create_si_pp.sh Url : /pipermail/attachments/20031001/afe84f80/attachment.sh -------------- next part -------------- A non-text attachment was scrubbed... Name: calc_si_bs.sh Type: application/x-sh Size: 1411 bytes Desc: calc_si_bs.sh Url : /pipermail/attachments/20031001/afe84f80/attachment-0001.sh From giannozz at nest.sns.it Wed Oct 1 15:14:31 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 1 Oct 2003 15:14:31 +0200 Subject: [Pw_forum] PP Generation/Conversion for Si In-Reply-To: References: Message-ID: <200310011514.31147.giannozz@nest.sns.it> On Wednesday 01 October 2003 15:03, Danilo Neuber wrote: > I have tried to create a Si pseudopotential with Paolo Giannozzi's > programs ld1/trou (accepting the suggested cutoff radii) and > converted it using ps2cp and cpmd2upf actually you can use the pseudopotential file produced by "trou" with no further conversion. If you want to convert it to upf format (which I recommand), use "ncpp2upf". Converting twice as you did should also work, though, but the probability of making a mistake increases at each conversion ... Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From giannozz at nest.sns.it Wed Oct 1 15:25:25 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 1 Oct 2003 15:25:25 +0200 Subject: [Pw_forum] pseudopotential generation In-Reply-To: References: Message-ID: <200310011525.25077.giannozz@nest.sns.it> On Wednesday 01 October 2003 11:16, Andrea Ferretti wrote: > [...] I always found that a ghost appears if I use lloc=2, desappears > with lloc=1 and also in this better case I'm not able to reproduce 7s > and 6d atomic eigenvalues (Pt has a 6s1 6p0 5d9 ground state > configuration).... try also lloc=0 > I'm not experienced with pseudopotential generation, so I 'm wondering if > this behaviour makes the pseudo completely unreliable or can be accepted > after a deep check the configuration you used in the check is rather extreme (all 5d electrons are missing) so expect some loss of accuracy in the Kleinman-Bylander projection. You can have a look at logarithmic derivatives in order to have a better feeling of how good or bad your pseudopotential is Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From neuber at itp.tu-graz.ac.at Wed Oct 1 15:41:45 2003 From: neuber at itp.tu-graz.ac.at (Danilo Neuber) Date: Wed, 1 Oct 2003 15:41:45 +0200 (CEST) Subject: [Pw_forum] PP Generation/Conversion for Si In-Reply-To: <200310011514.31147.giannozz@nest.sns.it> References: <200310011514.31147.giannozz@nest.sns.it> Message-ID: On Wed, 1 Oct 2003, Paolo Giannozzi wrote: > actually you can use the pseudopotential file produced by "trou" > with no further conversion. If you want to convert it to upf format > (which I recommand), use "ncpp2upf". Converting twice as you > did should also work, though, but the probability of making a > mistake increases at each conversion ... Both the original PP file from 'trou' as well as the file converted using ncpp2upf work fine for me - thank you very much! Danilo -- Dipl.-Ing. Danilo Neuber Technische Universit?t Graz Telefon: +43 / 316 / 873 - 8188 Institut f?r Theoretische Physik Telefax: +43 / 316 / 873 - 8678 Petersgasse 16, A-8010 Graz From ferretti.andrea at unimore.it Wed Oct 1 15:57:52 2003 From: ferretti.andrea at unimore.it (Andrea Ferretti) Date: Wed, 1 Oct 2003 15:57:52 +0200 (CEST) Subject: [Pw_forum] pseudopotential generation In-Reply-To: <200310011525.25077.giannozz@nest.sns.it> Message-ID: On Wed, 1 Oct 2003, Paolo Giannozzi wrote: > On Wednesday 01 October 2003 11:16, Andrea Ferretti wrote: > > [...] I always found that a ghost appears if I use lloc=2, desappears > > with lloc=1 and also in this better case I'm not able to reproduce 7s > > and 6d atomic eigenvalues (Pt has a 6s1 6p0 5d9 ground state > > configuration).... > > try also lloc=0 > > > I'm not experienced with pseudopotential generation, so I 'm wondering if > > this behaviour makes the pseudo completely unreliable or can be accepted > > after a deep check > > the configuration you used in the check is rather extreme (all 5d electrons > are missing) so expect some loss of accuracy in the Kleinman-Bylander > projection. You can have a look at logarithmic derivatives in order to have > a better feeling of how good or bad your pseudopotential is > > Paolo > thanks Paolo, the configuration was inteded to be 5d9 and not 5d0...... now it seems to well reproduce atomic energies thanks again andrea From lucre at ms1.hinet.net Wed Oct 1 19:56:44 2003 From: lucre at ms1.hinet.net (Lu Frederic) Date: Thu, 2 Oct 2003 01:56:44 +0800 Subject: [Pw_forum] Negative phonon frequencies? Message-ID: <000c01c38845$61c1fd10$ef3310ac@phys> Dear PWscf users, While calculating phonon frequencies at Gamma of zb/wz GaN using PWscf_v1.2.0, no matter what Ga 3d core electrons treat as valence electrons. I am always found negative frequencies in the output file. I guess that I am not the first people to ask why? but I don't know how to search old Pw_forum digest. If someone really know what's volume and number have talk about negative phonon frequencies, please tell me. Thank you Lu Fu-Fa, 10/02/2003 From eyvaz_isaev at yahoo.com Wed Oct 1 22:14:41 2003 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Wed, 1 Oct 2003 13:14:41 -0700 (PDT) Subject: [Pw_forum] Negative phonon frequencies? In-Reply-To: <000c01c38845$61c1fd10$ef3310ac@phys> Message-ID: <20031001201441.43085.qmail@web60304.mail.yahoo.com> Dear Fredric, Probably negative frequencies are due to lack of the ASR not imposed for dynamic matrix calculations at the \Gamma point. But they should not be too large. For example, as it is mentioned in the PWSCF manual (troubleshooting) it might be around 50cm^{-1} or 1.6THz. Nevertheless, you have to be careful about your input file. I remember, there was a mail from Gothenborg (Sweden) where negative frequences for ZnS were obtained due to an error in input file. I suggest, it depends also on the pseudopotential quality. Regards, Eyvaz. --- Lu Frederic wrote: > Dear PWscf users, > > While calculating phonon frequencies at Gamma of > zb/wz GaN using > PWscf_v1.2.0, > no matter what Ga 3d core electrons treat as valence > electrons. I am always > found > negative frequencies in the output file. > > I guess that I am not the first people to ask why? > but I don't know how to > search > old Pw_forum digest. If someone really know what's > volume and number have > talk about negative phonon frequencies, please tell > me. > > Thank you > > Lu Fu-Fa, 10/02/2003 > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com From Steven.Kirk at htu.se Wed Oct 1 22:27:14 2003 From: Steven.Kirk at htu.se (Steven Kirk) Date: Wed, 01 Oct 2003 22:27:14 +0200 Subject: [Pw_forum] Bader analysis, part II Message-ID: <3F7B38A2.9000003@htu.se> Hello all, Firstly, thank you for your responses to my previous enquiry regarding Bader analysis of the results produced by PWSCF. I am strongly considering writing some code to add this capability to, for example, the PP program from PWSCF. I am primarily interested in point properties rather than integration over atomic basins: finding the critical points and bond paths would be a good start, together with the values of the charge density, Laplacian etc. at these points. Implementing basin definition and integration is a much more demanding task ! In order to find, for example, the bond critical points (BCPs), one could assume a starting point halfway between 2 neighbouring atoms, then evaluate grad(rho) (the gradient of the charge density) at this point, and take small steps 'downhill' to exactly locate the BCP (where grad(rho) =0 ). This can be done using Newton-Raphson in 3D - I once implemented this in another code - you need to extract the Hessian matrix of the charge density. If the wave function is represented by plane waves of the form exp( ikx), then weighting each of the plane waves with ik should allow one to extract grad(rho), and weighting with -k^2 should allow one to evaluate the Laplacian, at an arbitrary location x. Using plane waves has advantages over relatively crude 3D interpolations of gridded data in this case. A good starting point, then, would be the ability to specify two atoms, find the BCP (if any) that exists between them, and extract the charge density and the Laplacian at the BCP. The bond path (the path linking the two atoms along which the charge density is maximal relative on any neigbouring path) would also be useful - sometimes these are *not* the expected straight lines ! I would be very grateful if any of the participants on this list or the developers can point me to an appropriate location in the code where it would be possible to 'splice in' this capability. Many thanks in advance for any feedback on this topic, Regards, Steve Kirk From pushpa at jncasr.ac.in Thu Oct 2 08:44:29 2003 From: pushpa at jncasr.ac.in (Raghani Pushpa) Date: Thu, 2 Oct 2003 12:14:29 +0530 (IST) Subject: [Pw_forum] Negative phonon frequencies? Message-ID: Dear Fredric, I was also getting the negative phonon frequencies at gamma pt for beta tin structure but then I found that I was using very coarse mesh for k pts for BZ integrations. When I increased the mesh size, the negative frequencies disappeared. You could try increasing the k pt mesh size. Pushpa -- From eyvaz_isaev at yahoo.com Fri Oct 3 10:45:35 2003 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 3 Oct 2003 01:45:35 -0700 (PDT) Subject: [Pw_forum] PWSCF_1.3 on ALPHA clusters In-Reply-To: Message-ID: <20031003084535.79311.qmail@web60304.mail.yahoo.com> Dear PWSCF users, I seems I have managed to work PWSCF_1.3 on ALPHA cluster. At least, all tests in /examples2 were passed properly. There is a little question with the running of the whole script and hope too fix it soon. Details of compilation will be available soon, too. Regards, Eyvaz. __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com From proffess at yandex.ru Fri Oct 3 11:02:17 2003 From: proffess at yandex.ru (Sergei Lisenkov) Date: Fri, 3 Oct 2003 13:02:17 +0400 (MSD) Subject: [Pw_forum] PWscf-1.3 on Compaq ALPHA clusters under Linux In-Reply-To: <20031003084535.79311.qmail@web60304.mail.yahoo.com> References: <20031003084535.79311.qmail@web60304.mail.yahoo.com> Message-ID: <3F7D3B19.000005.26461@pantene.yandex.ru> Dear PWscf authors and users, I think we should to separate make.sys files for DEC Alpha and Compaq Alpha machines. I using the last machines working under Linux. I have seen the Makefile provided by script configure does nor work on Compaq Alpha machines. I have seen the errors message and done the minor changes in file clib/cp.h (add one underscore): #if defined __ALPHA && defined __LINUX # define FFTW_INPLACE_DRV_1D fftw_inplace_drv_1d__ # define FFTW_INPLACE_DRV_2D fftw_inplace_drv_2d__ # define FFTW_INPLACE_DRV_3D fftw_inplace_drv_3d__ # define CREATE_PLAN_1D create_plan_1d__ # define DESTROY_PLAN_1D destroy_plan_1d__ # define CREATE_PLAN_2D create_plan_2d__ # define CREATE_PLAN_3D create_plan_3d__ # define DESTROY_PLAN_2D destroy_plan_2d__ # define DESTROY_PLAN_3D destroy_plan_3d__ # define FFT_X_STICK fft_x_stick__ # define FFT_XY_STICK fft_xy_stick_ # define FFT_XY fft_xy_ # define FFT_Y_STICK fft_y_stick__ # define FFT_Y_STICK2 fft_y_stick2_ # define FFT_Z_STICK fft_z_stick__ # define FFT_Z fft_z_ # define FFT_STICK fft_stick_ # define CP_GETENV cp_getenv_ # define CP_DATE cp_date__ # define CPFLUSH cpflush_ # define CPTIMER cptimer_ # define ELAPSED_SECONDS elapsed_seconds__ # define CCLOCK cclock_ # define FACTOR235 factor235_ # define FACTOR2 factor2_ # define LN_ALLOC ln_alloc__ # define LN_DEALLOC ln_dealloc__ # define LN_SET ln_set__ # define LN_ACTIVATE ln_activate__ # define LN_IND ln_ind__ # define MEMSTAT memstat_ # define READOCC readocc_ # define ROUND2 round2_ # define MYUNITNAME myunitname_ # define CP_ITOA cp_itoa_ #endif and redefine CPPFLAGS : CPPFLAGS = -D__ALPHA -D__LINUX ..... The compiler is "fort", i.e. F90=fort F77=fort Bets wishes, Sergey From proffess at yandex.ru Fri Oct 3 11:25:42 2003 From: proffess at yandex.ru (Sergei Lisenkov) Date: Fri, 3 Oct 2003 13:25:42 +0400 (MSD) Subject: [Pw_forum] PWscf-1.3 on HP-PA RISC machines In-Reply-To: <3F7D3B19.000005.26461@pantene.yandex.ru> References: <20031003084535.79311.qmail@web60304.mail.yahoo.com> <3F7D3B19.000005.26461@pantene.yandex.ru> Message-ID: <3F7D4096.000006.27509@pantene.yandex.ru> Dear PWscf authors, users, and Paolo, I would like to add some changes for succesfull compilation of new version of PWscf-1.3 on HP machines. Paolo, you was right in manual, that old make.sys file for version 1.2.0 does not work for version 1.3.0. I did some changes and it works. We should add in file clib/cp.h the next : #if defined __HP # define FFTW_INPLACE_DRV_1D fftw_inplace_drv_1d # define FFTW_INPLACE_DRV_2D fftw_inplace_drv_2d # define FFTW_INPLACE_DRV_3D fftw_inplace_drv_3d # define CP_GETENV cp_getenv # define CP_DATE cp_date # define CREATE_PLAN_2D create_plan_2d # define CREATE_PLAN_3D create_plan_3d # define CREATE_PLAN_1D create_plan_1d # define DESTROY_PLAN_1D destroy_plan_1d # define DESTROY_PLAN_2D destroy_plan_2d # define DESTROY_PLAN_3D destroy_plan_3d # define FFT_X_STICK fft_x_stick # define FFT_XY_STICK fft_xy_stick # define FFT_XY fft_xy # define FFT_Y_STICK fft_y_stick # define FFT_Y_STICK2 fft_y_stick2 # define FFT_Z_STICK fft_z_stick # define FFT_Z fft_z # define FFT_STICK fft_stick # define CPFLUSH cpflush # define CPTIMER cptimer # define ELAPSED_SECONDS elapsed_seconds # define CCLOCK cclock # define FACTOR235 factor235 # define FACTOR2 factor2 # define LN_ALLOC ln_alloc # define LN_DEALLOC ln_dealloc # define LN_SET ln_set # define LN_ACTIVATE ln_activate # define LN_IND ln_ind # define MEMSTAT memstat # define READOCC readocc # define ROUND2 round2 # define MYUNITNAME myunitname # define CP_ITOA cp_itoa and add -D__HP in CPPFLAGS. Now I present two separate files for Serial and Parallel version of PWscf-1.3.0 : Make.hp: ============================================================================================ OSHOME=/nethome/proffess/O-sesame # # System-dependent definitions for HP-PA RISC machines, HP-UX # Contributed by Sergey Lisenkov, with help from Eyvaz Isaev # and Joey Dieckhans. Software configuration: # HP-UX >=11.0, HP f90 compiler v >= 2.4, HP C compiler, MLIB # # Note that there is one version of MPI that requires the argument # "-stdio=i" in comman line for reading from standart output # # # Use precompiled fftw library (version <= 2.1.5, NOT v.3!) # In this case, specify also how to load the fftw library (FFTW_LIB) # and the path to the fftw.h include file (FFTW_INC_DIR). Example: # 32-bit version FFTW_LIB=-L/nethome/proffess/lib/fftw-2.1.5/32/lib -lfftw FFTW_INC_DIR=/nethome/proffess/lib/fftw-2.1.5/32/include # 64-bit version # FFTW_LIB=-L/nethome/proffess/lib/fftw-2.1.5/64/lib -lfftw # FFTW_INC_DIR=/nethome/proffess/lib/fftw-2.1.5/64/include CPPFLAGS = -I$(OSHOME)/include -D__HP -D__FFTW -I$(FFTW_INC_DIR) # # Use the local copy of fftw #CPPFLAGS = -I$(OSHOME)/include -D__HP -D__FFTW -D__USE_INTERNAL_FFTW # Fortran compiler: # F90 = f90 F77 = $(F90) CC = cc # # Fortran compiler flags # 32-bit FFLAGS = +O2 -w +cpp=yes +Odataprefetch +Onolimit +r8 +U77 +DA2.0 +DS2.0 # 64-bit #FFLAGS = +O2 -w +cpp=yes +Odataprefetch +Onolimit +r8 +U77 +DA2.0W +DS2.0W +noppu F90FLAGS = $(FFLAGS) $(CPPFLAGS) # # Fortran77 compiler flags # F77FLAGS = $(FFLAGS) F77FLAGS_NOOPT = -O0 # # C compiler flags # 32-bit CCFLAGS = $(CPPFLAGS) +O2 +DA2.0 +DS2.0 +DD32 +Odataprefetch +Onolimit # 64-bit #CCFLAGS = $(CPPFLAGS) +O2 +DA2.0W +DS2.0W +DD64 +Odataprefetch +Onolimit # This is needed to tell the compiler where modules are # Serial version MODULEFLAG= -I$(OSHOME)/Modules -I$(OSHOME)/PW -I$(OSHOME)/PH -I/opt/mpi/include # # Libraries: # 32-bit LIBS = -L/opt/mpi/lib/pa2.0 -lmpi -ldmpi -lmpio -lmtmpi -L/opt/mlib/lib/pa2.0 -lveclib -llapack $(FFTW_LIB) -lm # 64-bit #LIBS = -L/opt/mpi/lib/pa20_64 -lmpi -ldmpi -lmpio -lmtmpi -L/opt/mlib/lib/pa20_64 -lveclib -llapack $(FFTW_LIB) -lm # Loader: LD=$(F90) #32-bit LDFLAGS = $(OSHOME)/flib/ptools.a $(OSHOME)/flib/flib.a $(OSHOME)/clib/clib.a +U77 +DA2.0 +DS2.0 +O2 $(LIBS) # 64-bit #LDFLAGS = $(OSHOME)/flib/ptools.a $(OSHOME)/flib/flib.a $(OSHOME)/clib/clib.a +U77 +DA2.0W +DS2.0W +O2 $(LIBS) # # ar: # AR = ar ARFLAGS = ruv ========================================================================================== Make.hpMPI ========================================================================================== OSHOME=/nethome/proffess/O-sesame # # System-dependent definitions for HP-PA RISC machines, HP-UX # Contributed by Sergey Lisenkov, with help from Eyvaz Isaev # and Joey Dieckhans. Software configuration: # HP-UX >=11.0, HP f90 compiler v >= 2.4, HP C compiler, MLIB # # Note that there is one version of MPI that requires the argument # "-stdio=i" in comman line for reading from standart output # # # Use precompiled fftw library (version <= 2.1.5, NOT v.3!) # In this case, specify also how to load the fftw library (FFTW_LIB) # and the path to the fftw.h include file (FFTW_INC_DIR). Example: # 32-bit version FFTW_LIB=-L/nethome/proffess/lib/fftw-2.1.5/32/lib -lfftw FFTW_INC_DIR=/nethome/proffess/lib/fftw-2.1.5/32/include # 64-bit version # FFTW_LIB=-L/nethome/proffess/lib/fftw-2.1.5/64/lib -lfftw # FFTW_INC_DIR=/nethome/proffess/lib/fftw-2.1.5/64/include CPPFLAGS = -I$(OSHOME)/include -D__HP -D__FFTW -D__PARA -D__MPI -I$(FFTW_INC_DIR) # # Use the local copy of fftw #CPPFLAGS = -I$(OSHOME)/include -D__HP -D__FFTW -D__PARA -D__MPI -D__USE_INTERNAL_FFTW # Fortran compiler: # F90 = f90 F77 = $(F90) CC = mpicc # # Fortran compiler flags # 32-bit FFLAGS = +O2 -w +cpp=yes +Odataprefetch +Onolimit +r8 +U77 +DA2.0 +DS2.0 # 64-bit #FFLAGS = +O2 -w +cpp=yes +Odataprefetch +Onolimit +r8 +U77 +DA2.0W +DS2.0W +noppu F90FLAGS = $(FFLAGS) $(CPPFLAGS) # # Fortran77 compiler flags # F77FLAGS = $(FFLAGS) F77FLAGS_NOOPT = -O0 # # C compiler flags # 32-bit CCFLAGS = $(CPPFLAGS) +O2 +DA2.0 +DS2.0 +DD32 +Odataprefetch +Onolimit # 64-bit #CCFLAGS = $(CPPFLAGS) +O2 +DA2.0W +DS2.0W +DD64 +Odataprefetch +Onolimit # This is needed to tell the compiler where modules are # Serial version MODULEFLAG= -I$(OSHOME)/Modules -I$(OSHOME)/PW -I$(OSHOME)/PH # # Libraries: # 32-bit #LIBS = -L/opt/mlib/lib/pa2.0 -lveclib -llapack $(FFTW_LIB) -lm # 64-bit #LIBS = -L/opt/mlib/lib/pa20_64 -lveclib -llapack $(FFTW_LIB) -lm # Loader: LD=$(F90) # 32-bit LDFLAGS = $(OSHOME)/flib/ptools.a $(OSHOME)/flib/flib.a $(OSHOME)/clib/clib.a +U77 +DA2.0 +DS2.0 +O2 $(LIBS) -lpthread # 64-bit #LDFLAGS = $(OSHOME)/flib/ptools.a $(OSHOME)/flib/flib.a $(OSHOME)/clib/clib.a +U77 +DA2.0W +DS2.0W +O2 $(LIBS) -lpthread # ar: # AR = ar ARFLAGS = ruv =============================================================================================== Important notice: You must change in file PW/openfil.f90 "iunigk = 7" to "iunigk = 77". Also you must use ONLY old version of Pseudopotentials (not UPF format). I do not know why. Thanks, Sergey From giannozz at nest.sns.it Fri Oct 3 19:31:53 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 3 Oct 2003 19:31:53 +0200 Subject: [Pw_forum] PWscf-1.3 on HP-PA RISC machines In-Reply-To: <3F7D4096.000006.27509@pantene.yandex.ru> References: <20031003084535.79311.qmail@web60304.mail.yahoo.com> <3F7D3B19.000005.26461@pantene.yandex.ru> <3F7D4096.000006.27509@pantene.yandex.ru> Message-ID: <200310031931.53500.giannozz@nest.sns.it> On Friday 03 October 2003 11:25, Sergei Lisenkov wrote: > I would like to add some changes for succesfull compilation of > new version of PWscf-1.3 on HP machines. Your changes have been added to the current version.Thank you Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From proffess at yandex.ru Sun Oct 5 14:57:12 2003 From: proffess at yandex.ru (Sergei Lisenkov) Date: Sun, 5 Oct 2003 16:57:12 +0400 (MSD) Subject: [Pw_forum] error in rdiaghg Message-ID: <3F801528.000002.09885@tide.yandex.ru> Dear PWscf users, I got the error after 68 steps. How to solve it? ..... Entering Dynamics; it = 68 time = 0.00324 pico-seconds new lattice vectors (alat unit) : 1.000195559 -0.000001730 0.000000653 -0.000001730 1.000165067 0.000000632 0.000000653 0.000000632 1.000186077 new unit-cell volume = 9266.0639 (a.u.)^3 new positions in cryst coord C -0.071238291 0.306426380 -0.376621113 C 0.033239385 0.271946752 -0.439817355 ... Ekin = 0.77391739 Ryd T = 325.1 K Etot = -750.51524601 NEW-OLD atomic charge density approx. for the potential NEW K-POINTS 0.0000000 0.0000000 0.0000000 2.0000000 total cpu time spent up to now is 8115.42 secs iteration # 1 ecut= 35.00 ryd beta=0.10 Davidson diagonalization with overlap ethr = 9.43E-06, avg # of iterations = 14.0 total energy = -751.19329655 ryd estimated scf accuracy < 0.14888871 ryd total cpu time spent up to now is 8137.06 secs iteration # 2 ecut= 35.00 ryd beta=0.10 Davidson diagonalization with overlap ethr = 5.64E-05, avg # of iterations = 4.0 total energy = -751.24683236 ryd estimated scf accuracy < 0.06564089 ryd total cpu time spent up to now is 8152.35 secs iteration # 3 ecut= 35.00 ryd beta=0.10 Davidson diagonalization with overlap ethr = 2.49E-05, avg # of iterations = 6.0 total energy = -751.28577433 ryd estimated scf accuracy < 0.01460215 ryd total cpu time spent up to now is 8178.24 secs iteration # 4 ecut= 35.00 ryd beta=0.10 Davidson diagonalization with overlap %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from rdiaghg : error # 716 info =/= 0 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... Thanks for any help, Best wishes, Sergey From eyvaz_isaev at yahoo.com Sun Oct 5 21:39:26 2003 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Sun, 5 Oct 2003 12:39:26 -0700 (PDT) Subject: [Pw_forum] error in rdiaghg In-Reply-To: <3F801528.000002.09885@tide.yandex.ru> Message-ID: <20031005193926.60321.qmail@web60307.mail.yahoo.com> Dear Sergei, You can try to change the diagonalization key, i.e. "cg" instead of davidson. Regards, Eyvaz. --- Sergei Lisenkov wrote: > Dear PWscf users, > > I got the error after 68 steps. How to solve it? > > ..... > Entering Dynamics; it = 68 time = 0.00324 > pico-seconds > > new lattice vectors (alat unit) : > 1.000195559 -0.000001730 0.000000653 > -0.000001730 1.000165067 0.000000632 > 0.000000653 0.000000632 1.000186077 > new unit-cell volume = 9266.0639 (a.u.)^3 > new positions in cryst coord > C -0.071238291 0.306426380 -0.376621113 > C 0.033239385 0.271946752 -0.439817355 > ... > Ekin = 0.77391739 Ryd T = 325.1 K Etot = > -750.51524601 > > NEW-OLD atomic charge density approx. for the > potential > NEW K-POINTS > 0.0000000 0.0000000 0.0000000 2.0000000 > > total cpu time spent up to now is 8115.42 > secs > > iteration # 1 ecut= 35.00 ryd > beta=0.10 > Davidson diagonalization with overlap > ethr = 9.43E-06, avg # of iterations = 14.0 > > total energy = -751.19329655 ryd > estimated scf accuracy < 0.14888871 ryd > > total cpu time spent up to now is 8137.06 > secs > > iteration # 2 ecut= 35.00 ryd > beta=0.10 > Davidson diagonalization with overlap > ethr = 5.64E-05, avg # of iterations = 4.0 > > total energy = -751.24683236 ryd > estimated scf accuracy < 0.06564089 ryd > > total cpu time spent up to now is 8152.35 > secs > > iteration # 3 ecut= 35.00 ryd > beta=0.10 > Davidson diagonalization with overlap > ethr = 2.49E-05, avg # of iterations = 6.0 > > total energy = -751.28577433 ryd > estimated scf accuracy < 0.01460215 ryd > > total cpu time spent up to now is 8178.24 > secs > > iteration # 4 ecut= 35.00 ryd > beta=0.10 > Davidson diagonalization with overlap > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from rdiaghg : error # 716 > info =/= 0 > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > stopping ... > > Thanks for any help, > Best wishes, > Sergey > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com From giannozz at nest.sns.it Mon Oct 6 22:55:08 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 6 Oct 2003 22:55:08 +0200 Subject: [Pw_forum] error in rdiaghg In-Reply-To: <20031005193926.60321.qmail@web60307.mail.yahoo.com> References: <20031005193926.60321.qmail@web60307.mail.yahoo.com> Message-ID: <200310062255.09037.giannozz@nest.sns.it> On Sunday 05 October 2003 21:39, Eyvaz Isaev wrote: > Dear Sergei, > > You can try to change the diagonalization key, i.e. > "cg" instead of davidson. I don't think he can: "rdiaghg" is used in the Gamma-only version of the code, which uses the trick to perform two FFTs at the same time. Since the conjugate-gradient type diagonalization operates on one band at the time, it cannot work with two FFTs (on two bands) at the same time, unless some nontrivial change is made to the algorithm (everybody is welcome to try). Coming to the original question: in the manual there are some possible reasons for the similar error in routine "cdiaghg", which apply as well to this case. It seems to me however that the case shown is a case of failure of the lapack diagonalization routines used in *diaghg. Whether this is a problem of lapack (unlikely) or a problem of a specific machine-dependent implementation of lapack (more likely) or a compiler problem or some subtle problem in the code (typically, memory overrun), it is hard to say. Some time ago I spent several hours trying to pinpoint the origin of a similar problem with cdiaghg on a sp3, to no avail. So, the only advice I can offer is to restart the calculation from where it crashed and see what happens Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From giannozz at nest.sns.it Wed Oct 8 12:00:22 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 8 Oct 2003 12:00:22 +0200 Subject: [Pw_forum] Bader analysis, part II In-Reply-To: <3F7B38A2.9000003@htu.se> References: <3F7B38A2.9000003@htu.se> Message-ID: <200310081200.22912.giannozz@nest.sns.it> On Wednesday 01 October 2003 22:27, Steven Kirk wrote: > Firstly, thank you for your responses to my previous enquiry regarding > Bader analysis of the results produced by PWSCF. > > I am strongly considering writing some code to add this capability to, > for example, the PP program from PWSCF. consider yourself strongly encouraged. Once upon a time somebody tried to implement some Bader-analysis stuff. I heard that one problem specific to pseudopotentials is that the assumption that the charge density has maxima only on atoms is no longer true. I don't think this is a major problem, though. Also the presence of the augmentation charge in the ultrasoft case shouldn't be a problem. > I would be very grateful if any of the participants on this list or the > developers can point me to an appropriate location in the code > where it would be possible to 'splice in' this capability. "where", among the postprocessing codes. Start from a template like the one attached, add a line to PP/Makefile, for instance: bader.x: $(PPOBJS) bader.o $(LD) -o baders.x bader.o $(PPOBJS) $(PWOBJS) $(MODULES) $(LDFLAGS) (the first character is a tabulator!!!) "how", it's another story. The meaning of the relevant variables is (should be) described (sort of) in PW/pwcom.f90; other info is spread throughout the code. Try the "monkey style": look how things are done in the code and imitate!!! Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 -------------- next part -------------- ! !----------------------------------------------------------------------- program bader !----------------------------------------------------------------------- use pwcom use becmod use io_files, only: nd_nmbr, prefix, tmp_dir #ifdef __PARA use para, only: me use io_global, only: ionode_id use mp, only: mp_bcast #endif implicit none ! character(len=256) :: outdir integer :: ios namelist / inputpp / outdir, prefix ! call start_postproc (nd_nmbr) ! ! set default values for variables in namelist ! prefix = 'pwscf' outdir = './' ! #ifdef __PARA if (me == 1) then #endif read (5, inputpp, err = 200, iostat = ios) 200 call errore ('bader', 'reading inputpp namelist', abs (ios) ) ! tmp_dir = trim(outdir) ! #ifdef __PARA end if ! ! ... Broadcast variables ! CALL mp_bcast( tmp_dir, ionode_id ) CALL mp_bcast( prefix, ionode_id ) #endif ! ! Now allocate space for pwscf variables, read and check them. ! call read_file call openfil call init_us_1 ! call do_what_i_need ! call stop_pp stop end program bader ! !----------------------------------------------------------------------- subroutine do_what_i_need !----------------------------------------------------------------------- ! ! This routine does this and that ! #include "machine.h" (may or may not be needed, it makes no harm) #ifdef __PARA use para, only: me #endif use pwcom implicit none ... local variables here ...calculate ... #ifdef __PARA if (me == 1) then #endif ... write results to output ... #ifdef __PARA endif #endif ! return end subroutine do_what_i_need From giannozz at nest.sns.it Wed Oct 8 18:00:07 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 8 Oct 2003 18:00:07 +0200 Subject: [Pw_forum] PP Generation/Conversion for Si In-Reply-To: References: Message-ID: <200310081800.07029.giannozz@nest.sns.it> On Wednesday 01 October 2003 15:03, Danilo Neuber wrote: > I have tried to create a Si pseudopotenial with Paolo Giannozzi's > programs ld1/trou (accepting the suggested cutoff radii) and converted > it using ps2cp and cpmd2upf (pwscf v1.3; I've attached shell scripts > with input parameters). [...] What's going wrong? oops, CPMD pseudopotentials are in Hartree, UPF in Ry, there is a factor 2 missing in cpmd2upf in the local potential... thank you for bringing this to our attention Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From kcolak at gazi.edu.tr Thu Oct 9 08:54:11 2003 From: kcolak at gazi.edu.tr (=?iso-8859-9?B?S2VtYWwgx29sYWtv8Gx1?=) Date: Thu, 9 Oct 2003 09:54:11 +0300 Subject: [Pw_forum] (no subject) Message-ID: <002601c38e32$2512bd80$6200060a@gazi.edu.tr> Dear PWscf users, can PWscf(v.1.3) do elastic constants (Cij) calculations ? Thanks in advance for any answer. Kemal -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20031009/c0af404f/attachment.htm From giannozz at nest.sns.it Sat Oct 11 18:19:13 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sat, 11 Oct 2003 18:19:13 +0200 Subject: [Pw_forum] elastic constants In-Reply-To: <002601c38e32$2512bd80$6200060a@gazi.edu.tr> References: <002601c38e32$2512bd80$6200060a@gazi.edu.tr> Message-ID: <200310111819.13456.giannozz@nest.sns.it> On Thursday 09 October 2003 08:54, Kemal ?olako?lu wrote: > can PWscf(v.1.3) do elastic constants (Cij) calculations ? with stress calculations at finite displacements, yes. Within linear response [ PRL 59, 2662 (1987) ], no Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From ginzam at hotmail.com Mon Oct 13 22:53:33 2003 From: ginzam at hotmail.com (ZAM GIN) Date: Mon, 13 Oct 2003 20:53:33 +0000 Subject: [Pw_forum] question!!! Message-ID: i would like know how can i use the post processing projdos. I appreciate if you help me. Thank's. Zam _________________________________________________________________ MSN Fotos: la forma m?s f?cil de compartir e imprimir fotos. http://photos.msn.es/support/worldwide.aspx From francesco.antoniella at aquila.infn.it Tue Oct 14 09:55:08 2003 From: francesco.antoniella at aquila.infn.it (Francesco Antoniella) Date: 14 Oct 2003 09:55:08 +0200 Subject: [Pw_forum] question!!! In-Reply-To: References: Message-ID: <1066118118.1457.2.camel@altair> Il lun, 2003-10-13 alle 22:53, ZAM GIN ha scritto: > i would like know how can i use the post processing projdos. > I appreciate if you help me. Thank's. > > Zam Hi Zam, wich version of the code are you using? Maybe 1.1.2? If yes you will have projwfc.x and projdos programs in the package. If you send me more info on your code version I will sketch you the procedure to obtain the partial DOS with projdos. Wishes Francesco > > _________________________________________________________________ > MSN Fotos: la forma m?s f?cil de compartir e imprimir fotos. > http://photos.msn.es/support/worldwide.aspx > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From degironc at sissa.it Wed Oct 15 09:09:55 2003 From: degironc at sissa.it (Stefano de Gironcoli) Date: Wed, 15 Oct 2003 09:09:55 +0200 Subject: [Pw_forum] question!!! References: Message-ID: <3F8CF2C3.4040009@sissa.it> An example of use for program projwfc (that produces atomic-projected partial DOS) is given in example8. If io_choice is set to "files" or "both" (default vaule) it produces a number of files (one for each atomic wavefunction in the pseudopotential of each atom in the cell) that contains the projected dos on the (2l+1) angular momentum components, or 2*(2l+1) in case of spin polarized calculation, for the given atomic wfc of the given atom. stefano ZAM GIN wrote: > i would like know how can i use the post processing projdos. > I appreciate if you help me. Thank's. > > Zam > From ylyj2000 at yahoo.com Thu Oct 16 22:25:29 2003 From: ylyj2000 at yahoo.com (Li Yu) Date: Thu, 16 Oct 2003 13:25:29 -0700 (PDT) Subject: [Pw_forum] a problem about the output of example2 Message-ID: <20031016202529.13919.qmail@web9701.mail.yahoo.com> Hi, Every one, I am a beginner of PWscf. I have a problem about the output file of pw_example/example2. In file "si.phG.out", I can find the dielectric tensor of silicon. I wonder if it is the high frequency one, (I mean if the is only the electron contribution), or it includes the ions contribution. Thanks. Sincerely Li Yang __________________________________ Do you Yahoo!? The New Yahoo! Shopping - with improved product search http://shopping.yahoo.com From giannozz at nest.sns.it Thu Oct 16 22:38:55 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 16 Oct 2003 22:38:55 +0200 (CEST) Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: <20031016202529.13919.qmail@web9701.mail.yahoo.com> Message-ID: On Thu, 16 Oct 2003, Li Yu wrote: > [...] I can find the dielectric tensor of silicon. > I wonder if it is the high frequency one, (I mean > if the is only the electron contribution) it is > or it includes the ions contribution it doesn't Paolo From felipe at titan.ipicyt.edu.mx Fri Oct 17 01:32:47 2003 From: felipe at titan.ipicyt.edu.mx (FELIPE VALENCIA) Date: Thu, 16 Oct 2003 18:32:47 -0500 (CDT) Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: Message-ID: Dear Paolo, I am a little confused. As long as I knew, pwscf only dealt with the phonon contribution to the dielectric response, but it seems, from your reply, that there have been major changes and now you can study electronic contributions... ?Is that true or, perhaps, I just misunderstood your reply? Felipe > On Thu, 16 Oct 2003, Li Yu wrote: > > > [...] I can find the dielectric tensor of silicon. > > I wonder if it is the high frequency one, (I mean > > if the is only the electron contribution) > > it is > > > or it includes the ions contribution > > it doesn't > > Paolo > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From baroni at sissa.it Thu Oct 16 18:38:04 2003 From: baroni at sissa.it (Stefano Baroni) Date: Thu, 16 Oct 2003 18:38:04 +0200 Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: <20031016202529.13919.qmail@web9701.mail.yahoo.com> References: <20031016202529.13919.qmail@web9701.mail.yahoo.com> Message-ID: <3F8EC96C.9010501@sissa.it> You can find the answer in any solid-state physics textbook. What is the ionic contribution to the dielectric constant of Silicon? ;-) Stefano Baroni Li Yu wrote: > Hi, Every one, > I am a beginner of PWscf. I have a problem about > the output file of pw_example/example2. In file > "si.phG.out", I can find the dielectric tensor of > silicon. I wonder if it is the high frequency one, (I > mean if the is only the electron contribution), or it > includes the ions contribution. Thanks. > > Sincerely Li Yang > > __________________________________ > Do you Yahoo!? > The New Yahoo! Shopping - with improved product search > http://shopping.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- Stefano Baroni SISSA and DEMOCRITOS National Simulation Center via Beirut 2-4, I34014 Trieste, Italy, [+39] 040 3787 406 From baroni at sissa.it Thu Oct 16 18:38:41 2003 From: baroni at sissa.it (Stefano Baroni) Date: Thu, 16 Oct 2003 18:38:41 +0200 Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: <20031016202529.13919.qmail@web9701.mail.yahoo.com> References: <20031016202529.13919.qmail@web9701.mail.yahoo.com> Message-ID: <3F8EC991.5020203@sissa.it> You can find the answer in any solid-state physics textbook. What is the ionic contribution to the dielectric constant of Silicon? ;-) Stefano Baroni Li Yu wrote: > Hi, Every one, > I am a beginner of PWscf. I have a problem about > the output file of pw_example/example2. In file > "si.phG.out", I can find the dielectric tensor of > silicon. I wonder if it is the high frequency one, (I > mean if the is only the electron contribution), or it > includes the ions contribution. Thanks. > > Sincerely Li Yang > > __________________________________ > Do you Yahoo!? > The New Yahoo! Shopping - with improved product search > http://shopping.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- Stefano Baroni SISSA and DEMOCRITOS National Simulation Center via Beirut 2-4, I34014 Trieste, Italy, [+39] 040 3787 406 From giannozz at nest.sns.it Fri Oct 17 10:03:08 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 17 Oct 2003 10:03:08 +0200 Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: References: Message-ID: <200310171003.08906.giannozz@nest.sns.it> On Friday 17 October 2003 01:32, FELIPE VALENCIA wrote: > I am a little confused. As long as I knew, pwscf only dealt with the > phonon contribution to the dielectric response, but it seems, from > your reply, that there have been major changes and now you can > study electronic contributions... there are major changes all the time :-) but PWscf has always calculated what is called "epsilon infinity", that is, the electronic contribution to the dielectric tensor at clamped ions (see: Rev. Mod. Phys. 73, 515 (2001) ). The ionic contribution can be easily added knowing the force constants and effective charges (both calculated by PWscf). Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From bungaro at physics.rutgers.edu Fri Oct 17 16:20:00 2003 From: bungaro at physics.rutgers.edu (CLAUDIA BUNGARO) Date: Fri, 17 Oct 2003 10:20:00 -0400 (EDT) Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: References: Message-ID: Dear Felipe there are a number of references in the literature about ab initio calculations of electronic + phonon contributions to the static dielectric constant that you may find helpfull. For example: Cockayne E and Burton BP PHYS REV B {\bf 62}, 3735 (2000) X. Zhao and D. Vanderbilt PRB {\bf 65}, 75105 (2002) As a book I like for example: Peter Y. Yu and Manuel Cardona "Fundamentals of Semiconductors", Springer Claudia __________________________________________________ Dr. Claudia Bungaro Dept. of Physics and Astronomy, Rutgers University 136 Frelinghuysen Road, Piscataway, NJ 08854-8019 tel: +1 732 445 4197 fax: +1 732 445 4343 email: bungaro at physics.rutgers.edu __________________________________________________ On Thu, 16 Oct 2003, FELIPE VALENCIA wrote: > > Dear Paolo, > > I am a little confused. As long as I knew, pwscf only dealt with the > phonon contribution to the > dielectric response, but it seems, from your reply, that there have > been major changes and now you can study > electronic contributions... ?Is that true or, perhaps, I just > misunderstood your reply? > > Felipe > > On Thu, 16 Oct 2003, Li Yu wrote: > > > > > [...] I can find the dielectric tensor of silicon. > > > I wonder if it is the high frequency one, (I mean > > > if the is only the electron contribution) > > > > it is > > > > > or it includes the ions contribution > > > > it doesn't > > > > 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 > > From felipe at titan.ipicyt.edu.mx Fri Oct 17 17:25:55 2003 From: felipe at titan.ipicyt.edu.mx (FELIPE VALENCIA) Date: Fri, 17 Oct 2003 10:25:55 -0500 (CDT) Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: <200310171003.08906.giannozz@nest.sns.it> Message-ID: Ok, thanks, now I have got it... I guess I was just thinking about the dielectric response (dielectric function), thanks anyway :- )). By the way : which do you think that would be the easier way to calculate electronic contributions to the dynamic response within PWSCF (i.e. without resorting to TDDFT) ? Felipe > On Friday 17 October 2003 01:32, FELIPE VALENCIA wrote: > > > I am a little confused. As long as I knew, pwscf only dealt with the > > phonon contribution to the dielectric response, but it seems, from > > your reply, that there have been major changes and now you can > > study electronic contributions... > > there are major changes all the time :-) but PWscf has always > calculated what is called "epsilon infinity", that is, the electronic > contribution to the dielectric tensor at clamped ions (see: > Rev. Mod. Phys. 73, 515 (2001) ). The ionic contribution can be > easily added knowing the force constants and effective charges > (both calculated by PWscf). > > Paolo > > -- > Paolo Giannozzi e-mail: giannozz at nest.sns.it > Scuola Normale Superiore Phone: +39/050509412 > Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 > I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From ylyj2000 at yahoo.com Sat Oct 18 18:29:03 2003 From: ylyj2000 at yahoo.com (Li Yu) Date: Sat, 18 Oct 2003 09:29:03 -0700 (PDT) Subject: [Pw_forum] Get it Message-ID: <20031018162903.14127.qmail@web9707.mail.yahoo.com> Thanks for your answers, I get it. Li yang __________________________________ Do you Yahoo!? The New Yahoo! Shopping - with improved product search http://shopping.yahoo.com From sprokuda at dnk.ru Mon Oct 20 10:19:12 2003 From: sprokuda at dnk.ru (sprokuda at dnk.ru) Date: Mon, 20 Oct 2003 12:19:12 +0400 Subject: [Pw_forum] a problem about the output of example2 In-Reply-To: <200310171003.08906.giannozz@nest.sns.it> References: <200310171003.08906.giannozz@nest.sns.it> Message-ID: <1288475953.20031020121912@dnk.ru> Dear Sirs! I've compiled PWscf.1.3 on RedHat 7.3 with ifc-7.1, mkl-6.1, and with some routimes from netlib.org blas-lapack lib. In run of the standrard examples on Si phonon calculation I got "Fortran Stop 2", and in the out file I have : Program PHONON v.1.3.0 starts ... Today is 18Oct2003 at 20:46:56 Ultrasoft (Vanderbilt) Pseudopotentials %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from phq_readin : error # 146 reading inputph namelist %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... what could it mean and how to fight with it ? Best wishes, Sergey Prokudaylo. From Steven.Kirk at htu.se Wed Oct 22 22:39:23 2003 From: Steven.Kirk at htu.se (Steven Kirk) Date: Wed, 22 Oct 2003 22:39:23 +0200 Subject: [Pw_forum] Trouble with SCF calculation in PW Message-ID: <3F96EAFB.5090202@htu.se> Hello, I'm trying to get a simple single-point calculation from the PW input file shown below. I want to use an ultrasoft potential (C.pbe-rrkjus.UPF), but when I try to use that, the program stops in cdiagh and gives me a cryptic error #131. So I switch to something simpler (C.blyp-mt.UPF) and at least the program seems to be doing SCF iterations now. But after the 11th iteration, things seem to go wrong (est. change leaps up to < 0.9 Ry, never really settles back, and the calculation runs out of iterations ). The structure should look like a slightly squashed nanotube. Can anyone tell me what I'm doing wrong ? I have taken the fractional coordinates straight from a CIF file which looks fine when I preview it. The pseudopotentials used were downloaded straight from the website. Many thanks in advance, Steve Kirk ===== PW input file follows ====== &control calculation='scf' restart_mode='from_scratch', title='Trial scf run on 8_25 structure', prefix='8_25', pseudo_dir='/home/steve/pwscf1.3/pseudo//', outdir = '/home/steve/pwscf1.3/8_25//' / &system ibrav=8, A = 14.0289, B = 11.2105, C = 4.216681, COSAB = 0, COSAC = 0, COSBC = 0, nat = 32, ntyp = 1, ecutwfc = 36.749309 / &electrons / ATOMIC_SPECIES C 12.0107 C.blyp-mt.UPF ATOMIC_POSITIONS crystal C 0.23321806 0.10541259 -0.83477178 C 0.08680611 0.21666170 -0.83347361 C -0.08680611 0.21666170 -0.83347361 C -0.23321806 0.10541259 -0.83477178 C -0.23321806 -0.10541259 -0.83477178 C -0.08680611 -0.21666170 -0.83347361 C 0.08680611 -0.21666170 -0.83347361 C 0.23321806 -0.10541259 -0.83477178 C 0.16896599 0.17979883 -0.66617051 C 0.00000000 0.22347250 -0.66665291 C -0.16896599 0.17979883 -0.66617051 C -0.25898970 0.00000000 -0.66448979 C -0.16896599 -0.17979883 -0.66617051 C 0.00000000 -0.22347250 -0.66665291 C 0.16896599 -0.17979883 -0.66617051 C 0.25898970 0.00000000 -0.66448979 C 0.16896599 0.17979883 0.66617051 C 0.00000000 0.22347250 0.66665291 C -0.16896599 0.17979883 0.66617051 C -0.25898970 0.00000000 0.66448979 C -0.16896599 -0.17979883 0.66617051 C 0.00000000 -0.22347250 0.66665291 C 0.16896599 -0.17979883 0.66617051 C 0.25898970 0.00000000 0.66448979 C 0.23321806 0.10541259 0.83477178 C 0.08680611 0.21666170 0.83347361 C -0.08680611 0.21666170 0.83347361 C -0.23321806 0.10541259 0.83477178 C -0.23321806 -0.10541259 0.83477178 C -0.08680611 -0.21666170 0.83347361 C 0.08680611 -0.21666170 0.83347361 C 0.23321806 -0.10541259 0.83477178 K_POINTS gamma ====== PW input file ends ====== From giannozz at nest.sns.it Thu Oct 23 09:02:11 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 23 Oct 2003 09:02:11 +0200 (CEST) Subject: [Pw_forum] Trouble with SCF calculation in PW In-Reply-To: <3F96EAFB.5090202@htu.se> Message-ID: On Wed, 22 Oct 2003, Steven Kirk wrote: > A = 14.0289, B = 11.2105, C = 4.216681, COSAB = 0, COSAC = 0, COSBC = 0, there is a problem with "crystallographic" input in the version of PWscf available on the web. In PW/input.f90, replace the line celldm_ (1) = a * bohr_radius_angs with celldm_ (1) = a / bohr_radius_angs (a is in A, celldm(1) in Bohr radii: the latter is printed in the output, so you may easily verify what the code is using). Alternatively, you can download the last version from CVS. Please let us know if this solves your problem Paolo Giannozzi From sprokuda at dnk.ru Thu Oct 23 09:07:55 2003 From: sprokuda at dnk.ru (sprokuda at dnk.ru) Date: Thu, 23 Oct 2003 11:07:55 +0400 Subject: [Pw_forum] a problem about the run of example2 In-Reply-To: <3F8EC991.5020203@sissa.it> References: <20031016202529.13919.qmail@web9701.mail.yahoo.com> <3F8EC991.5020203@sissa.it> Message-ID: <1376138609.20031023110755@dnk.ru> Dear Sirs! I've compiled PWscf.1.3 on RedHat 7.3 with ifc-7.1, mkl-6.1. In run of the standrard examples on Si phonon calculation I got "Fortran Stop 2", and in the out file I have : Program PHONON v.1.3.0 starts ... Today is 18Oct2003 at 20:46:56 Ultrasoft (Vanderbilt) Pseudopotentials %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from phq_readin : error # 146 reading inputph namelist %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... what could it mean and how to fight with it ? Best wishes, Sergey Prokudaylo. From giannozz at nest.sns.it Thu Oct 23 10:51:24 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 23 Oct 2003 10:51:24 +0200 Subject: [Pw_forum] a problem about the run of example2 In-Reply-To: <1376138609.20031023110755@dnk.ru> References: <20031016202529.13919.qmail@web9701.mail.yahoo.com> <3F8EC991.5020203@sissa.it> <1376138609.20031023110755@dnk.ru> Message-ID: <200310231051.25043.giannozz@nest.sns.it> On Thursday 23 October 2003 09:07, sprokuda at dnk.ru wrote: > from phq_readin : error # 146 > reading inputph namelist are you by any chance executing examples for the preceding version ? Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From goranka.bilalbegovic at zg.hinet.hr Sun Oct 26 21:52:53 2003 From: goranka.bilalbegovic at zg.hinet.hr (Goranka Bilalbegovic) Date: Sun, 26 Oct 2003 21:52:53 +0100 Subject: [Pw_forum] chdens for clusters Message-ID: <001001c39c03$53d04960$f7361dc3@gost> Hello, I am not able to plot charge densities for the cluster geometry (i.e. for free boundary conditions). This applies for both XCrysDen and gOpenMol, and for both pw.1.2.0 and pw.1.3.0. (1) gOpenMol Charge plots are OK from technical point of view, but the coordinate files produced by chdens and wanted by gOpenMol do not contain all particles and the structures are different than initial ones. (2) XCrySDen In *.xsf file, the numbers after PRIMCOORD are written as "34, 1" instead of "34 1". Using an editor to delete "," did not help. Then XCrySDen exits with "error reading DATAGRID_3D_SECTION", although this part of the file looks OK. (I use *.xsf examples of Tone on my installation without any problems.) Thanks for help. Best regards, Goranka -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20031026/26a79990/attachment.htm From giannozz at nest.sns.it Mon Oct 27 22:43:23 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 27 Oct 2003 22:43:23 +0100 Subject: [Pw_forum] chdens for clusters In-Reply-To: <001001c39c03$53d04960$f7361dc3@gost> References: <001001c39c03$53d04960$f7361dc3@gost> Message-ID: <200310272243.23992.giannozz@nest.sns.it> On Sunday 26 October 2003 21:52, Goranka Bilalbegovic wrote: > I am not able to plot charge densities for the cluster geometry (i.e. for > free boundary conditions). the code doesn't make any distinction between "cluster" and "supercell" geometry > (1) gOpenMol > Charge plots are OK from technical point of view, but the coordinate files > produced by chdens and wanted by gOpenMol do not contain all particles only atoms that are present in the visualization region are written. Is this that you do not like, or is there something else? by the way, the algorithm that determines which atoms are in the visualization region is not especially smart and may occasionally miss atoms. > and the structures are different than initial ones. you mean: after an optimization run? the code writes the final charge density and coordinates. Paolo From sprokuda at dnk.ru Tue Oct 28 14:34:33 2003 From: sprokuda at dnk.ru (sprokuda at dnk.ru) Date: Tue, 28 Oct 2003 16:34:33 +0300 Subject: [Pw_forum] ATLAS compilation In-Reply-To: <200310272243.23992.giannozz@nest.sns.it> References: <001001c39c03$53d04960$f7361dc3@gost> <200310272243.23992.giannozz@nest.sns.it> Message-ID: <14223878781.20031028163433@dnk.ru> Dear Sirs! I'm trying to compile PWscf (versions 1.3.0 and 1.2.0) on Duron-ChainTech hardware, ifc 7.1, fftw 2.3. Igot rathe strange result on the liking stage - ATLAS can't find his ownm functions. Can anybody say me something on it?. Thak you all in advance. Prokudaylo Sergey. That is what I got: /home/ATLAS/lib/Linux_ATHLONSSE1/liblapack.a(dgetrf.o): In function `dgetrf_': dgetrf.o(.text+0x55): undefined reference to `atl_f77wrap_dgetrf_' /home/ATLAS/lib/Linux_ATHLONSSE1/liblapack.a(dgetri.o): In function `dgetri_': dgetri.o(.text+0xa4): undefined reference to `atl_f77wrap_dgetri_' /home/ATLAS/lib/Linux_ATHLONSSE1/liblapack.a(ilaenv.o): In function `ilaenv_': ilaenv.o(.text+0x6ed): undefined reference to `atl_f77wrap_zgetnb_' ilaenv.o(.text+0x6fa): undefined reference to `atl_f77wrap_cgetnb_' ilaenv.o(.text+0x707): undefined reference to `atl_f77wrap_dgetnb_' ilaenv.o(.text+0x714): undefined reference to `atl_f77wrap_sgetnb_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(dscal.o): In function `dscal_': dscal.o(.text+0x18): undefined reference to `atl_f77wrap_dscal_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(dnrm2.o): In function `dnrm2_': dnrm2.o(.text+0x25): undefined reference to `atl_f77wrap_dnrm2_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(daxpy.o): In function `daxpy_': daxpy.o(.text+0x20): undefined reference to `atl_f77wrap_daxpy_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(dcopy.o): In function `dcopy_': dcopy.o(.text+0x1c): undefined reference to `atl_f77wrap_dcopy_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(ddot.o): In function `ddot_': ddot.o(.text+0x2d): undefined reference to `atl_f77wrap_ddot_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(dgemv.o): In function `dgemv_': dgemv.o(.text+0x130): undefined reference to `atl_f77wrap_dgemv_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(dgemm.o): In function `dgemm_': dgemm.o(.text+0x1f5): undefined reference to `atl_f77wrap_dgemm_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zscal.o): In function `zscal_': zscal.o(.text+0x18): undefined reference to `atl_f77wrap_zscal_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zdscal.o): In function `zdscal_': zdscal.o(.text+0x18): undefined reference to `atl_f77wrap_zdscal_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(dznrm2.o): In function `dznrm2_': dznrm2.o(.text+0x25): undefined reference to `atl_f77wrap_dznrm2_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(izamax.o): In function `izamax_': izamax.o(.text+0x1d): undefined reference to `atl_f77wrap_izamax_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zaxpy.o): In function `zaxpy_': zaxpy.o(.text+0x20): undefined reference to `atl_f77wrap_zaxpy_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zcopy.o): In function `zcopy_': zcopy.o(.text+0x1c): undefined reference to `atl_f77wrap_zcopy_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zswap.o): In function `zswap_': zswap.o(.text+0x1c): undefined reference to `atl_f77wrap_zswap_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zdotc.o): In function `zdotc_': zdotc.o(.text+0x2e): undefined reference to `atl_f77wrap_zdotc_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zdotu.o): In function `zdotu_': zdotu.o(.text+0x2e): undefined reference to `atl_f77wrap_zdotu_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zgemv.o): In function `zgemv_': zgemv.o(.text+0x130): undefined reference to `atl_f77wrap_zgemv_' /home/ATLAS/lib/Linux_ATHLONSSE1/libf77blas.a(zgemm.o): In function `zgemm_': zgemm.o(.text+0x1f5): undefined reference to `atl_f77wrap_zgemm_' From profeta at sissa.it Tue Oct 28 14:39:43 2003 From: profeta at sissa.it (Mickael Profeta) Date: Tue, 28 Oct 2003 14:39:43 +0100 Subject: [Pw_forum] ATLAS compilation In-Reply-To: <14223878781.20031028163433@dnk.ru> References: <001001c39c03$53d04960$f7361dc3@gost> <200310272243.23992.giannozz@nest.sns.it> <14223878781.20031028163433@dnk.ru> Message-ID: <3F9E719F.5060906@sissa.it> Hi, Do you compile atlas with the same compiler (ifc ?) If Atlas was compile with g77 for instance, there is certainly the problem of these underscores which are somtime added, sometime not depending on the compiler and options. I recommend compiling Atlas with the same compiler use to compile pwscf Sincerely Mickael sprokuda at dnk.ru wrote: > Dear Sirs! > I'm trying to compile PWscf (versions 1.3.0 and 1.2.0) on > Duron-ChainTech hardware, ifc 7.1, fftw 2.3. > Igot rathe strange result on the liking stage - ATLAS can't find his > ownm functions. Can anybody say me something on it?. Thak you all in > advance. > Prokudaylo Sergey. From giannozz at nest.sns.it Tue Oct 28 14:58:51 2003 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 28 Oct 2003 14:58:51 +0100 Subject: [Pw_forum] ATLAS compilation In-Reply-To: <14223878781.20031028163433@dnk.ru> References: <001001c39c03$53d04960$f7361dc3@gost> <200310272243.23992.giannozz@nest.sns.it> <14223878781.20031028163433@dnk.ru> Message-ID: <200310281458.51146.giannozz@nest.sns.it> On Tuesday 28 October 2003 14:34, sprokuda at dnk.ru wrote: > That is what I got: if you explain how you got it, maybe somebody will have a clue > /home/ATLAS/lib/Linux_ATHLONSSE1/liblapack.a(dgetrf.o): In function > `dgetrf_': dgetrf.o(.text+0x55): undefined reference to`atl_f77wrap_dgetrf_' did you load all the pieces? something like: -L"directory where ATLAS reside" -llapack -lf77blas -lcblas -latlas ? Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050509412 Piazza dei Cavalieri 7 Fax: +39/050509417, 050563513 I-56126 Pisa, Italy Office: Lab. NEST, Via della Faggiola 19 From sprokuda at dnk.ru Tue Oct 28 15:25:46 2003 From: sprokuda at dnk.ru (sprokuda at dnk.ru) Date: Tue, 28 Oct 2003 17:25:46 +0300 Subject: [Pw_forum] ATLAS compilation In-Reply-To: <3F9E719F.5060906@sissa.it> References: <001001c39c03$53d04960$f7361dc3@gost> <200310272243.23992.giannozz@nest.sns.it> <14223878781.20031028163433@dnk.ru> <3F9E719F.5060906@sissa.it> Message-ID: <7826952328.20031028172546@dnk.ru> Dear Mickael! Sure I've compiled ATLAS iwth the same ifc. Nevetheless there is a question conserning the compilation flags - if you could give me some advises here, it would be great. Best Wishes, Sergey Prokudaylo. MP> Do you compile atlas with the same compiler (ifc ?) MP> If Atlas was compile with g77 for instance, there is certainly the MP> problem of these underscores which are somtime added, sometime not MP> depending on the compiler and options. MP> I recommend compiling Atlas with the same compiler use to compile pwscf MP> Sincerely MP> Mickael MP> sprokuda at dnk.ru wrote: >> Dear Sirs! >> I'm trying to compile PWscf (versions 1.3.0 and 1.2.0) on >> Duron-ChainTech hardware, ifc 7.1, fftw 2.3. >> Igot rathe strange result on the liking stage - ATLAS can't find his >> ownm functions. Can anybody say me something on it?. Thak you all in >> advance. >> Prokudaylo Sergey. MP> _______________________________________________ MP> Pw_forum mailing list MP> Pw_forum at pwscf.org MP> http://www.democritos.it/mailman/listinfo/pw_forum -- ? ?????????, sprokuda mailto:sprokuda at dnk.ru From sprokuda at dnk.ru Tue Oct 28 15:33:12 2003 From: sprokuda at dnk.ru (sprokuda at dnk.ru) Date: Tue, 28 Oct 2003 17:33:12 +0300 Subject: [Pw_forum] ATLAS compilation In-Reply-To: <200310281458.51146.giannozz@nest.sns.it> References: <001001c39c03$53d04960$f7361dc3@gost> <200310272243.23992.giannozz@nest.sns.it> <14223878781.20031028163433@dnk.ru> <200310281458.51146.giannozz@nest.sns.it> Message-ID: <2627397968.20031028173312@dnk.ru> Dear Paolo! Yes, I've included all thing like : -llapack -lf77blas -lcblas -latlas. and I did all correction to xerbla.o and would do it for dlamch.o, but the last doesn't resides in ATLAS, and has to be included extra, what, to be honest, adds some more problems to compilation. Best wishes, Sergey Prokudaylo PG> if you explain how you got it, maybe somebody will have a clue >> /home/ATLAS/lib/Linux_ATHLONSSE1/liblapack.a(dgetrf.o): In function >> `dgetrf_': dgetrf.o(.text+0x55): undefined reference to`atl_f77wrap_dgetrf_' PG> did you load all the pieces? something like: PG> -L"directory where ATLAS reside" -llapack -lf77blas -lcblas -latlas PG> ? PG> Paolo -- ? ?????????, sprokuda mailto:sprokuda at dnk.ru From profeta at sissa.it Tue Oct 28 15:36:27 2003 From: profeta at sissa.it (Mickael Profeta) Date: Tue, 28 Oct 2003 15:36:27 +0100 Subject: [Pw_forum] ATLAS compilation In-Reply-To: <7826952328.20031028172546@dnk.ru> References: <001001c39c03$53d04960$f7361dc3@gost> <200310272243.23992.giannozz@nest.sns.it> <14223878781.20031028163433@dnk.ru> <3F9E719F.5060906@sissa.it> <7826952328.20031028172546@dnk.ru> Message-ID: <3F9E7EEB.8030400@sissa.it> sprokuda at dnk.ru wrote: > Dear Mickael! > > Sure I've compiled ATLAS iwth the same ifc. Nevetheless there is a > question conserning the compilation flags - if you could give me some > advises here, it would be great. well in that case there is nothing particular... if you use the same option in pwscf and atlas, it should work -Vaxlib -O2 -tpp6 should do the job I think (do not have a duron to test...) but as told Paolo, in that case it can be an error in the link option of pwscf, can you send us the LIBS parameter, don't you forget to link against lapack end blas? this is an example in another code compile with atlas: LIBS = -L/usr/local/lib -llapack \ -lf77blas -latlas -lcblas -ltstatlas \ -lfftw -lrfftw -lg2c if it can help... Mickael From proffess at yandex.ru Wed Oct 29 07:53:16 2003 From: proffess at yandex.ru (Sergei Lisenkov) Date: Wed, 29 Oct 2003 09:53:16 +0300 (MSK) Subject: [Pw_forum] ATLAS compilation In-Reply-To: <3F9E7EEB.8030400@sissa.it> References: <001001c39c03$53d04960$f7361dc3@gost> <200310272243.23992.giannozz@nest.sns.it> <14223878781.20031028163433@dnk.ru> <3F9E719F.5060906@sissa.it> <7826952328.20031028172546@dnk.ru> <3F9E7EEB.8030400@sissa.it> Message-ID: <3F9F63DC.000005.23547@tide.yandex.ru> Dear Sergey, Some time ago I have this problem. You must include in link the next options: "-lpthread ", LIBS = ........ (libraries) -lpthread -Vaxlib Good luck, Sergey From novikov.dmitri at tiax.biz Wed Oct 29 22:20:53 2003 From: novikov.dmitri at tiax.biz (novikov.dmitri at tiax.biz) Date: Wed, 29 Oct 2003 16:20:53 -0500 Subject: [Pw_forum] vc-relax Message-ID: Could someone send me an example of an input file for a simple "vc-relax" case. Thanks! From novikov.dmitri at tiax.biz Fri Oct 31 18:14:53 2003 From: novikov.dmitri at tiax.biz (novikov.dmitri at tiax.biz) Date: Fri, 31 Oct 2003 12:14:53 -0500 Subject: [Pw_forum] what atoms are counted by ntype Message-ID: Could someone tell me please what atoms are counted as "different type": crystallographycally different or chemically(and spin direction) different? From eyvaz_isaev at yahoo.com Fri Oct 31 18:46:01 2003 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 31 Oct 2003 09:46:01 -0800 (PST) Subject: [Pw_forum] what atoms are counted by ntype In-Reply-To: Message-ID: <20031031174601.974.qmail@web60303.mail.yahoo.com> Dear Dmitry, PWSCF input requires all atoms in the unit cell, including all crystallographycally equivalent and different atoms. I will be happy if it is not true (then there was not neseccary using other codes like LMTO46 to generate equivalent atoms. Even VASP is not able to do that, at least I am not aware of this). So, it means chemically (and spin direction) different atoms. For example, NiO has 2 type of atoms in the paramagnetic case and 4 type of atoms for the antiferromagnetic case. Regards, Eyvaz. --- novikov.dmitri at tiax.biz wrote: > > > > > Could someone tell me please what atoms are counted > as "different type": > crystallographycally different or chemically(and > spin direction) different? > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com From michele.lazzeri at lmcp.jussieu.fr Fri Oct 31 17:25:31 2003 From: michele.lazzeri at lmcp.jussieu.fr (Michele Lazzeri) Date: Fri, 31 Oct 2003 17:25:31 +0100 Subject: [Pw_forum] what atoms are counted by ntype References: Message-ID: <3FA28CFB.E4E61A5B@lmcp.jussieu.fr> novikov.dmitri at tiax.biz wrote: > Could someone tell me please what atoms are counted as "different type": > crystallographycally different or chemically(and spin direction) different? Two atoms are of the same "type" if you are describing them with the same pseudopotential. Two atom of the same "type" can be crystallographycally equivalent or not. At the end of the calculation they can result as being chemically different. You can decide to describe two crystallographically equivalent atoms with two different pseudopotentials. In this case the code will not recognize the two atoms as equivalent. Let's make a stupid example, if you want to compute a molecule of CH3-CH3 you will use two different pseudopotentials: one for C and one for H, and you will have ntyp=2. You can decide (for some weird reason) to use two different pseudopotentials to describe the two C atoms. In this case you will have ntyp=3. Best regards,