From a.soon at auckland.ac.nz Thu Jul 1 05:37:49 2004 From: a.soon at auckland.ac.nz (Aloysius Soon Sze Lok) Date: Thu, 1 Jul 2004 15:37:49 +1200 Subject: [Pw_forum] Warning message - negative or imaginary core charge Message-ID: <200407011537.49190.a.soon@auckland.ac.nz> Dear all, I'm running a slab calculation on Cu2O and found this warning message: number of k points= 13 gaussian broad. (ryd)= 0.0070 ngauss = 0 cart. coord. in units 2pi/a_0 k( 1) = ( 0.0000000 0.0000000 0.0000000), wk = 0.0800000 k( 2) = ( 0.0000000 0.2309401 0.0000000), wk = 0.1600000 k( 3) = ( 0.0000000 0.4618802 0.0000000), wk = 0.1600000 k( 4) = ( 0.2000000 0.3464102 0.0000000), wk = 0.1600000 k( 5) = ( 0.2000000 0.5773503 0.0000000), wk = 0.1600000 k( 6) = ( -0.2000000 0.1154701 0.0000000), wk = 0.1600000 k( 7) = ( 0.2000000 0.1154701 0.0000000), wk = 0.1600000 k( 8) = ( -0.4000000 0.2309401 0.0000000), wk = 0.1600000 k( 9) = ( 0.4000000 0.2309401 0.0000000), wk = 0.1600000 k( 10) = ( -0.2000000 0.3464102 0.0000000), wk = 0.1600000 k( 11) = ( 0.4000000 0.0000000 0.0000000), wk = 0.1600000 k( 12) = ( -0.4000000 0.4618802 0.0000000), wk = 0.1600000 k( 13) = ( 0.6000000 0.1154701 0.0000000), wk = 0.1600000 G cutoff = 326.7719 ( 59669 G-vectors) FFT grid: ( 40, 40,108) nbndx = 428 nbnd = 107 natomwfc = 140 npwx = 7475 nelec = 178.00 nkb = 280 ngl = 2816 warning: negative or imaginary core charge -0.000003 0.000000 Initial potential from superposition of free atoms starting charge = 177.85806 Starting wfc are random Is this a serious warning? Or is the value of -0.000003 small enough to ignore? May I know the implications, please? Thank you in advance. -- Aloysius Soon SL Postgraduate Student Structural & Computational Chemistry Chemistry Department The University Of Auckland, New Zealand Phone: +64 9 373 7599 ext 88291 Email: a.soon at auckland.ac.nz From eyvaz_isaev at yahoo.com Thu Jul 1 10:31:07 2004 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Thu, 1 Jul 2004 01:31:07 -0700 (PDT) Subject: [Pw_forum] Warning message - negative or imaginary core charge In-Reply-To: <200407011537.49190.a.soon@auckland.ac.nz> Message-ID: <20040701083107.37887.qmail@web60305.mail.yahoo.com> Dear Aloysius, Go ahead. It is just warning. I see now you are using gaussian broadening which is more safely in your case. Bests, Eyvaz. --- Aloysius Soon Sze Lok wrote: > Dear all, > I'm running a slab calculation on Cu2O and found > this warning message: > > number of k points= 13 gaussian broad. > (ryd)= 0.0070 ngauss = 0 > cart. coord. in units 2pi/a_0 > k( 1) = ( 0.0000000 0.0000000 > 0.0000000), wk = 0.0800000 > k( 2) = ( 0.0000000 0.2309401 > 0.0000000), wk = 0.1600000 > k( 3) = ( 0.0000000 0.4618802 > 0.0000000), wk = 0.1600000 > k( 4) = ( 0.2000000 0.3464102 > 0.0000000), wk = 0.1600000 > k( 5) = ( 0.2000000 0.5773503 > 0.0000000), wk = 0.1600000 > k( 6) = ( -0.2000000 0.1154701 > 0.0000000), wk = 0.1600000 > k( 7) = ( 0.2000000 0.1154701 > 0.0000000), wk = 0.1600000 > k( 8) = ( -0.4000000 0.2309401 > 0.0000000), wk = 0.1600000 > k( 9) = ( 0.4000000 0.2309401 > 0.0000000), wk = 0.1600000 > k( 10) = ( -0.2000000 0.3464102 > 0.0000000), wk = 0.1600000 > k( 11) = ( 0.4000000 0.0000000 > 0.0000000), wk = 0.1600000 > k( 12) = ( -0.4000000 0.4618802 > 0.0000000), wk = 0.1600000 > k( 13) = ( 0.6000000 0.1154701 > 0.0000000), wk = 0.1600000 > > G cutoff = 326.7719 ( 59669 G-vectors) > FFT grid: ( 40, 40,108) > > nbndx = 428 nbnd = 107 natomwfc = > 140 npwx = 7475 > nelec = 178.00 nkb = 280 ngl = > 2816 > warning: negative or imaginary core charge > -0.000003 0.000000 > > Initial potential from superposition of free > atoms > > starting charge = 177.85806 > Starting wfc are random > > Is this a serious warning? Or is the value of > -0.000003 small enough to > ignore? May I know the implications, please? Thank > you in advance. > > > > -- > Aloysius Soon SL > Postgraduate Student > Structural & Computational Chemistry > Chemistry Department > The University Of Auckland, New Zealand > Phone: +64 9 373 7599 ext 88291 > Email: a.soon at auckland.ac.nz > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Do you Yahoo!? Take Yahoo! Mail with you! Get it on your mobile phone. http://mobile.yahoo.com/maildemo From subhra at uiuc.edu Thu Jul 1 15:26:07 2004 From: subhra at uiuc.edu (Subhradip Ghosh) Date: Thu, 1 Jul 2004 09:26:07 -0400 Subject: [Pw_forum] Setting the scale in plotrho Message-ID: <369c0a1.59169d89.81a0900@expms6.cites.uiuc.edu> Hi, I have a question about setting the scale in the ps output file of charge density generated by plotrho.x. I am working on a Cscl structure. I plotted out the charge densities along [110] and [100] directions. The orthogonal vectors used in [110] case are: e1= 1 1 0 e2= 0 0 1 The orthogonal vectors used in [100] case are: e1= 0 1 0 e2= 0 0 1 When I plotted out the charge densities I found that the size of the [110] figure is 13.25x8.75 cm which is sqrt(2)x1 essentially. The size of [100] figure is 10.7x10.7 cm which is 1x1. This is consistent with the length of the orthogonal vectors chosen. But, I want the plots in the same scale i.e. I want the size of the [100] figure to be 8.75x8.75 cm so that I can have consistency with the size of the [110] figure. Is it possible to do? Subhradip ******************************************************************* Subhradip Ghosh Post doctoral Fellow Department of Materials Science & Engineering University of Illinois at Urbana-Champaign ******************************************************************** Office Home ********************************************************************** MSEB 312C * 2001 South Orchard Street Apt D 1304 W Green street * Urbana, IL-61801 Urbana, IL-61801 * Phone # 2173841824 Phone # 2173336584 * E-mail: subhradip at yahoo.com E-mail: subhra at uiuc.edu * ************************************************************************ From profeta at sissa.it Thu Jul 1 15:38:01 2004 From: profeta at sissa.it (Mickael Profeta) Date: Thu, 01 Jul 2004 15:38:01 +0200 Subject: [Pw_forum] Setting the scale in plotrho In-Reply-To: <369c0a1.59169d89.81a0900@expms6.cites.uiuc.edu> References: <369c0a1.59169d89.81a0900@expms6.cites.uiuc.edu> Message-ID: <40E413B9.8040804@sissa.it> Hi > The orthogonal vectors used in [110] case are: > > e1= 1 1 0 > e2= 0 0 1 > > The orthogonal vectors used in [100] case are: > > e1= 0 1 0 > e2= 0 0 1 > > When I plotted out the charge densities I found that the size > of the [110] figure is 13.25x8.75 cm which is sqrt(2)x1 > essentially. > > The size of [100] figure is 10.7x10.7 cm which is 1x1. > > This is consistent with the length of the orthogonal vectors > chosen. But, I want the plots in the same scale i.e. I want > the size of the [100] figure to be 8.75x8.75 cm so that I can > have consistency with the size of the [110] figure. > > Is it possible to do? If you scale the e vectors it should do the job: leave [110] unchanged and use e1= 0 0.8177 0 e2 = 0 0 0.8177 Hope it can help Mickael From subhra at uiuc.edu Thu Jul 1 15:46:31 2004 From: subhra at uiuc.edu (Subhradip Ghosh) Date: Thu, 1 Jul 2004 09:46:31 -0400 Subject: [Pw_forum] Setting the scale in plotrho Message-ID: <4c62a943.59187bc4.81b5500@expms6.cites.uiuc.edu> Hi, I tried this. It does not change the size of the plot. The only difference it does is that it washes out some of the structures in the previous plot i.e. the one with e1=0 1 0 and e2= 0 0 1 Subhradip ---- Original message ---- >Date: Thu, 01 Jul 2004 15:38:01 +0200 >From: Mickael Profeta >Subject: Re: [Pw_forum] Setting the scale in plotrho >To: pw_forum at pwscf.org > >Hi > >> The orthogonal vectors used in [110] case are: >> >> e1= 1 1 0 >> e2= 0 0 1 >> >> The orthogonal vectors used in [100] case are: >> >> e1= 0 1 0 >> e2= 0 0 1 >> >> When I plotted out the charge densities I found that the size >> of the [110] figure is 13.25x8.75 cm which is sqrt(2)x1 >> essentially. >> >> The size of [100] figure is 10.7x10.7 cm which is 1x1. >> >> This is consistent with the length of the orthogonal vectors >> chosen. But, I want the plots in the same scale i.e. I want >> the size of the [100] figure to be 8.75x8.75 cm so that I can >> have consistency with the size of the [110] figure. >> >> Is it possible to do? > >If you scale the e vectors it should do the job: > >leave [110] unchanged > >and use > >e1= 0 0.8177 0 >e2 = 0 0 0.8177 > >Hope it can help > >Mickael >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum ******************************************************************* Subhradip Ghosh Post doctoral Fellow Department of Materials Science & Engineering University of Illinois at Urbana-Champaign ******************************************************************** Office Home ********************************************************************** MSEB 312C * 2001 South Orchard Street Apt D 1304 W Green street * Urbana, IL-61801 Urbana, IL-61801 * Phone # 2173841824 Phone # 2173336584 * E-mail: subhradip at yahoo.com E-mail: subhra at uiuc.edu * ************************************************************************ From profeta at sissa.it Thu Jul 1 16:00:46 2004 From: profeta at sissa.it (Mickael Profeta) Date: Thu, 01 Jul 2004 16:00:46 +0200 Subject: [Pw_forum] Setting the scale in plotrho In-Reply-To: <4c62a943.59187bc4.81b5500@expms6.cites.uiuc.edu> References: <4c62a943.59187bc4.81b5500@expms6.cites.uiuc.edu> Message-ID: <40E4190E.1070602@sissa.it> Subhradip Ghosh wrote: > Hi, Hi, > > I tried this. It does not change the size of the plot. The > only difference it does is that it washes out some of the > structures in the previous plot i.e. the one with > e1=0 1 0 and e2= 0 0 1 this is normal, as it scales the size of the plane it calculates. I fear you mean the size of the produced drawing, and not the size of the region of the unit cell. Sorry for this misunderstanding. Can't you just rescale the picture with an alternate drawing program? (gimp for instance) I do not know how to do it in another way Mickael From subhra at uiuc.edu Thu Jul 1 16:06:13 2004 From: subhra at uiuc.edu (Subhradip Ghosh) Date: Thu, 1 Jul 2004 10:06:13 -0400 Subject: [Pw_forum] Setting the scale in plotrho Message-ID: <92cd3eff.591a493f.8293000@expms6.cites.uiuc.edu> Sorry for the confusion. I meant the size of the drawing, not the unit cell. Do you mean that there is no way to change the size of the pictures produced by plotrho.x? If I understand it correctly, the size of the picture is set quite arbitrarily, therefore. Thanks for the help. I will try to find out something to rescale my pictures. ---- Original message ---- >Date: Thu, 01 Jul 2004 16:00:46 +0200 >From: Mickael Profeta >Subject: Re: [Pw_forum] Setting the scale in plotrho >To: pw_forum at pwscf.org > >Subhradip Ghosh wrote: >> Hi, > >Hi, >> >> I tried this. It does not change the size of the plot. The >> only difference it does is that it washes out some of the >> structures in the previous plot i.e. the one with >> e1=0 1 0 and e2= 0 0 1 > >this is normal, as it scales the size of the plane it calculates. > >I fear you mean the size of the produced drawing, and not the size of >the region of the unit cell. Sorry for this misunderstanding. >Can't you just rescale the picture with an alternate drawing program? >(gimp for instance) I do not know how to do it in another way > >Mickael >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum ******************************************************************* Subhradip Ghosh Post doctoral Fellow Department of Materials Science & Engineering University of Illinois at Urbana-Champaign ******************************************************************** Office Home ********************************************************************** MSEB 312C * 2001 South Orchard Street Apt D 1304 W Green street * Urbana, IL-61801 Urbana, IL-61801 * Phone # 2173841824 Phone # 2173336584 * E-mail: subhradip at yahoo.com E-mail: subhra at uiuc.edu * ************************************************************************ From cvasse at fy.chalmers.se Thu Jul 1 15:58:54 2004 From: cvasse at fy.chalmers.se (Vasile Chis) Date: Thu, 01 Jul 2004 15:58:54 +0200 Subject: [Pw_forum] wrong phonon eigenvectors. Message-ID: <40E4189E.9010203@fy.chalmers.se> Dear all, I'm performing phonon calculations of a 7-layer Cu(111) slab using a PBE ultrasoft pseudo potential. ecutwfc=30 Ry, ecutrho=480 Ry, smearing=mp, conv-thr=1.0d-12 over a 12x12x1 MP-mesh, ibrav=4. The slab has been relaxed untill total force in the slab was 0.77meV/Angstrom. The phonon calculation converges with no problem. I'm using phonon thresholt of 1.0d-14 and a alpha-mix of 0.1 The main problem is not at Gamma, but at the K-point of the SBZ. K = (0.0, 1/sqrt(3), 0.0) in a_lat units. At this point the frequencies ar correct comparing with experimental results. What bothers me is that the eigenvectors of the lowest mode are incorrect. A sample of the lowest mode: q = ( 0.000000000 0.577350300 0.000000000 ) **************************************************** omega( 1) = 3.195483 [THz] = 106.590533 [cm-1] ( 0.000000 0.000000 0.067560 0.000000 0.458240 0.000000 ) ( 0.000000 0.000000 -0.139989 0.000000 -0.322556 0.000000 ) ( 0.000000 0.000000 -0.115517 0.000000 -0.308861 0.000000 ) ( 0.000000 0.000000 -0.108002 0.000000 -0.307379 0.000000 ) ( 0.000000 0.000000 0.115517 0.000000 0.308861 0.000000 ) ( 0.000000 0.000000 0.139989 0.000000 0.322556 0.000000 ) ( 0.000000 0.000000 0.067560 0.000000 0.458240 0.000000 ) I have not check if the rotational ivariance holdes. If it does not hold, that means that the Rayleigh wave have been lost. Still if the RW is lost how does one find it back? What could be the problem of wrong eigenvectors and how does one solve it? Is the problem of to few layers in the slab? One other thing, I would expect, during the phonon calculation of a high symmetry point on the boundary edge, to only have 14 irreducible representations. But in my case a have all 21. I would be delighted if some one have some pointers to this problem. Thank you in advance. -- """""""""""""""""""""""""""""""""""" Vasile Chis Interface Modelling Group Solid State Physics Department of Experimental Physics G?teborgs Universitet/Chalmers """"""""""""""""""""""""""""""""""""" From giannozz at nest.sns.it Thu Jul 1 16:15:14 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Jul 2004 16:15:14 +0200 Subject: [Pw_forum] Setting the scale in plotrho In-Reply-To: <200407011613.04324.p.giannozzi@sns.it> References: <369c0a1.59169d89.81a0900@expms6.cites.uiuc.edu> <200407011613.04324.p.giannozzi@sns.it> Message-ID: <200407011615.14686.giannozz@nest.sns.it> On Thursday 01 July 2004 15:26, Subhradip Ghosh wrote: > When I plotted out the charge densities I found that the size > of the [110] figure is 13.25x8.75 cm which is sqrt(2)x1 > essentially. > > The size of [100] figure is 10.7x10.7 cm which is 1x1. > > This is consistent with the length of the orthogonal vectors > chosen. But, I want the plots in the same scale i.e. I want > the size of the [100] figure to be 8.75x8.75 cm so that I can > have consistency with the size of the [110] figure. look inside PP/plotrho.f90, routine cplot: ! xdim and ydim are the physical dimensions of the figure in cm ! xs and ys determine a shift of the origin in cm xdim and ydim are set in the calling program to 15*(x/r) and 15*(y/r) cm, where x = x-axis length, y=y-axis length, r = sqrt(x^2+y^2) (xs and ys are set to 4 and 3 cm). Set them to what you like (or modify the code to read them from input) and recompile. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Thu Jul 1 16:27:53 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Jul 2004 16:27:53 +0200 Subject: [Pw_forum] Warning message - negative or imaginary core charge In-Reply-To: <200407011537.49190.a.soon@auckland.ac.nz> References: <200407011537.49190.a.soon@auckland.ac.nz> Message-ID: <200407011627.53670.giannozz@nest.sns.it> On Thursday 01 July 2004 05:37, Aloysius Soon Sze Lok wrote: > warning: negative or imaginary core charge -0.000003 0.000000 > [...] Is this a serious warning? no > Or is the value of -0.000003 small enough to ignore? definitely > May I know the implications, please? in this particular case, none. In general: if you transform a positive function in real space to Fourier space and truncate at some finite cutoff, your positive function is no longer guaranteed to be positive when you transform back to real space. This happens only with core corrections and with ultrasoft pseudopotentials. It can be a a source of trouble, but it is usually solved by increasing the cutoff for the charge density Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From g.ballabio at cineca.it Fri Jul 2 11:11:17 2004 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Fri, 2 Jul 2004 11:11:17 +0200 (MEST) Subject: [Pw_forum] PWscf on Windows? In-Reply-To: References: Message-ID: <1088759476.1706.31.camel@pc-ballabio> Hi all, is it possible to compile PWscf on Windows? Anyone has ever tried? Gerardo From giannozz at nest.sns.it Fri Jul 2 11:32:39 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 2 Jul 2004 11:32:39 +0200 Subject: [Pw_forum] PWscf on Windows? In-Reply-To: <1088759476.1706.31.camel@pc-ballabio> References: <1088759476.1706.31.camel@pc-ballabio> Message-ID: <200407021132.39065.giannozz@nest.sns.it> On Friday 02 July 2004 11:11, Gerardo Ballabio wrote: > is it possible to compile PWscf on Windows? Anyone has ever tried? see file install/Make.cygwin, and the following paragraph in the manual: ----------------------------------------------------------- Note for MS-Windows users: The Makefile for Windows PCs is based on a Makefile written for an earlier version of PWscf (1.2.0), contributed by Lu Fu-Fa, CCIT, Taiwan. Since there have been many changes to the installation procedure, the provided Makefile -- which has never been tested -- may not work. You will need the CygWin package (a UNIX environment for PC which runs in Windows). The provided Makefiles assumes that you have the Intel compiler with MKL libraries installed. Another possibility is to install Linux in dual-boot mode. This can be done without touching Windows (the only somewhat delicate operations are creating a partition for Linux and installing a boot loader). ------------------------------------------------------------ Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From sprokuda at dnk.ru Fri Jul 2 11:42:49 2004 From: sprokuda at dnk.ru (sprokuda) Date: Fri, 2 Jul 2004 13:42:49 +0400 Subject: [Pw_forum] PWscf on Windows? In-Reply-To: <1088759476.1706.31.camel@pc-ballabio> References: <1088759476.1706.31.camel@pc-ballabio> Message-ID: <1713921265.20040702134249@dnk.ru> I've tried - practicaly without any success. I've tried whole way and found that it is waste of time - without any knowledge of Linux I've spent much less time to study Linux from zero to work with PWscf than to compile PWscf on Windows. There is some sence to compile q2r and around codes to Windows to be able to have simpler post-processing of main PWscf results. I've started, but I've not completed. best wishes, Prokudaylo Sergey. > Hi all, > is it possible to compile PWscf on Windows? Anyone has ever tried? > Gerardo > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum Best wishes, Sergey Prokudaylo. From goranka.bilalbegovic at zg.htnet.hr Fri Jul 2 19:12:58 2004 From: goranka.bilalbegovic at zg.htnet.hr (Goranka Bilalbegovic) Date: Fri, 02 Jul 2004 19:12:58 +0200 Subject: [Pw_forum] PWscf on Windows? In-Reply-To: <1088759476.1706.31.camel@pc-ballabio> References: <1088759476.1706.31.camel@pc-ballabio> Message-ID: <40E5979A.9000200@zg.htnet.hr> Gerardo Ballabio wrote: >Hi all, >is it possible to compile PWscf on Windows? Anyone has ever tried? > >Gerardo > > > >see file install/Make.cygwin, and the following paragraph in the manual: >... >Another possibility is to >install Linux in dual-boot mode. This can be done without touching >Windows (the only somewhat delicate operations are creating a partition >for Linux and installing a boot loader). And there is another possibility for someone who does not want to touch Windows on some computer, does not want to create a partition for Linux and install a boot loader, but would like to do something quickly with PWscf on the same computer. There is a possibility to use a "live" Linux: Knoppix (www.knoppix.org), or MandrakeMove, or BitDefender... They run from CD, there is no installation... Best regards, Goranka From breezejd at lsbu.ac.uk Sat Jul 3 02:02:27 2004 From: breezejd at lsbu.ac.uk (breezejd at lsbu.ac.uk) Date: Sat, 3 Jul 2004 01:02:27 +0100 (BST) Subject: [Pw_forum] PWscf on Windows? In-Reply-To: <40E5979A.9000200@zg.htnet.hr> References: <1088759476.1706.31.camel@pc-ballabio> <40E5979A.9000200@zg.htnet.hr> Message-ID: <3265.81.153.6.109.1088812947.squirrel@webmail.lsbu.ac.uk> I have a knoppix/quantian based live CD with pwscf installed, but I have to say that although it is handy being able to run pwscf on any computer without leaving a trace, it can sometimes be annoying when the CDrom keeps spinning up to load programs like gawk or grep in script files. Also, due to pwscfs reaonably high I/O requirements, and that knoppix type CDs rely purely on ramdisks, that memory can run out quite quickly. > And there is another possibility for someone who does not want to touch > Windows on some computer, does not want to create a partition for Linux > and install a boot loader, but would like to do something quickly with > PWscf on the same computer. There is a possibility to use a "live" > Linux: Knoppix (www.knoppix.org), or MandrakeMove, or BitDefender... They > run from CD, there is no installation... > > Best regards, > Goranka > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From prendergast2 at poptop.llnl.gov Wed Jul 7 20:41:37 2004 From: prendergast2 at poptop.llnl.gov (David Prendergast) Date: Wed, 07 Jul 2004 11:41:37 -0700 Subject: [Pw_forum] vc-relax example Message-ID: <40EC43E1.9030105@poptop.llnl.gov> Hi, Is there a working example input file which applies the variable cell relaxation (vc-relax) option in PWscf? I am having some problems converging the cell parameters for a hexagonal system and the output does not seem to reflect what I would expect. David. -- OoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOo David Prendergast Quantum Simulations Group phone: (925) 423-2202 Lawrence Livermore National Laboratory fax: (925) 422-6594 P.O. Box 808, L-415, Livermore, CA 94551 email: prendergast2 at llnl.gov OoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOoOo From a.soon at auckland.ac.nz Thu Jul 8 08:12:54 2004 From: a.soon at auckland.ac.nz (Aloysius Soon Sze Lok) Date: Thu, 8 Jul 2004 18:12:54 +1200 Subject: [Pw_forum] Applying dipole correction to slab calculations Message-ID: <200407081812.54345.a.soon@auckland.ac.nz> Dear all, I'm working on adsorption studies on metal oxides and would like to use PWSCF to run calculations on these surfaces. I have gone through the manual and the pw_docs and found a few input variables, like 'tefield', 'emax' and others, were referred to applying an efield to supercell. However, I cannot understand how this is used for dipole-correction in slab calculations. Are we to specify the experimental dipole value to correspond to the efield strength but this doesn't seem to be the point as the dipole will change as the calculation runs. And by specifying the artificial efield strength, will we over/under-correct the dipole in the cell? Please pardon my questions as I'm not sure how to implement this in my calculations. Much guidance and direction will be deeply appreciated. Many thanks in advance. -- Aloysius Soon SL Postgraduate Student Structural & Computational Chemistry Chemistry Department The University Of Auckland, New Zealand Phone: +64 9 373 7599 ext 88291 Email: a.soon at auckland.ac.nz From bungaro at physics.rutgers.edu Thu Jul 8 17:46:55 2004 From: bungaro at physics.rutgers.edu (CLAUDIA BUNGARO) Date: Thu, 8 Jul 2004 11:46:55 -0400 (EDT) Subject: [Pw_forum] Applying dipole correction to slab calculations In-Reply-To: <200407081812.54345.a.soon@auckland.ac.nz> References: <200407081812.54345.a.soon@auckland.ac.nz> Message-ID: As you rightgully say the dipole of the slab changes during the self-consistency run. You can use the option tefield=.true. to add a saw-like potential along the direction normal to the surface, V(z). The saw-like potential changes during selfconsistency so to keep the EXTERNAL electric field equal to a fixed value, given in the input parameter eamp. To simulate an isolated slab set eamp=0 (zero external electric field). The saw-like potential (i.e. the dipole potential) is a function of both the dipole moment of the slab and the desired external electric field. Beside tefield and eamp you should set also: edir, emaxpos, and eopreg. The edir parameter is the direction perpendicular to the slab And emaxpos and eopreg are two parameters to define the detail of the saw-like shape. The position of the maximum of the sawlike potential within the unit cell ( emaxpos) should be in the middle of the vacuum region. And the part of the unit cell where the sawlike potential has an inverse slope than where the slab is (eopreg) should be a small percentage of the vacuum region. Usefull refs are the one cited in the routine add_efield.f90 ! (See Bengtsson PRB 59, 12 301 (1999) and ! Meyer and Vanderbilt, PRB 63, 205426 (2001). A comment for the pwscf group :) The emaxpos and eopreg parameters and their relationship to the saw-like potential are not really clear, I think a reference to some paper or some more hint would be helpful. Hope this helps __________________________________________________ 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, 8 Jul 2004, Aloysius Soon Sze Lok wrote: > Dear all, > I'm working on adsorption studies on metal oxides and would like to use PWSCF > to run calculations on these surfaces. > I have gone through the manual and the pw_docs and found a few input > variables, like 'tefield', 'emax' and others, were referred to applying an > efield to supercell. > However, I cannot understand how this is used for dipole-correction in slab > calculations. Are we to specify the experimental dipole value to correspond > to the efield strength but this doesn't seem to be the point as the dipole > will change as the calculation runs. And by specifying the artificial efield > strength, will we over/under-correct the dipole in the cell? > > Please pardon my questions as I'm not sure how to implement this in my > calculations. Much guidance and direction will be deeply appreciated. > > > Many thanks in advance. > > > > -- > Aloysius Soon SL > Postgraduate Student > Structural & Computational Chemistry > Chemistry Department > The University Of Auckland, New Zealand > Phone: +64 9 373 7599 ext 88291 > Email: a.soon at auckland.ac.nz > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From goranka.bilalbegovic at zg.htnet.hr Thu Jul 8 20:25:05 2004 From: goranka.bilalbegovic at zg.htnet.hr (Goranka Bilalbegovic) Date: Thu, 08 Jul 2004 20:25:05 +0200 Subject: [Pw_forum] vc-relax example In-Reply-To: <40EC43E1.9030105@poptop.llnl.gov> References: <40EC43E1.9030105@poptop.llnl.gov> Message-ID: <40ED9181.8000709@zg.htnet.hr> David Prendergast wrote: > Hi, > > Is there a working example input file which applies the variable cell > relaxation (vc-relax) option in PWscf? > > I am having some problems converging the cell parameters for a > hexagonal system and the output does not seem to reflect what I would > expect. > > David. > Hi David, I am posting an example for the vc-relax input file which was presented at the PWscf workshop. It was prepared for 1.2.0. You should look at pwdocs/ChangeLog to see whether there is something new for vc-relax after PWscf.1.2.0. I did not use this input file and vc-relax in PWscf, but the choice of dt in the namelist control should be the most sensitive point. Best regards, Goranka -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: pw.GaAs.VC.in Url: /pipermail/attachments/20040708/c2661733/attachment.txt From dchrobak at us.edu.pl Fri Jul 9 09:11:49 2004 From: dchrobak at us.edu.pl (Dariusz Chrobak) Date: Fri, 09 Jul 2004 09:11:49 +0200 Subject: [Pw_forum] P4HT problem? Message-ID: <40EE4535.1070008@us.edu.pl> Dear PWscf users, Do you have experience with running paraller version of the pw.x on P4 HT processor? My machines file contains e.g. the following line: my_host :2 or my_host my_host but after running parraler pw.x : mpirun -np 2 pw.x -npool 2 < input_file CPU time is nearly equal to the CPU time for sequensial version of pw.x. Do you have any comments? Best regards Dariusz Chrobak From giannozz at nest.sns.it Fri Jul 9 10:22:57 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 9 Jul 2004 10:22:57 +0200 Subject: [Pw_forum] P4HT problem? In-Reply-To: <40EE4535.1070008@us.edu.pl> References: <40EE4535.1070008@us.edu.pl> Message-ID: <200407091022.57641.giannozz@nest.sns.it> On Friday 09 July 2004 09:11, Dariusz Chrobak wrote: > Do you have experience with running paraller version of the pw.x > on P4 HT processor? [...] CPU time is nearly equal to the CPU time > for sequential version of pw.x. Do you have any comments? From the manual: Note also that Beowulf-style machines (PC clusters) may have disappointing parallelization performances unless they have a decent communication hardware (at least Gigabit ethernet). Do not expect good scaling with cheap hardware: plane-wave calculations are not at all an "embarrassing parallel" problem. Lemon alert: some multiprocessor motherboards for Intel CPUs have just one memory bus for all processors. This dramatically slows down any code doing massive access to memory (such as PWscf and other plane-wave codes) that runs on processors belonging on the same motherboard. Hardware produced by well-known brands and sold at premium price belongs to this category. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From dchrobak at us.edu.pl Fri Jul 9 10:29:10 2004 From: dchrobak at us.edu.pl (Dariusz Chrobak) Date: Fri, 09 Jul 2004 10:29:10 +0200 Subject: [Pw_forum] P4HT problem? In-Reply-To: <200407091022.57641.giannozz@nest.sns.it> References: <40EE4535.1070008@us.edu.pl> <200407091022.57641.giannozz@nest.sns.it> Message-ID: <40EE5756.6050106@us.edu.pl> Thank you for interesting comments. Best regards Dariusz Chrobak From wahab at mail.ouargla-univ.dz Fri Jul 9 13:09:33 2004 From: wahab at mail.ouargla-univ.dz (ouahab) Date: Fri, 09 Jul 2004 12:09:33 +0100 Subject: [Pw_forum] Parallel execution of pwscf Message-ID: <40EE7CED.8030103@mail.ouargla-univ.dz> Dear users; I am preparing to run pwscf on linux cluter, when reading the manuel, the bacth query system recommended is mpiexec, but this one needs pbs which is not available to donwnload from its page on the net. Can any one help to get it. Thanks in advance. Ouahab A. Laboratoire des Energies Nouvelles et Renuovleables dans les Zones Arides Universit? de Ouargla Alg?rie From profeta at sissa.it Fri Jul 9 15:15:19 2004 From: profeta at sissa.it (Mickael Profeta) Date: Fri, 09 Jul 2004 15:15:19 +0200 Subject: [Pw_forum] Parallel execution of pwscf In-Reply-To: <40EE7CED.8030103@mail.ouargla-univ.dz> References: <40EE7CED.8030103@mail.ouargla-univ.dz> Message-ID: <40EE9A67.2010702@sissa.it> Good morning > > I am preparing to run pwscf on linux cluter, when reading the manuel, > the bacth query system recommended is mpiexec, but this one needs pbs > which is not available to donwnload from its page on the net. Can any > one help to get it. the way to run a mpi job, depend on your mpi environment (mpich/lampi compiler options). You may use mpirun in spite of mpiexec depending on the environnement. A scheduler like pbs is not "necessary" to run mpi job, its goal is to use at best the ressource of the cluster when running multiple jobs. But if you need one, you can have a look there: http://www.openpbs.org/ Yours Sincerely Mickael From marzari at MIT.EDU Fri Jul 9 15:39:57 2004 From: marzari at MIT.EDU (Nicola Marzari) Date: Fri, 09 Jul 2004 09:39:57 -0400 Subject: [Pw_forum] P4HT problem? In-Reply-To: <200407091022.57641.giannozz@nest.sns.it> References: <40EE4535.1070008@us.edu.pl> <200407091022.57641.giannozz@nest.sns.it> Message-ID: <40EEA02D.5070602@mit.edu> Dear All, these are tests made by Matteo Cococcioni on our own cluster. (The code used is CP, the Car-Parrinello molecular dynamics code distributed by www.democritos.it - input and parts of the architecture are common with PWSCF). Our notes are that the bus speed is very important, as is having DDR3200 memory; DDR2700 leads to a 10-15% decrease in performance. Until last week, fastest PIV had 800 MHz front side bus, and fastest Xeons only 533 MHz FSB. For this reason, one dual Xeon performs only between 0% to 40% better than a single Xeon on memory intensive application, and a dual Xeon is on par or even worse with a single PIV with the faster bus. Note that CP is a very optimized code, that relies heavily on the blas/lapack zgemm as implemented by Intel in the MKL library. Gigabit ethernet seems to do poorly above 4 or maybe 6 nodes. Intel has just released Nocona, the Xeon with 800 MHz FSB on the Tumwater E7525 chipset, and this would probably improve considerably the performance of dual-processor machines. Best, nicola > Matteo's tests - opensesame fast cp (uqbar: cluster of PIV 3.2 GHz, > 800 MHz FSB, DDR3200 or DDR2700; dx6: dual xeon 2.4 GHz, 533 MHz FSB, > DDR2100; grant: dual xeon 2.8 (?) GHz, 400 (?) MHz FSB) > ifc 7.1, MKL 6.0, fftw 2.1.3. AgI with 108 atoms, 486 bands. > > ------------------------------------------------- > uqbar (DDR3200 over 8-port gigabit) > 5 iteraz 10 iteraz (10-5, in sec) > 1 28:34 47:49 1155s > 2 15:24 26:31 667s > 3 10:34 17:59 445s > 4 7:55 13:22 327s > 5 7:10 12:14 304s > 6 6:11 10:15 244s > 7 5:50 9:42 232s > 8 5:23 9:25 242s > uqbar (DDR3200 over 4-port gigabit) > 5 iteraz 10 iteraz 5 iteraz 10 iteraz > 1 proc 29:27 49:59 (10-5=1232) 42:51 (dx6) 71:12 (dx6) > 2 proc 15:15 26:19 (10-5=664s) 39:59 (grant) 68:51 (grant) > 3 proc 10:30 17:57 (10-5=447s) > 4 proc 7:55 13:22 (10-5=327s) 22:03 (grant) 38:13 (grant) > uqbar (DDR3200 over 48-port ethernet) > 2 proc 18:45 31:58 > 3 proc 15:12 26:04 > uqbar (DDR2700 over 4-port gigabit) > 1 proc 32:31 55:00 > 2 proc 17:00 29:26 > 3 proc 10:43 18:15 > 4 proc 8:50 14:55 (For some reason, the two 1 proc tests on uqbar seems to be slightly different). We have also a variety of tests on machines with unknown FSBs and RAM speeds > dual Xeon 2.4 GHz (6 cpus - myrinet) DDRAM macarthur 24:27 > dual Xeon 2.4 GHz (4 cpus - myrinet) DDRAM macarthur 36:21 > dual Xeon 2.8 GHz (4 cpus - grant 7,8 lam ethernet) DDRAM 50:45 +/- 15s > dual Xeon 2.4 GHz (2 cpus) RDRAM 63 to 70 minutes > dual Xeon 2.4 GHz (2 cpus - myrinet) DDRAM macarthur 68:30 > dual Xeon 2.8 GHz (2 cpus - grant node 9) DDRAM (200 MHz?) 74:45 +/- 15s > dual Xeon 2.4 GHz (1 cpu) RDRAM 107: minutes > dual Xeon 2.0 GHz (1 cpu) DDR 127: minutes --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From aaron at nemo.physics.ncsu.edu Fri Jul 9 16:48:52 2004 From: aaron at nemo.physics.ncsu.edu (aaron at nemo.physics.ncsu.edu) Date: Fri, 9 Jul 2004 10:48:52 -0400 (EDT) Subject: [Pw_forum] CP question In-Reply-To: <40EEA02D.5070602@mit.edu> Message-ID: Hi, Can I ask if there is any avenue of interoptibility between the CP code and the nice postprocessing (PP) geared toward pw.x? Thanks in advance for your consideration, (not to mention the development!) Aaron From ps at ned.sims.nrc.ca Fri Jul 9 13:59:27 2004 From: ps at ned.sims.nrc.ca (Serguei Patchkovskii) Date: Fri, 9 Jul 2004 07:59:27 -0400 (EDT) Subject: [Pw_forum] P4HT problem? In-Reply-To: <200407091022.57641.giannozz@nest.sns.it> Message-ID: On Fri, 9 Jul 2004, Paolo Giannozzi wrote: > On Friday 09 July 2004 09:11, Dariusz Chrobak wrote: > > > Do you have experience with running paraller version of the pw.x > > on P4 HT processor? [...] CPU time is nearly equal to the CPU time > > for sequential version of pw.x. Do you have any comments? > > Lemon alert: > > some multiprocessor motherboards for Intel CPUs have just one > memory bus for all processors. This dramatically slows down any > code doing massive access to memory (such as PWscf and other > plane-wave codes) that runs on processors belonging on the same > motherboard. While this is quite true (in fact, -all- dual-P4 motherboards share the same front-side bus, which also runs at a slower speed than a single-P4 FSB - typically 533MHz effective vs. 800), there is likely another reason for the results seen by Dariusz. A P4 with "hyperthreading" enabled is simply one CPU, pretending to be two CPUs, not a proper dual-CPU system. Enabling hyperthreading on a P4-class CPU does not add any additional computational resources, it just splits the existing resources differently. It does bring some benefits to multi-user systems, by allowing two threads of execution with -dissimilar- resource requirements to share these resources, instead of taking turns at using them. Running two execution threads with -similar- resource requirements on a P4 HT CPU will simply make each one of them run at half the speed. For a typical high-performance computing workload, the benefits of HT are usually marginal, or even negative. Serguei From giannozz at nest.sns.it Sat Jul 10 17:38:54 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sat, 10 Jul 2004 17:38:54 +0200 Subject: [Pw_forum] CP question In-Reply-To: References: Message-ID: <200407101738.54413.giannozz@nest.sns.it> On Friday 09 July 2004 16:48, aaron at nemo.physics.ncsu.edu wrote: > Can I ask if there is any avenue of interoptibility between the > CP code and the nice postprocessing (PP) geared toward pw.x? of course there is. In principle it is possible to use the postprocessing from PWscf with data files produced by CP or FPMD: the file format is quite the same and the ordering of plane waves and G-vectors is also the same. Moving from principles to practice requires action, though: unless somebody starts to use this possibility and to report success or failure or, better, to fix what doesn't work and to add what is missing, very little will happen. I think that all postprocessing not involving wavefunctions, or involving wavefunctions and implemented for the k=0 (Gamma) case, should work right now or with minimal action. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From aaron at nemo.physics.ncsu.edu Sat Jul 10 16:32:10 2004 From: aaron at nemo.physics.ncsu.edu (aaron at nemo.physics.ncsu.edu) Date: Sat, 10 Jul 2004 10:32:10 -0400 (EDT) Subject: [Pw_forum] CP question In-Reply-To: <200407101738.54413.giannozz@nest.sns.it> Message-ID: Great! Thanks. On Sat, 10 Jul 2004, Paolo Giannozzi wrote: > On Friday 09 July 2004 16:48, aaron at nemo.physics.ncsu.edu wrote: > > > Can I ask if there is any avenue of interoptibility between the > > CP code and the nice postprocessing (PP) geared toward pw.x? > > of course there is. In principle it is possible to use the postprocessing > from PWscf with data files produced by CP or FPMD: the file format > is quite the same and the ordering of plane waves and G-vectors is > also the same. Moving from principles to practice requires action, > though: unless somebody starts to use this possibility and to report > success or failure or, better, to fix what doesn't work and to add what > is missing, very little will happen. I think that all postprocessing not > involving wavefunctions, or involving wavefunctions and implemented > for the k=0 (Gamma) case, should work right now or with minimal action. > > Paolo > > From a.soon at auckland.ac.nz Sun Jul 11 07:45:06 2004 From: a.soon at auckland.ac.nz (Aloysius Soon Sze Lok) Date: Sun, 11 Jul 2004 17:45:06 +1200 Subject: [Pw_forum] Using projwfc.x and plotrho.x Message-ID: <200407111745.06842.a.soon@auckland.ac.nz> Dear all, I have completed a pw.x calculation and used projwfc.x to do some post-processing work. I believe this calculates the pDOS and Lowdin charges. However, the output files for each orbital of each element does not specify which specific orbital it is. For example, the file will show that the pDOS is for oxygen 2 p orbitals, with 3 individual dos values (same heading), but NOT specifying which p(i) it is, ie, p(x), p(y) or p(z). Perhaps it is written somewhere but I did not notice. So if anyone could help, I'll deeply appreciate. Also, with regards to plotting the ildos, is there a way to reduce the 'dot' that represents the atomic position? Many thanks. -- Aloysius Soon SL Postgraduate Student Structural & Computational Chemistry Chemistry Department The University Of Auckland, New Zealand Phone: +64 9 373 7599 ext 88291 Email: a.soon at auckland.ac.nz From giannozz at nest.sns.it Mon Jul 12 10:04:13 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 12 Jul 2004 10:04:13 +0200 Subject: [Pw_forum] Using projwfc.x and plotrho.x In-Reply-To: <200407111745.06842.a.soon@auckland.ac.nz> References: <200407111745.06842.a.soon@auckland.ac.nz> Message-ID: <200407121004.13300.giannozz@nest.sns.it> On Sunday 11 July 2004 07:45, Aloysius Soon Sze Lok wrote: > [...] the file will show that the pDOS is for oxygen 2 p orbitals, with 3 > individual dos values (same heading), but NOT specifying which p(i) > it is, ie, p(x), p(y) or p(z). > Perhaps it is written somewhere but I did not notice. the order of m components of Y_{lm} is determined by routine "ylmr2": 1, cos(phi), sin(phi), cos(2*phi), sin(2*phi), .., cos(m*phi), sin(m*phi) where phi is the polar angle (x=r cos(theta)cos(phi), y=r cos(theta)sin(phi)) > Also, with regards to plotting the ildos, is there a way to reduce the > 'dot' that represents the atomic position? see subroutine "atomi" in plotrho.f90. It should be sufficient to reduce the parameter "delta" Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From dchrobak at us.edu.pl Tue Jul 13 16:51:55 2004 From: dchrobak at us.edu.pl (Dariusz Chrobak) Date: Tue, 13 Jul 2004 16:51:55 +0200 Subject: [Pw_forum] configure: error: cannot compute sizeof (int), 77 Message-ID: <40F3F70B.8000208@us.edu.pl> Dear PWscf users, I noticed the following error during configuration of the PWscf with MPI (MPICH): checking build system type... i686-pc-linux-gnu checking for mpif90... mpif90 ... checking for int... yes checking size of int... configure: error: cannot compute sizeof (int), 77 See `config.log' for more details. (this error not appears in case of sequensial version of the PWscf) However, compilation process (with make.sys file taken from /PWscf_home/install directory) goes without errors. Do you have any comments? Best regards Dariusz Chrobak From g.ballabio at cineca.it Tue Jul 13 18:51:45 2004 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Tue, 13 Jul 2004 18:51:45 +0200 (MEST) Subject: [Pw_forum] configure: error: cannot compute sizeof (int), 77 In-Reply-To: <40F3F70B.8000208@us.edu.pl> References: <40F3F70B.8000208@us.edu.pl> Message-ID: <1089737504.32222.12.camel@pc-ballabio> On Tue, 2004-07-13 at 16:51, Dariusz Chrobak wrote: > checking size of int... configure: error: cannot compute sizeof > (int), 77 > See `config.log' for more details. I know that. It's been fixed in the current development version. Gerardo From dchrobak at us.edu.pl Thu Jul 15 09:43:55 2004 From: dchrobak at us.edu.pl (Dariusz Chrobak) Date: Thu, 15 Jul 2004 09:43:55 +0200 Subject: [Pw_forum] configure: error: cannot compute sizeof (int), 77 In-Reply-To: <1089737504.32222.12.camel@pc-ballabio> References: <40F3F70B.8000208@us.edu.pl> <1089737504.32222.12.camel@pc-ballabio> Message-ID: <40F635BB.9040703@us.edu.pl> U?ytkownik Gerardo Ballabio napisa?: >On Tue, 2004-07-13 at 16:51, Dariusz Chrobak wrote: > > >> checking size of int... configure: error: cannot compute sizeof >>(int), 77 >> See `config.log' for more details. >> >> > >I know that. It's been fixed in the current development version. > >Gerardo > > >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > > > Sorry, I do not check previous messages. D.Ch From cbarreteau at cea.fr Fri Jul 16 13:49:20 2004 From: cbarreteau at cea.fr (Cyrille Barreteau) Date: Fri, 16 Jul 2004 13:49:20 +0200 (CEST) Subject: [Pw_forum] DOS in parallel mode Message-ID: <53186.132.166.21.1.1089978560.squirrel@132.166.21.1> Dear pwscf users, I know that the dos.x (and other post processing codes) are not available in parallel mode. However is it possible to obtain the file.pun with pw.x (nscf calculation) in parallel mode and use this file.pun to obtain the dos in serial mode with dos.x. I have tried this but the code stops with the error %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from read_file : error # 154 reading eigenvalues %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... Does it mean this is not possible, or did I make a mistake? thanks, cyrille -- ================================================ Cyrille Barreteau | phone:+33 (0)1 69 08 29 51 CEA Saclay | fax : +33 (0)1 69 08 84 46 DSM/DRECAM/SPCSI | email: cbarreteau at cea.fr Batiment 462 | 91191 Gif sur Yvette Cedex FRANCE ~~~~~~~~~~~~~~~~~~~~~~~~ web : http://www-drecam.cea.fr/spcsi/groupe1.htm ================================================ From goranka.bilalbegovic at zg.htnet.hr Sat Jul 17 09:46:20 2004 From: goranka.bilalbegovic at zg.htnet.hr (Goranka Bilalbegovic) Date: Sat, 17 Jul 2004 09:46:20 +0200 Subject: [Pw_forum] DOS in parallel mode In-Reply-To: <53186.132.166.21.1.1089978560.squirrel@132.166.21.1> References: <53186.132.166.21.1.1089978560.squirrel@132.166.21.1> Message-ID: <40F8D94C.3010106@zg.htnet.hr> Cyrille Barreteau wrote: >I know that the dos.x (and other post processing codes) are >not available in parallel mode. However is it possible to >obtain the file.pun with pw.x (nscf calculation) in parallel mode and >use this file.pun to obtain the dos in serial mode with dos.x. >I have tried this but the code stops with the error > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from read_file : error # 154 > reading eigenvalues > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > stopping ... > > >Does it mean this is not possible, or did I make a mistake? > > Hello Cyrille, Although it is true that many postprocessing routines do not run in the parallel mode, according to the example 8 ("shows how to use pw.x and dos.x to calculate the Density of States of Ni"), dos.x runs in parallel: $PARA_PREFIX $PW_ROOT/bin/dos.x $PARA_POSTFIX < ni.dos2.in > ni.dos2.out However, I was using dos.x last year (old versions of PWscf) together with some tricks. On old SGI O2K I had a line as (please note that there is no < for input file, pw.x ran on the same number of processors): mpirun -np 2 dos.x -npool 1 inp.dos.in > inp.dos.out On the Linux cluster (pw.x ran on several processors): mpirun -np 1 dos.x -npool 1 < inp.dos.in > inp.dos.out Best regards, Goranka From ashima at cat.hokudai.ac.jp Sat Jul 17 13:08:20 2004 From: ashima at cat.hokudai.ac.jp (ashima) Date: Sat, 17 Jul 2004 20:08:20 +0900 Subject: [Pw_forum] Surface Phonon and vib. of molecule Message-ID: <20040717110820.976D33181D@green.sys.hokudai.ac.jp> Hi all, I'd like to calculate vibrational properties of crystal surfaces and molecules. But I don't have the knowlege about the way to calculate the surface phonon. So, I have the following two questions: (1) Can ph.x code calculate phonon of slab model ?. If, 'Yes', please teach the methods. (2) Can ph.x code calculate vibrational modes of molecule, such as Benzene ring ?. Does it needs a special manners ?. Thank you in advance. Best regards, H. Ashima */*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/ Hidenori Ashima Catalysis Research Center Hokkaido University, Japan e-mail : ashima at cat.hokudai.ac.jp */*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/ From baroni at sissa.it Sat Jul 17 13:19:27 2004 From: baroni at sissa.it (Stefano Baroni) Date: Sat, 17 Jul 2004 13:19:27 +0200 Subject: [Pw_forum] Surface Phonon and vib. of molecule In-Reply-To: <20040717110820.976D33181D@green.sys.hokudai.ac.jp> References: <20040717110820.976D33181D@green.sys.hokudai.ac.jp> Message-ID: <2A194AE5-D7E3-11D8-AB96-000A95CDDD16@sissa.it> On Jul 17, 2004, at 1:08 PM, ashima wrote: > Hi all, > I'd like to calculate vibrational properties > of crystal surfaces and molecules. > > But I don't have the knowlege about the way to calculate the surface > phonon. > > So, I have the following two questions: > > (1) > Can ph.x code calculate phonon of slab model ?. > If, 'Yes', please teach the methods. as a first approximation, the computer does not know the difference between a "slab" and a "crystal". A "slab" is simply a crystal with lots of empty space stuffed into it. So, just set the supercell and calculate its vibrational modes "as though" it were a physical crystal. Of course, the number of modes thus obtained is limited by the thickness of the slab. When you're done with this excercise, please revert to this forum and we will teach you how to simulate a "semi-infinite" slab by patching interatomic force constants calculated for the bulk onto the results obtained for the finite slab. > (2) > Can ph.x code calculate vibrational modes of molecule, such as Benzene > ring ?. > Does it needs a special manners ?. Same as above. Just build a fake crystal with a benzene ring at the origin (say) of the unit cell, and lots of empty space surrounding it (i.e. choose the size of the unit cell large enough so that the charge densities of neighboring molecules do not overlap). Hope this helps Stefano Baroni --- Stefano Baroni --- SISSA & DEMOCRITOS National Simulation Center via Beirut 2-4 34014 Trieste Grignano / [+39] 040 3787 406 (tel) -528 (fax) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html From ashima at cat.hokudai.ac.jp Sun Jul 18 05:53:00 2004 From: ashima at cat.hokudai.ac.jp (ashima) Date: Sun, 18 Jul 2004 12:53:00 +0900 Subject: [Pw_forum] Surface Phonon and vib. of molecule In-Reply-To: <2A194AE5-D7E3-11D8-AB96-000A95CDDD16@sissa.it> References: <2A194AE5-D7E3-11D8-AB96-000A95CDDD16@sissa.it> Message-ID: <20040718035300.2213F3181D@green.sys.hokudai.ac.jp> Stefano Baroni wrote: > > On Jul 17, 2004, at 1:08 PM, ashima wrote: > as a first approximation, the computer does not know the difference > between a "slab" and a "crystal". A "slab" is simply a crystal with > lots of empty space stuffed into it. So, just set the supercell and > calculate its vibrational modes "as though" it were a physical crystal. > Of course, the number of modes thus obtained is limited by the > thickness of the slab. When you're done with this excercise, please > revert to this forum and we will teach you how to simulate a > "semi-infinite" slab by patching interatomic force constants calculated > for the bulk onto the results obtained for the finite slab. > Thank you for your advise. I will charange the calculations in the way you told. Best regards, H. Ashima */*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/ Hidenori Ashima Catalysis Research Center Hokkaido University e-mail : ashima at cat.hokudai.ac.jp */*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/ From ashima at cat.hokudai.ac.jp Sun Jul 18 06:10:54 2004 From: ashima at cat.hokudai.ac.jp (ashima) Date: Sun, 18 Jul 2004 13:10:54 +0900 Subject: [Pw_forum] Error message : ph.x Message-ID: <20040718041054.B2DFC31771@blue.sys.hokudai.ac.jp> Hi all, I got the following error message in the vibrational calculations of a molecule by ph.x : %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from cinterpolate : error # 1 not implemented %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... This model of molecule was positioned at the center of a large cell enough not to overlap neighboring molecules. What's the meaning of above error message ?. Best regards, H. Ashima */*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/ Hidenori Ashima Catalysis Research Center Hokkaido University, Japan e-mail : ashima at cat.hokudai.ac.jp */*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/*/ From giannozz at nest.sns.it Sun Jul 18 21:53:54 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Sun, 18 Jul 2004 21:53:54 +0200 Subject: [Pw_forum] Error message : ph.x In-Reply-To: <20040718041054.B2DFC31771@blue.sys.hokudai.ac.jp> References: <20040718041054.B2DFC31771@blue.sys.hokudai.ac.jp> Message-ID: <200407182153.54499.giannozz@nest.sns.it> On Sunday 18 July 2004 06:10, ashima wrote: > > I got the following error message in the vibrational calculations of > a molecule by ph.x : > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from cinterpolate : error # 1 > not implemented > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > What's the meaning of above error message ?. it is a confusing error message - now fixed - meaning that the phonon code contained in Gamma/ (specific for molecules) needs data produced by pw.x using Gamma-point algorithms; or vice versa, the phonon code in PH (the generic one) cannot process data produced by pw.x using Gamma-point algorithms Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From giannozz at nest.sns.it Mon Jul 19 11:42:12 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 19 Jul 2004 11:42:12 +0200 Subject: [Pw_forum] DOS in parallel mode In-Reply-To: <53186.132.166.21.1.1089978560.squirrel@132.166.21.1> References: <53186.132.166.21.1.1089978560.squirrel@132.166.21.1> Message-ID: <200407191142.12031.giannozz@nest.sns.it> On Friday 16 July 2004 13:49, Cyrille Barreteau wrote: > I know that the dos.x (and other post processing codes) are > not available in parallel mode. postprocessing codes that read data files produced by pw.x are available in parallel mode. In version 2.0 and earlier you may need to run the postprocessing code on exactly the same number of processors and pools, though Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From dchrobak at us.edu.pl Mon Jul 19 15:15:46 2004 From: dchrobak at us.edu.pl (Dariusz Chrobak) Date: Mon, 19 Jul 2004 15:15:46 +0200 Subject: [Pw_forum] DOS in example8 Message-ID: <40FBC982.2000406@us.edu.pl> Dear PWscf users, My question is about DOS presented in example8. Could you tell mi why total DOS from ni.dos differs from total DOS from ni.pos_tot ? Do you know how to obtain similar results ? Than you in advanse Dariusz Chrobak From a.soon at auckland.ac.nz Tue Jul 20 05:41:23 2004 From: a.soon at auckland.ac.nz (Aloysius Soon Sze Lok) Date: Tue, 20 Jul 2004 15:41:23 +1200 Subject: [Pw_forum] Workfunction and Magnetism Message-ID: <200407201541.23178.a.soon@auckland.ac.nz> Dear all, I have 2 questions I would like to consult: 1. After calculating surface energy for a slab calculation, can we use PWSCF to extract the workfunction of the surface? If so, is it specify somewhere? I'm interested in comparing the change in the workfunction between the clean surface and the adsorbate-surface. 2. As I will be working on magnetic systems very soon, I would like to ask how to specify it in the run. I understand that some of the parameters like nspin will be set to 2 and the starting_magnetization(1) will be needed to be set between -1 and 1. So what would the values 1 and -1 mean? I will be working specifically on iron sulphide systems. Many thanks. -- Aloysius Soon SL Postgraduate Student Structural & Computational Chemistry Chemistry Department The University Of Auckland, New Zealand Phone: +64 9 373 7599 ext 88291 Email: a.soon at auckland.ac.nz From baroni at sissa.it Tue Jul 20 08:02:53 2004 From: baroni at sissa.it (Stefano Baroni) Date: Tue, 20 Jul 2004 08:02:53 +0200 Subject: [Pw_forum] Workfunction and Magnetism In-Reply-To: <200407201541.23178.a.soon@auckland.ac.nz> References: <200407201541.23178.a.soon@auckland.ac.nz> Message-ID: <701D124E-DA12-11D8-9F42-000A95CDDD16@sissa.it> On Jul 20, 2004, at 5:41 AM, Aloysius Soon Sze Lok wrote: > Dear all, > I have 2 questions I would like to consult: > > 1. After calculating surface energy for a slab calculation, can we use > PWSCF > to extract the workfunction of the surface? If so, is it specify > somewhere? > I'm interested in comparing the change in the workfunction between the > clean > surface and the adsorbate-surface. What is the work function? - The minimim work (energy) to extract an electron from a solid work = (final energy) - (initial energy) in order to minimize a difference, you have to minimize the first term and maximize the second What is the maximum initial energy that an electron being ionized can have? - The Fermi energy (provided by the code) What is the minimum final energy that an electron being ionized can have? - The "vacuum level" i.e. the value that the potential felt by an electron (the Kohn-Sham potential, also calculated and used by the code) has far from the surface. How far? - The farthest (i.e. in the middle of the vacuum region in between two adjacent slabs) Beware: * The Kohn-Sham potential and the Hartree potential converge to the same value far from the origin; the second converges faster. You see why? * A good estimate is not necessarily the best that you can extract from a given calculation. Convergence has always to be checked. i.e. you may want to plot the potential (or rather the planar average of it parallel to the surface) as a function of the the distance from the surface, and see if and to which value it converges ... In other words, the vacuum region may be thick enough for calculating the surface energy, but not so for the work function ... > 2. As I will be working on magnetic systems very soon, I would like to > ask how > to specify it in the run. I understand that some of the parameters > like nspin > will be set to 2 and the starting_magnetization(1) will be needed to > be set > between -1 and 1. So what would the values 1 and -1 mean? I will be > working > specifically on iron sulphide systems. most probably, (rho_up-rho_down)/(rho_up+rho_down) Hope this helps Stefano Baroni --- Stefano Baroni --- SISSA & DEMOCRITOS National Simulation Center via Beirut 2-4 34014 Trieste Grignano / [+39] 040 3787 406 (tel) -528 (fax) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html From cvasse at fy.chalmers.se Mon Jul 19 17:48:37 2004 From: cvasse at fy.chalmers.se (Vasile Chis) Date: Mon, 19 Jul 2004 17:48:37 +0200 Subject: [Pw_forum] ?semi-infinite slab? Message-ID: <40FBED55.9030800@fy.chalmers.se> Dear users, I'm curious about how the technique of expanding a slab by introducing slab-adapted bulk interatomic force constants is done. --Stefano Baroni wrote: >Of course, the number of modes thus obtained is limited by the=20 >thickness of the slab. When you're done with this excercise, please=20 >revert to this forum and we will teach you how to simulate a=20 >"semi-infinite" slab by patching interatomic force constants calculated=20= >for the bulk onto the results obtained for the finite slab. Thank you in advance. Regards/ """""""""""""""""""""""""""""""""""" Vasile Chis Interface Modelling Group Solid State Physics Department of Experimental Physics G?teborgs Universitet/Chalmers """"""""""""""""""""""""""""""""""""" From FB153746 at ATIL.CEA.FR Tue Jul 20 10:30:22 2004 From: FB153746 at ATIL.CEA.FR (=?iso-8859-1?Q?BOUYER_Fr=E9d=E9ric_153746?=) Date: Tue, 20 Jul 2004 10:30:22 +0200 Subject: [Pw_forum] Allocation memory Message-ID: <601BA98F9E8C7645B7A0052ADBFCC714066EB8BC@atil.intra.cea.fr> Dear PWscf users, A particular job (on our NEC) with few atoms stops immediately after few minutes, and the error is the following : 194 Could not allocate in ALLOCATE PROG=allocate_nlpot ELN=186(400016768) Called from init_run ELN=31(40010145c) Called from pwscf ELN=82(4000001154) An other one stops with the error : 194 Could not allocate in ALLOCATE PROG=wfcinit.wfcinit_k ELN=365(...) Called from wfcinit ELN=125 (...) Called from init_run ELN=42 (...) Called from pwscf ELN=82 (...) Does anybody can help me to pass the jobs ? Do I have to change something in the parameter.f90 file, or somewhere else ? Many thanks in advance, Fr?d?ric Bouyer -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20040720/4281bc31/attachment.htm From a.soon at auckland.ac.nz Wed Jul 21 07:45:21 2004 From: a.soon at auckland.ac.nz (Aloysius Soon Sze Lok) Date: Wed, 21 Jul 2004 17:45:21 +1200 Subject: [Pw_forum] Workfunction (reprise) Message-ID: <200407211745.21842.a.soon@auckland.ac.nz> Dear Prof. Baroni, You mentioned that the KS potential is calculated and used in the code. Could you please tell me how to extract it from the code at the middle of the vacuum? Many thanks, Sir. -- Aloysius Soon SL Postgraduate Student Structural & Computational Chemistry Chemistry Department The University Of Auckland, New Zealand Phone: +64 9 373 7599 ext 88291 Email: a.soon at auckland.ac.nz From sbraccia at sissa.it Wed Jul 21 09:40:16 2004 From: sbraccia at sissa.it (carlo sbraccia) Date: Wed, 21 Jul 2004 09:40:16 +0200 Subject: [Pw_forum] Workfunction (reprise) In-Reply-To: <200407211745.21842.a.soon@auckland.ac.nz> References: <200407211745.21842.a.soon@auckland.ac.nz> Message-ID: <1090395616.2551.27.camel@dhpc-4-04.sissa.it> On Wed, 2004-07-21 at 07:45, Aloysius Soon Sze Lok wrote: > Dear Prof. Baroni, > You mentioned that the KS potential is calculated and used in the code. Could > you please tell me how to extract it from the code at the middle of the > vacuum? > > Many thanks, Sir. > > To extract the value of the KS potential in the middle of the vacuum you should follow this strategy: 1) do a self-consistent calculation for your system (this you probably have already done) 2) do a post-processing of output files with the pp.x code, setting plot_num = 1 (for the meaning of the this an other input variables see the file pwdoc/INPUT_PP) 3) compute the planar average of the KS potential using the code average.x (for the meaning of the input variables see the header of the source file PP/average.x). 4) plot the result and extract what you need (column 3 gives you an useful macroscopic average). Here are reported minimal templates for the input files: For pp.x code: &inputPP prefix = " ", outdir = " ", filplot = "example", plot_num = 1, / For average.x code: 1 example 1.D0 200 3 0.1D0 carlo sbraccia From baroni at sissa.it Wed Jul 21 14:04:47 2004 From: baroni at sissa.it (Stefano Baroni) Date: Wed, 21 Jul 2004 14:04:47 +0200 Subject: [Pw_forum] Workfunction (reprise) In-Reply-To: <1090395616.2551.27.camel@dhpc-4-04.sissa.it> References: <200407211745.21842.a.soon@auckland.ac.nz> <1090395616.2551.27.camel@dhpc-4-04.sissa.it> Message-ID: <29550159-DB0E-11D8-8AB8-000A95CDDD16@sissa.it> Thank you, Carlo. Beware, though, that extracting the "vacuum level" from the Kohn-Sham potential (which includes xc) rather then from Hartree-only could be numerically less efficient. I think we should leave it to Aloysius as a useful exercise: 1) To understand why this is so. 2) To figure our how to extract the Hartree potential from the output of a PWscf calculation. Have fun! Stefano B. On Jul 21, 2004, at 9:40 AM, carlo sbraccia wrote: > On Wed, 2004-07-21 at 07:45, Aloysius Soon Sze Lok wrote: >> Dear Prof. Baroni, >> You mentioned that the KS potential is calculated and used in the >> code. Could >> you please tell me how to extract it from the code at the middle of >> the >> vacuum? >> >> Many thanks, Sir. >> >> > > To extract the value of the KS potential in the middle of the vacuum > you > should follow this strategy: > > 1) do a self-consistent calculation for your system (this you probably > have already done) > > 2) do a post-processing of output files with the pp.x code, setting > plot_num = 1 (for the meaning of the this an other input variables see > the file pwdoc/INPUT_PP) > > 3) compute the planar average of the KS potential using the code > average.x (for the meaning of the input variables see the header of the > source file PP/average.x). > > 4) plot the result and extract what you need (column 3 gives you an > useful macroscopic average). > > Here are reported minimal templates for the input files: > > For pp.x code: > > &inputPP > prefix = " ", > outdir = " ", > filplot = "example", > plot_num = 1, > / > > For average.x code: > > 1 > example > 1.D0 > 200 > 3 > 0.1D0 > > carlo sbraccia > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > --- Stefano Baroni --- SISSA & DEMOCRITOS National Simulation Center via Beirut 2-4 34014 Trieste Grignano / [+39] 040 3787 406 (tel) -528 (fax) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html From eesantis at unity.ncsu.edu Wed Jul 21 23:46:27 2004 From: eesantis at unity.ncsu.edu (Erik Santiso) Date: Wed, 21 Jul 2004 17:46:27 -0400 (EDT) Subject: [Pw_forum] neb_routines.f90 In-Reply-To: <200407182153.54499.giannozz@nest.sns.it> References: <20040718041054.B2DFC31771@blue.sys.hokudai.ac.jp> <200407182153.54499.giannozz@nest.sns.it> Message-ID: <3600.152.14.40.124.1090446387.squirrel@webmail.ncsu.edu> Dear pwscf developers: I've been having some trouble compiling the newest version of pwscf on both a Linux machine with ifc and an SGI with PGI. The compiler seems to be having trouble when compiling neb_routines.f90, because the variable nat is called both from module 'basis' and module 'input_parameters'. Is this the way it was intended to be, or should it rather be 'USE basis , ONLY : nat', removing nat from line 33? Thanks for your help! Erik. ------------------------------------- Vir prudens non contra ventum mingit. From giannozz at nest.sns.it Thu Jul 22 12:23:18 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 22 Jul 2004 12:23:18 +0200 (CEST) Subject: [Pw_forum] neb_routines.f90 In-Reply-To: <3600.152.14.40.124.1090446387.squirrel@webmail.ncsu.edu> References: <20040718041054.B2DFC31771@blue.sys.hokudai.ac.jp> <200407182153.54499.giannozz@nest.sns.it> <3600.152.14.40.124.1090446387.squirrel@webmail.ncsu.edu> Message-ID: On Wed, 21 Jul 2004, Erik Santiso wrote: > I've been having some trouble compiling the newest version of pwscf on > both a Linux machine with ifc and an SGI with PGI. The compiler seems to > be having trouble when compiling neb_routines.f90, because the variable > nat is called both from module 'basis' and module 'input_parameters'. > Is this the way it was intended to be no, it isn't > or should it rather be 'USE basis, ONLY: nat', removing nat from line 33? correct. This problem was previously remarked and fixed in the version under development, but not in the stable version (sorry) Paolo From eesantis at unity.ncsu.edu Thu Jul 22 17:26:02 2004 From: eesantis at unity.ncsu.edu (Erik Santiso) Date: Thu, 22 Jul 2004 11:26:02 -0400 (EDT) Subject: [Pw_forum] neb_routines.f90 In-Reply-To: References: <20040718041054.B2DFC31771@blue.sys.hokudai.ac.jp> <200407182153.54499.giannozz@nest.sns.it> <3600.152.14.40.124.1090446387.squirrel@webmail.ncsu.edu> Message-ID: <3065.66.26.237.218.1090509962.squirrel@webmail.ncsu.edu> Thanks! I fixed this but I found another problem. This time is in the restart.f90 file under the PWNC directory. The declaration: USE wavefunctions_module , ONLY: evc => evc_nc appears in all three subroutines in the module restart_spin. This gives me a compiling error. I changed it so that the declaration appears under the module and not the subroutines and it seems to compile correctly now. Is that OK? Thanks again for your help! Erik. ------------------------------------- Vir prudens non contra ventum mingit. > On Wed, 21 Jul 2004, Erik Santiso wrote: > >> I've been having some trouble compiling the newest version of pwscf on >> both a Linux machine with ifc and an SGI with PGI. The compiler seems to >> be having trouble when compiling neb_routines.f90, because the variable >> nat is called both from module 'basis' and module 'input_parameters'. >> Is this the way it was intended to be > > no, it isn't > >> or should it rather be 'USE basis, ONLY: nat', removing nat from line >> 33? > > correct. This problem was previously remarked and fixed in the > version under development, but not in the stable version (sorry) > > Paolo > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From a.soon at auckland.ac.nz Fri Jul 23 08:00:22 2004 From: a.soon at auckland.ac.nz (Aloysius Soon Sze Lok) Date: Fri, 23 Jul 2004 18:00:22 +1200 Subject: [Pw_forum] Projected DOS Message-ID: <200407231800.22326.a.soon@auckland.ac.nz> Dear all, I have asked the question on this subject before but I have yet to resolve this. The following pdos is extracted from a Cu+ ion (using projwfc.x) # E (eV) dos(E) dos(E) dos(E) dos(E) dos(E) -18.746 0.927E-14 0.360E-17 0.238E-13 0.460E-13 0.137E-19 -18.646 0.833E-10 0.630E-13 0.214E-09 0.414E-09 0.241E-15 -18.546 0.111E-06 0.151E-09 0.286E-06 0.552E-06 0.586E-12 -18.446 0.230E-04 0.518E-07 0.594E-04 0.115E-03 0.215E-09 ...... From a simplistic point of view I can understand that this d-element has 5 d-orbitals but till now, I still can't figure out which is which. The same case for p orbitals. Prof. Paolo did mention about spherical harmonics, but please pardon my lack of knowledge and understanding, but I can't figure out how the combination of l and m can help to decipher which orbital is which. To me, it looks like when the magnetic quantum number m = -2, -1, 0, 1 and 2, the "m" that is mentioned in the output file corresponds to 1 to 5 respectively. But sometimes this seems quite wrong. I'm totally confused. So if there's someone that can help me out in this, I am TOTALLY grateful. As I have been dwelling in this issue for about a few weeks, I totally need some help. Many thanks in advance. -- Aloysius Soon SL Postgraduate Student Structural & Computational Chemistry Chemistry Department The University Of Auckland, New Zealand Phone: +64 9 373 7599 ext 88291 Email: a.soon at auckland.ac.nz From giannozz at nest.sns.it Fri Jul 23 12:16:09 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 23 Jul 2004 12:16:09 +0200 (CEST) Subject: [Pw_forum] neb_routines.f90 In-Reply-To: <3065.66.26.237.218.1090509962.squirrel@webmail.ncsu.edu> References: <20040718041054.B2DFC31771@blue.sys.hokudai.ac.jp> <200407182153.54499.giannozz@nest.sns.it> <3600.152.14.40.124.1090446387.squirrel@webmail.ncsu.edu> <3065.66.26.237.218.1090509962.squirrel@webmail.ncsu.edu> Message-ID: On Thu, 22 Jul 2004, Erik Santiso wrote: > [...] in the restart.f90 file under the PWNC directory. The declaration: > > USE wavefunctions_module , ONLY: evc => evc_nc > > appears in all three subroutines in the module restart_spin. This gives me > a compiling error. I changed it so that the declaration appears under the > module and not the subroutines and it seems to compile correctly now. > Is that OK? I think so Paolo From pushpa at jncasr.ac.in Sat Jul 24 09:21:01 2004 From: pushpa at jncasr.ac.in (Raghani Pushpa) Date: Sat, 24 Jul 2004 12:51:01 +0530 (IST) Subject: [Pw_forum] (no subject) Message-ID: <41771.202.41.111.151.1090653661.squirrel@mercury.jncasr.ac.in> dear users, i was doing the slab calculations and wanted to fix a few middle layers of the slab, but i don't find the option 'fixatom' in the new code (pw-2.0.1), as it used to be in the old version. could u tell me is there any option related to this in the version 2.0.1? pushpa From Giovanni.Cantele at na.infn.it Sat Jul 24 09:31:27 2004 From: Giovanni.Cantele at na.infn.it (Giovanni Cantele) Date: Sat, 24 Jul 2004 09:31:27 +0200 Subject: [Pw_forum] (no subject) In-Reply-To: <41771.202.41.111.151.1090653661.squirrel@mercury.jncasr.ac.in> References: <41771.202.41.111.151.1090653661.squirrel@mercury.jncasr.ac.in> Message-ID: <4102104F.3050303@na.infn.it> Raghani Pushpa wrote: > dear users, > i was doing the slab calculations and wanted to fix a few middle layers of > the slab, but i don't find the option 'fixatom' in the new code > (pw-2.0.1), as it used to be in the old version. could u tell me is there > any option related to this in the version 2.0.1? > > pushpa > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From pwdocs/INPUT_PW: ATOMIC_POSITIONS { alat | bohr | crystal | angstrom } There are "nat" cards like the following X x y z {if_pos(1) if_pos(2) if_pos(3)} where : X Character: label of the atom as specified in ATOMIC_SPECIES x, y, z Real: atomic positions if_pos: Integer: component i of the force for this atom is multiplied by if_pos(i), which must be 0 or 1. Used to keep selected atoms and/or selected components fixed in a NEB, MD or optimization run A few examples: Si 0.1 0.2 0.3 This Si atom is relaxed during the optimization Si 0.1 0.2 0.3 0 0 0 This Si atom is kept fixed during the optimization Si 0.1 0.2 0.3 0 1 1 In this case only the x coordinate is kept fixed Hope this helps Giovanni -- Dr. Giovanni Cantele INFM-Coherentia and Dipartimento di Scienze Fisiche Universita' di Napoli "Federico II" Complesso Universitario di Monte S. Angelo - Ed. G Via Cintia, I-80126, Napoli, Italy Phone: +39 081 676910 Fax: +39 081 676346 E-mail: Giovanni.Cantele at na.infn.it Web: http://people.na.infn.it/~cantele From francesco.antoniella at poste.it Fri Jul 23 10:42:26 2004 From: francesco.antoniella at poste.it (Francesco Antoniella) Date: Fri, 23 Jul 2004 10:42:26 +0200 Subject: [Pw_forum] Projected DOS References: <200407231800.22326.a.soon@auckland.ac.nz> Message-ID: <001201c47090$fd349620$2d10a8c0@alhambra.local> ----- Original Message ----- From: "Aloysius Soon Sze Lok" To: "PWSCF Forum" Sent: Friday, July 23, 2004 8:00 AM Subject: [Pw_forum] Projected DOS > Dear all, > I have asked the question on this subject before but I have yet to resolve > this. > The following pdos is extracted from a Cu+ ion (using projwfc.x) > > # E (eV) dos(E) dos(E) dos(E) dos(E) dos(E) > -18.746 0.927E-14 0.360E-17 0.238E-13 0.460E-13 0.137E-19 > -18.646 0.833E-10 0.630E-13 0.214E-09 0.414E-09 0.241E-15 > -18.546 0.111E-06 0.151E-09 0.286E-06 0.552E-06 0.586E-12 > -18.446 0.230E-04 0.518E-07 0.594E-04 0.115E-03 0.215E-09 > ...... > > >From a simplistic point of view I can understand that this d-element has 5 > d-orbitals but till now, I still can't figure out which is which. The same > case for p orbitals. > > Prof. Paolo did mention about spherical harmonics, but please pardon my lack > of knowledge and understanding, but I can't figure out how the combination of > l and m can help to decipher which orbital is which. To me, it looks like > when the magnetic quantum number m = -2, -1, 0, 1 and 2, the "m" that is > mentioned in the output file corresponds to 1 to 5 respectively. But > sometimes this seems quite wrong. I'm totally confused. The m is not the m of COMPLEX spherical harmonics but is the \mu of the REAL spherical harmonics (or CUBIC harmonics) so the order is p_y p_z p_x for the p orbitals. For d orbitals this is more complex but always the "central" value is the \mu=0. Regards Ing. Francesco Antoniella > > So if there's someone that can help me out in this, I am TOTALLY grateful. As > I have been dwelling in this issue for about a few weeks, I totally need some > help. > > Many thanks in advance. > > > > > -- > Aloysius Soon SL > Postgraduate Student > Structural & Computational Chemistry > Chemistry Department > The University Of Auckland, New Zealand > Phone: +64 9 373 7599 ext 88291 > Email: a.soon at auckland.ac.nz > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From pushpa at jncasr.ac.in Sat Jul 24 17:41:54 2004 From: pushpa at jncasr.ac.in (Raghani Pushpa) Date: Sat, 24 Jul 2004 21:11:54 +0530 (IST) Subject: [Pw_forum] (no subject) Message-ID: <43932.202.41.111.151.1090683714.squirrel@mercury.jncasr.ac.in> dear users, i was doing the slab calculations and wanted to fix some of the middle layers, so at the end of the atomic positions i gave "iforce(i=1-3)=0". but in the output the warning comes as "Warning: card IFORCE(I=1-4)=0 ignored". what does this warning mean? and is the code going to keep the last three atoms fixed? thanks, pushpa my input file is as follows: &control calculation='relax', restart_mode='from_scratch', tstress = .true. tprnfor = .true. prefix='ru', pseudo_dir = '../../pseudi/', outdir='./tmp/' / &system ibrav= 4, celldm(1) =5.0, celldm(3) = 10.0527, nat= 6, ntyp= 2, ecutwfc =30.0, occupations='smearing', smearing='methfessel-paxton', degauss=0.002, nbnd=30, / &electrons conv_thr = 1.0d-10 mixing_beta = 0.7 / &ions upscale =10 / ATOMIC_SPECIES Ru 101.07 Ru.pbe-n-van.UPF Cu 63.55 Cu.pbe-d-rrkjus.UPF ATOMIC_POSITIONS Cu 0.0 0.0 2.09431 Cu 0. 0.577350 -2.09431 Ru 0.5 0.288675 1.25659 Ru 0. 0.577350 0.418863 Ru 0. 0.0 -0.418863 Ru 0.5 0.288675 -1.25659 iforce(i=1-3)=0 K_POINTS {automatic} 12 12 4 0 0 0 From Tone.Kokalj at ijs.si Mon Jul 26 09:36:43 2004 From: Tone.Kokalj at ijs.si (Tone Kokalj) Date: Mon, 26 Jul 2004 09:36:43 +0200 Subject: [Pw_forum] (no subject) In-Reply-To: <43932.202.41.111.151.1090683714.squirrel@mercury.jncasr.ac.in> References: <43932.202.41.111.151.1090683714.squirrel@mercury.jncasr.ac.in> Message-ID: <20040726073642.GA24004@crysden.ijs.si> On Sat, Jul 24, 2004 at 09:11:54PM +0530, Raghani Pushpa wrote: > dear users, > i was doing the slab calculations and wanted to fix some of the middle > layers, so at the end of the atomic positions i gave "iforce(i=1-3)=0". > but in the output the warning comes as "Warning: card IFORCE(I=1-4)=0 > ignored". what does this warning mean? and is the code going to keep the > last three atoms fixed? > > ATOMIC_POSITIONS > Cu 0.0 0.0 2.09431 > Cu 0. 0.577350 -2.09431 > Ru 0.5 0.288675 1.25659 > Ru 0. 0.577350 0.418863 > Ru 0. 0.0 -0.418863 > Ru 0.5 0.288675 -1.25659 > iforce(i=1-3)=0 No it will not, as the "iforce" line will be ignored. Instead, you should use ATOMIC_POSITIONS Cu 0.0 0.0 2.09431 Cu 0. 0.577350 -2.09431 Ru 0.5 0.288675 1.25659 Ru 0. 0.577350 0.418863 0 0 0 Ru 0. 0.0 -0.418863 0 0 0 Ru 0.5 0.288675 -1.25659 0 0 0 In the last three lines the zeros at 5th/6th/7th column mean: "don't move this atom in x/y/z direction". Regards, Tone -- +------------------------------------------------------------------------+ | Anton Kokalj Email: Tone.Kokalj at ijs.si | | Department of Physical and Organic Chemistry Phone: x 386 1 477 3523 | | Jozef Stefan Institute Fax: x 386 1 477 3811 | | Jamova 39, SI-1000 Ljubljana | | SLOVENIA | +------------------------------------------------------------------------+ !!! *** please, do not send me attachments in any Microsoft format *** !!! From proffess at yandex.ru Mon Jul 26 15:00:01 2004 From: proffess at yandex.ru (Sergei Lisenkov) Date: Mon, 26 Jul 2004 17:00:01 +0400 (MSD) Subject: [Pw_forum] Reading input file in MPI version (Linux cluster) In-Reply-To: <40ED9181.8000709@zg.htnet.hr> References: <40EC43E1.9030105@poptop.llnl.gov> <40ED9181.8000709@zg.htnet.hr> Message-ID: <41050051.000019.15117@pantene.yandex.ru> Dear PWscf authors and users, I have a problem with running PWscf code on PC Linux cluster with SCALI MPI. In this implementation pw.x code cannot read input file. I asked of system people of this cluster but they could not suggest anything. I tried a different input keys such "-stdin" but it does't helps. It seems that in such mpi mode system cannot read the input file using "<" command. But other programs requested input file without such sign (i.e. mpirun -np 4 ./a.x input ) does work. Because system people cannot help me I think about changing of procedure of input. Is it possible to rewrite something in the source code that code could work such "./pw.x input" ? Or there are other solutions? Thanks a lot, Best wishes, Sergey From sbraccia at sissa.it Mon Jul 26 15:13:02 2004 From: sbraccia at sissa.it (Carlo Sbraccia) Date: Mon, 26 Jul 2004 15:13:02 +0200 (CEST) Subject: [Pw_forum] Reading input file in MPI version (Linux cluster) In-Reply-To: <41050051.000019.15117@pantene.yandex.ru> References: <40EC43E1.9030105@poptop.llnl.gov> <40ED9181.8000709@zg.htnet.hr> <41050051.000019.15117@pantene.yandex.ru> Message-ID: Dear Sergey, you have simply to run the pw.x executable in this way: pw.x -input input_file > output_file .or. pw.x -inp input_file > output_file .or. pw.x -in input_file > output_file Let me know if it doesn't work. carlo sbraccia On Mon, 26 Jul 2004, Sergei Lisenkov wrote: > Dear PWscf authors and users, > > I have a problem with running PWscf code on PC Linux cluster with SCALI MPI. In this implementation pw.x code cannot read input file. I asked of system people of this cluster but they could not suggest anything. I tried a different input keys such "-stdin" but it does't helps. It seems that in such mpi mode system cannot read the input file using "<" command. But other programs requested input file without such sign (i.e. mpirun -np 4 ./a.x input ) does work. > > Because system people cannot help me I think about changing of procedure of input. Is it possible to rewrite something in the source code that code could work such "./pw.x input" ? Or there are other solutions? > > Thanks a lot, > Best wishes, > Sergey > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From proffess at yandex.ru Mon Jul 26 15:21:52 2004 From: proffess at yandex.ru (Sergei Lisenkov) Date: Mon, 26 Jul 2004 17:21:52 +0400 (MSD) Subject: [Pw_forum] Reading input file in MPI version (Linux cluster) In-Reply-To: References: <40EC43E1.9030105@poptop.llnl.gov> <40ED9181.8000709@zg.htnet.hr> Message-ID: <41050570.000002.16912@pantene.yandex.ru> Dear Carlo, Thank you very much for your promt reply. I tried the first - it works!!! But such "input flags" does not explain in the manual. Thanks a lot, Best wishes, Sergey From subhra at uiuc.edu Mon Jul 26 17:21:28 2004 From: subhra at uiuc.edu (Subhradip Ghosh) Date: Mon, 26 Jul 2004 11:21:28 -0400 Subject: [Pw_forum] Question on charge density plot Message-ID: <89fdf28b.66011332.81a0700@expms6.cites.uiuc.edu> Hi, I was trying to understand the details of 2-D charge density plot using chdens and plotrho. What I understand is that using chdens we only get rho in G-space and plotrho draws the contrours of rho(G). I want to have contour plots of rho(r). Is there a way to get it? Also, what units are these charge densities? Thanks, Subhradip ******************************************************************* Subhradip Ghosh Post doctoral Fellow Department of Materials Science & Engineering University of Illinois at Urbana-Champaign ******************************************************************** Office Home ********************************************************************** MSEB 312C * 2001 South Orchard Street Apt D 1304 W Green street * Urbana, IL-61801 Urbana, IL-61801 * Phone # 2173841824 Phone # 2173336584 * E-mail: subhradip at yahoo.com E-mail: subhra at uiuc.edu * ************************************************************************ From giannozz at nest.sns.it Tue Jul 27 08:46:10 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 27 Jul 2004 08:46:10 +0200 (CEST) Subject: [Pw_forum] Question on charge density plot In-Reply-To: <89fdf28b.66011332.81a0700@expms6.cites.uiuc.edu> References: <89fdf28b.66011332.81a0700@expms6.cites.uiuc.edu> Message-ID: On Mon, 26 Jul 2004, Subhradip Ghosh wrote: > I understand is that using chdens we only get rho in G-space > and plotrho draws the contours of rho(G) not really: chdens produces rho(r) in the selected region of space and plotrho draws contours of rho(r) > Also, what units are these charge densities? 1/(Bohr radii)^3 : the charge density yields the number of electrons (not the total charge) when integrated over the unit cell. Paolo From subhra at uiuc.edu Tue Jul 27 15:29:08 2004 From: subhra at uiuc.edu (Subhradip Ghosh) Date: Tue, 27 Jul 2004 09:29:08 -0400 Subject: [Pw_forum] Question on charge density plot Message-ID: <1618f5b6.667aa062.81a0c00@expms6.cites.uiuc.edu> Dear Paolo, Thank you very much. Yes, I understand the thing now. I was initially confused at seeing 'rhoG' being computed in the subroutine 'plot2d'. Thanks for your help. Subhradip ---- Original message ---- >Date: Tue, 27 Jul 2004 08:46:10 +0200 (CEST) >From: Paolo Giannozzi >Subject: Re: [Pw_forum] Question on charge density plot >To: pw_forum at pwscf.org > >On Mon, 26 Jul 2004, Subhradip Ghosh wrote: > >> I understand is that using chdens we only get rho in G-space >> and plotrho draws the contours of rho(G) > >not really: chdens produces rho(r) in the selected region of space >and plotrho draws contours of rho(r) > >> Also, what units are these charge densities? > >1/(Bohr radii)^3 : the charge density yields the number of electrons >(not the total charge) when integrated over the unit cell. > >Paolo >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum ******************************************************************* Subhradip Ghosh Post doctoral Fellow Department of Materials Science & Engineering University of Illinois at Urbana-Champaign ******************************************************************** Office Home ********************************************************************** MSEB 312C * 2001 South Orchard Street Apt D 1304 W Green street * Urbana, IL-61801 Urbana, IL-61801 * Phone # 2173841824 Phone # 2173336584 * E-mail: subhradip at yahoo.com E-mail: subhra at uiuc.edu * ************************************************************************ From subhra at uiuc.edu Tue Jul 27 16:30:58 2004 From: subhra at uiuc.edu (Subhradip Ghosh) Date: Tue, 27 Jul 2004 10:30:58 -0400 Subject: [Pw_forum] Gradient of charge density Message-ID: Dear PWSCF users, I am interested in plotting the gradient of charge density and the laplacian of charge density on the same plane as the charge density itself. Can somebody suggest an algorithm available on this? Subhradip ******************************************************************* Subhradip Ghosh Post doctoral Fellow Department of Materials Science & Engineering University of Illinois at Urbana-Champaign ******************************************************************** Office Home ********************************************************************** MSEB 312C * 2001 South Orchard Street Apt D 1304 W Green street * Urbana, IL-61801 Urbana, IL-61801 * Phone # 2173841824 Phone # 2173336584 * E-mail: subhradip at yahoo.com E-mail: subhra at uiuc.edu * ************************************************************************ From giannozz at nest.sns.it Wed Jul 28 08:19:07 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 28 Jul 2004 08:19:07 +0200 (CEST) Subject: [Pw_forum] Projected DOS In-Reply-To: <001201c47090$fd349620$2d10a8c0@alhambra.local> References: <200407231800.22326.a.soon@auckland.ac.nz> <001201c47090$fd349620$2d10a8c0@alhambra.local> Message-ID: On Fri, 23 Jul 2004, Francesco Antoniella wrote: > The m is not the m of COMPLEX spherical harmonics but is the \mu of the > REAL spherical harmonics (or CUBIC harmonics) so the order is p_y p_z > p_x for the p orbitals. For d orbitals this is more complex but always > the "central" value is the \mu=0. it used to be that way, but in recent version the order is, I think: 1 P_{0,0}(t) 2 P_{1,0}(t) 3 P_{1,1}(t)*cos(phi) 4 P_{1,1}(t)*sin(phi) 5 P_{2,0}(t) 6 P_{2,1}(t)*cos(phi) 7 P_{2,1}(t)*sin(phi) 8 P_{2,2}(t)*cos(2*phi) 9 P_{2,2}(t)*sin(2*phi) .... P_{l,m}=Legendre Polynomials, t=cos(theta)=z/r, phi=atan(y/x). No warranty. Anybody really interested in knowing for sure which spherical harmonic combination is which should look into routine "ylmr2" Paolo From pushpa at jncasr.ac.in Wed Jul 28 08:49:14 2004 From: pushpa at jncasr.ac.in (Raghani Pushpa) Date: Wed, 28 Jul 2004 12:19:14 +0530 (IST) Subject: [Pw_forum] (no subject) Message-ID: <51758.202.41.111.151.1090997354.squirrel@mercury.jncasr.ac.in> dear users, i was doing the slab calculations on a 9 layer slab with pw.2.0.1. i was relaxing the top most and bottom most layers. following are the energies after each ionic move ie. when the self consistency is achieved after each ionic move. 1 -526.81315909 2 -526.82109709 3 -526.81619935 4 -526.82131390 5 -526.82123272 6 -526.82130692 7 -526.82131312 8 -526.82131381 9 -526.82131383 10 -526.82131383 11 -526.82131383 12 -526.82131383 13 -526.82131383 14 -526.82131383 15 -526.82131383 16 -526.82131383 17 -526.82131383 18 -526.82131383 19 -526.82131383 20 -526.82131383 21 -526.82131383 22 -526.82131383 23 -526.82131383 24 -526.82131383 25 -526.82131383 26 -526.82131383 27 -526.82131383 28 -526.82131383 29 -526.82131383 30 -526.82131383 31 -526.82131383 32 -526.82131383 33 -526.82131383 34 -526.82131383 35 -526.82131383 36 -526.82131383 37 -526.82131383 though the energy is not decreasing further from 9th to 37th iteration but the code is still running, it does not stop. could u please tell me why is it happening? pushpa From proffess at yandex.ru Wed Jul 28 12:01:26 2004 From: proffess at yandex.ru (Sergei Lisenkov) Date: Wed, 28 Jul 2004 14:01:26 +0400 (MSD) Subject: [Pw_forum] compiling on IBM PPC970 In-Reply-To: References: <200407231800.22326.a.soon@auckland.ac.nz> Message-ID: <41077976.000001.12391@tide.yandex.ru> Dear PWscf authors and users, I try to compile PW code on IBM PPC970 SMP cluster, using xlf90 fortran (8.1 version), gcc , essl library and mpich-1.2.5.2. The OS is Suse Linux. I use the Makefile for ibmsp configuration and made some changes. Because I have not any experience with IBM I have a problems. First, I tried to compile the serial version. I got : .... cdiaghg.o: In function `cdiaghg': cdiaghg.o(.text+0x138): undefined reference to `ilaenv' cdiaghg.o(.text+0x4b0): undefined reference to `zhegv' cdiaghg.o(.text+0x59c): undefined reference to `zhegvx' input.o: In function `verify_tmpdir': input.o(.text+0x9930): undefined reference to `c_mkdir' mix_pot.o: In function `mix_potential': mix_pot.o(.text+0x110c): undefined reference to `dsytrf' mix_pot.o(.text+0x118c): undefined reference to `dsytri' mix_rho.o: In function `mix_rho': mix_rho.o(.text+0x2f74): undefined reference to `dsytrf' mix_rho.o(.text+0x2ff4): undefined reference to `dsytri' mix_rho.o: In function `approx_screening2': mix_rho.o(.text+0x93e0): undefined reference to `dsytrf' mix_rho.o(.text+0x945c): undefined reference to `dsytri' rdiaghg.o: In function `rdiaghg': rdiaghg.o(.text+0x124): undefined reference to `ilaenv' update_pot.o: In function `extrapolate_wfcs': update_pot.o(.text+0x23d8): undefined reference to `zgesvd' ../Modules/berry_phase.o: In function `indi_of_ig': ../Modules/berry_phase.o(.text+0xcc): undefined reference to `ln_ind' ../Modules/berry_phase.o(.text+0x10c): undefined reference to `ln_ind' ../Modules/berry_phase.o(.text+0x14c): undefined reference to `ln_ind' ../Modules/berry_phase.o(.text+0x170): undefined reference to `ln_ind' ../Modules/berry_phase.o(.text+0x19c): undefined reference to `ln_ind' ../Modules/berry_phase.o(.text+0x20c): more undefined references to `ln_ind' follow ../Modules/berry_phase.o: In function `__berry_phase_MOD_berry_setup': ../Modules/berry_phase.o(.text+0x618): undefined reference to `ln_alloc' ../Modules/berry_phase.o(.text+0x660): undefined reference to `ln_set' ../Modules/berry_phase.o(.text+0x67c): undefined reference to `ln_activate' ../Modules/berry_phase.o(.text+0x2ddc): undefined reference to `ln_dealloc' ../Modules/parser.o: In function `__parser_MOD_p_err_l': ../Modules/parser.o(.text+0x248): undefined reference to `cpflush' ../Modules/parser.o: In function `__parser_MOD_p_err_s': ../Modules/parser.o(.text+0x4ac): undefined reference to `cpflush' ../Modules/parser.o: In function `__parser_MOD_p_err_r': ../Modules/parser.o(.text+0x710): undefined reference to `cpflush' ../Modules/parser.o: In function `__parser_MOD_p_err_i': ../Modules/parser.o(.text+0x968): undefined reference to `cpflush' ../Modules/parser.o: In function `__parser_MOD_field_compare': ../Modules/parser.o(.text+0x11b8): undefined reference to `cpflush' ../Modules/ptoolkit.o: In function `pzhptrd': ../Modules/ptoolkit.o(.text+0x8b4): undefined reference to `dlapy3' ../Modules/ptoolkit.o(.text+0x8e8): undefined reference to `dlamch' ../Modules/ptoolkit.o(.text+0x8f8): undefined reference to `dlamch' ../Modules/ptoolkit.o(.text+0x9d4): undefined reference to `dlapy3' ../Modules/ptoolkit.o(.text+0xa24): undefined reference to `zladiv' ../Modules/ptoolkit.o(.text+0xad4): undefined reference to `zladiv' ../Modules/ptoolkit.o: In function `pzsteqr': ../Modules/ptoolkit.o(.text+0x31a4): undefined reference to `lsame' ../Modules/ptoolkit.o(.text+0x31d4): undefined reference to `lsame' ../Modules/ptoolkit.o(.text+0x3208): undefined reference to `lsame' ../Modules/ptoolkit.o(.text+0x32fc): undefined reference to `xerbla' ../Modules/ptoolkit.o(.text+0x35cc): undefined reference to `dlamch' ../Modules/ptoolkit.o(.text+0x35e8): undefined reference to `dlamch' ../Modules/ptoolkit.o(.text+0x3664): undefined reference to `zlaset' ../Modules/ptoolkit.o(.text+0x3c0c): undefined reference to `dlanst' ../Modules/ptoolkit.o(.text+0x3c8c): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x3cd0): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x3d3c): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x3d80): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x3f54): undefined reference to `dlaev2' ../Modules/ptoolkit.o(.text+0x4350): undefined reference to `dlae2' ../Modules/ptoolkit.o(.text+0x43c4): undefined reference to `dlapy2' ../Modules/ptoolkit.o(.text+0x4490): undefined reference to `dlartg' ../Modules/ptoolkit.o(.text+0x4530): undefined reference to `dlartg' ../Modules/ptoolkit.o(.text+0x4bc0): undefined reference to `dlaev2' ../Modules/ptoolkit.o(.text+0x4fc0): undefined reference to `dlae2' ../Modules/ptoolkit.o(.text+0x5034): undefined reference to `dlapy2' ../Modules/ptoolkit.o(.text+0x50f0): undefined reference to `dlartg' ../Modules/ptoolkit.o(.text+0x518c): undefined reference to `dlartg' ../Modules/ptoolkit.o(.text+0x56c4): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x570c): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x5788): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x57d0): undefined reference to `dlascl' ../Modules/ptoolkit.o(.text+0x5b14): undefined reference to `dlasrt' /usr/lib/gcc-lib/powerpc-suse-linux/3.2.2/../../../libessl.so: undefined reference to `pthread_once' /usr/lib/gcc-lib/powerpc-suse-linux/3.2.2/../../../libessl.so: undefined reference to `_xlf_create_threadlocal' make[1]: *** [memory.x] Error 1 It seems there are problems with LAPACK library. But I use the ESSL library. Does this library contain LAPACK? And another. In all makefiles for ibm machine I did not find anything for parallel version. When I added in CPP flags -D__MPI -D__PARA such for other clusters I got a warning that these flags are not recognized. So, How to compile a parallel version? Thanks a lot, Best wishes, Sergey From profeta at sissa.it Wed Jul 28 13:21:23 2004 From: profeta at sissa.it (Mickael Profeta) Date: Wed, 28 Jul 2004 13:21:23 +0200 Subject: [Pw_forum] compiling on IBM PPC970 In-Reply-To: <41077976.000001.12391@tide.yandex.ru> References: <200407231800.22326.a.soon@auckland.ac.nz> <41077976.000001.12391@tide.yandex.ru> Message-ID: <41078C33.5020009@sissa.it> Hi > I try to compile PW code on IBM PPC970 SMP cluster, using xlf90 fortran (8.1 version), gcc , essl library and mpich-1.2.5.2. The OS is Suse Linux. I use the Makefile for ibmsp configuration and made some changes. Because I have not any experience with IBM I have a problems. which version of PW are you compiling? in the last version, the configure script should do all the job for you... > /usr/lib/gcc-lib/powerpc-suse-linux/3.2.2/../../../libessl.so: undefined reference to `pthread_once' > /usr/lib/gcc-lib/powerpc-suse-linux/3.2.2/../../../libessl.so: undefined reference to `_xlf_create_threadlocal' you should link againt libpthread > make[1]: *** [memory.x] Error 1 > > It seems there are problems with LAPACK library. But I use the ESSL library. Does this library contain LAPACK? no, essl implement part of lapack and with another API, so you should link also with lapack. > And another. In all makefiles for ibm machine I did not find anything for parallel version. When I added in CPP flags -D__MPI -D__PARA such for other clusters I got a warning that these flags are not recognized. So, How to compile a parallel version? if you can, use the configure script, should solve all the problems. depending on your architecture, to compile parallel version, you should use a wrapper for the compiler, like mpi_f77 or something like that. Mike From proffess at yandex.ru Wed Jul 28 14:12:28 2004 From: proffess at yandex.ru (Sergei Lisenkov) Date: Wed, 28 Jul 2004 16:12:28 +0400 (MSD) Subject: [Pw_forum] compiling on IBM PPC970 In-Reply-To: <41078C33.5020009@sissa.it> References: <200407231800.22326.a.soon@auckland.ac.nz> <41077976.000001.12391@tide.yandex.ru> <41078C33.5020009@sissa.it> Message-ID: <4107982C.00000C.19094@tide.yandex.ru> Dear Mike, Thank you very much for your help. But the Configure script cannot configure my system ( I use pw-2.0.4): ./configure checking build system type... powerpc64-unknown-linux-gnu configure: WARNING: unsupported architecture checking for f90... f90 checking for Fortran 77 compiler default output... configure: error: Fortran 77 compiler cannot create executables See `config.log' for more details. I looked at the file config.log. There is a problem with f77: ... configure:1521: $? = 249 configure:1534: checking for Fortran 77 compiler default output configure:1537: f90 conftest.f >&5 f90: 1501-218 file conftest.f contains an incorrect file suffix /usr/bin/ld:conftest.f: file format not recognized; treating as linker script /usr/bin/ld:conftest.f:1: parse error configure:1540: $? = 1 configure: failed program was: | program main | | end configure:1579: error: Fortran 77 compiler cannot create executables See `config.log' for more details. But xlf works well. What is the problem? Thanks a lot, Best wishes, Sergey Lisenkov From a.soon at auckland.ac.nz Thu Jul 29 07:48:39 2004 From: a.soon at auckland.ac.nz (Aloysius Soon Sze Lok) Date: Thu, 29 Jul 2004 17:48:39 +1200 Subject: [Pw_forum] Projected DOS Message-ID: <200407291748.39545.a.soon@auckland.ac.nz> Dear Prof. Palo, Many thanks to your reply. This makes it clear for me to look into it. Thanks to many others who have contributed to my understanding of this. -- Aloysius Soon SL Postgraduate Student Structural & Computational Chemistry Chemistry Department The University Of Auckland, New Zealand Phone: +64 9 373 7599 ext 88291 Email: a.soon at auckland.ac.nz From timtro at rogers.com Wed Jul 28 23:34:37 2004 From: timtro at rogers.com (Timothy A.V. Teatro) Date: Wed, 28 Jul 2004 17:34:37 -0400 Subject: [Pw_forum] Temperature changeing and REAL CP Dynamics Message-ID: <1091050477.12558.6.camel@localhost.localdomain> Hello. We have a system that must be temperature controlled. Not only do we need to be able to set the temperature, but we will need to change it, and dictate the rate of change of temperature (approximately). It seems that setting the ion_temperature to 'rescaling' works, but the tolp (tolerance) does not. Also, when we have mode set to 'restart' and attempt to change the temperature, it goes crazy and vaporizes our sample! On a second note. Is it possible to use real Car-Parinello MD calculations? I see in the modules directory, there is a file that suggests that you can set the calculation to 'cp', but when I look at the input.f90 source file, there is no CASE to allow for 'cp'. Thank you all very much From g.ballabio at cineca.it Thu Jul 29 11:06:21 2004 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Thu, 29 Jul 2004 11:06:21 +0200 (MEST) Subject: [Pw_forum] Temperature changeing and REAL CP Dynamics In-Reply-To: <1091050477.12558.6.camel@localhost.localdomain> References: <1091050477.12558.6.camel@localhost.localdomain> Message-ID: <1091091980.15576.7.camel@pc-ballabio> On Wed, 2004-07-28 at 23:34, Timothy A.V. Teatro wrote: > On a second note. Is it possible to use real Car-Parinello MD > calculations? I see in the modules directory, there is a file that > suggests that you can set the calculation to 'cp', but when I look at > the input.f90 source file, there is no CASE to allow for 'cp'. The PWscf distribution includes two programs that specifically do Car-Parrinello dynamics, "par2.x" and "cp.x" (either has some features that the other one has not). To compile them, do "make fpmd" (builds par2.x) and "make cp" (builds cp.x). Documentation is in the "cpdocs" directory. Gerardo Ballabio From georg.heimel at chemistry.gatech.edu Thu Jul 29 19:03:52 2004 From: georg.heimel at chemistry.gatech.edu (Georg Heimel) Date: Thu, 29 Jul 2004 13:03:52 -0400 Subject: [Pw_forum] Workfunction from slab calculations? Message-ID: <001801c4758e$09a439e0$8d24cf82@georglaptop> Hi everyone, Having tried to extract the workfunction from slab calculation performed with PWscf i ran into some unexpected troubles. In principle, the simple version should go like this: - perform slab calculations - get fermi energy from output - plot plane-averaged full electrostatic potential (v_ion+v_hartree+v_c) - get value of potential in the middle of the vacuum region - subtract fermi energy from that - done doing this with pwscf (specifying plot_num=1 in the pp.x input file and using average.x thereafter) yields systematically wrong numbers. What can one do? Seeing that the eigenvalues and fermi energy can be positive and that the values sclae with the inverse of the unit cell volume, i guess that some constants have been dropped in PWscf (like v_c(G=0) or the "alpha"-term from the pseudopotentials). Is there any straightforward way to extract the right numbers? many thanks! PS: by the way: what units is the electrostatic potential given in? It does'nt say anywhere. Georg ======================================== DI Dr. Georg Heimel School of Chemistry & Biochemistry Georgia Institute of Technology Atlanta, GA 30332-0400 USA Phone: ++1/404/894 6456 Fax: ++1/404/894 7452 EMail: georg.heimel at chemistry.gatech.edu Web: http://www.muckisnspirit.com ======================================== From konstantin_kudin at yahoo.com Thu Jul 29 22:52:31 2004 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Thu, 29 Jul 2004 13:52:31 -0700 (PDT) Subject: [Pw_forum] Workfunction from slab calculations? In-Reply-To: <001801c4758e$09a439e0$8d24cf82@georglaptop> Message-ID: <20040729205231.79873.qmail@web52001.mail.yahoo.com> Georg, I am not sure if this is helpful or not, but do take a look at the following work: "Deriving accurate work functions from thin-slab calculations" CJ Fall, N Binggeli, A Baldereschi, J. Phys.: Condens. Matter, 11, 2689, 1999. Perhaps, that paper can clarify things a bit. You could also trace the papers that cite this one with several examples of such calculations. Good luck! Kostya __________________________________ Do you Yahoo!? Yahoo! Mail is new and improved - Check it out! http://promotions.yahoo.com/new_mail From giannozz at nest.sns.it Fri Jul 30 03:54:54 2004 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 30 Jul 2004 03:54:54 +0200 (CEST) Subject: [Pw_forum] Allocation memory In-Reply-To: <601BA98F9E8C7645B7A0052ADBFCC714066EB8BC@atil.intra.cea.fr> References: <601BA98F9E8C7645B7A0052ADBFCC714066EB8BC@atil.intra.cea.fr> Message-ID: On Tue, 20 Jul 2004, BOUYER Fr?d?ric 153746 wrote: > A particular job (on our NEC) with few atoms stops immediately > after few minutes "immediately" or "after few minutes" ?!? if it stops after several scf or optimization steps, it might be a memory leak (memory that it is allocated but not deallocated). Check if the memory usage grows at each scf or optimization step. Either it is something wrong in the FFT driver for NEC, or it is a compiler problem. Paolo From dchrobak at us.edu.pl Fri Jul 30 10:54:54 2004 From: dchrobak at us.edu.pl (Dariusz Chrobak) Date: Fri, 30 Jul 2004 10:54:54 +0200 Subject: [Pw_forum] DOS Message-ID: <410A0CDE.9040308@us.edu.pl> Dear PWscf developers, I wold like to calculate DOS (with tetrahedra) projected on the atomic states. For each energy e and for each k two loops should make which I need: # this can be inserted into attached dos0.f90 (line 234) e=e*13.6058 proj1=0.0; proj2=0.0 inter1=0.0; inter2=0.0 l=0 do i=1,nkpts do j=1,nbnds-1 if ((ebnds(i,j)<=e).and.(ebnds(i,j+1)>=e)) then amp(:)=psi(i,j,:)+(e-ebnds(i,j))/(ebnds(i,j+1)-ebnds(i,j))*(psi(i,j+1,:)-psi(i,j,:)) proj1=proj1+DOSofE(1)*amp/13.6058 proj2=proj2+DOSofE(2)*amp/13.6058 inter1=inter1+DOSofE(1)*(1-sum(amp))/13.6058 inter2=inter2+DOSofE(2)*(1-sum(amp))/13.6058 l=l+1 endif enddo enddo if (l>0) then proj1=proj1/l; inter1=inter1/l proj2=proj2/l; inter2=inter2/l else proj1=0.0; inter1=0.0 proj2=0.0; inter2=0.0 endif # next writing proj1, proj2 and inter to the output file where: nkpts - number of k points nbnds - number of bands ebnds(i,j) - energy psi(i) - coefficient appears in expansion of wave function on atomic states inter - nonlocalized part of DOS What do you think about this small code? Is this a proper way? Best regards Dariusz Chrobak -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: projwfc0.f90 Url: /pipermail/attachments/20040730/a67172c4/attachment.txt -------------- next part -------------- An embedded and charset-unspecified text was scrubbed... Name: dos0.f90 Url: /pipermail/attachments/20040730/a67172c4/attachment-0001.txt From proffess at yandex.ru Fri Jul 30 11:37:39 2004 From: proffess at yandex.ru (Sergei Lisenkov) Date: Fri, 30 Jul 2004 13:37:39 +0400 (MSD) Subject: [Pw_forum] running error on IBM PPC970 In-Reply-To: <4107A421.3050902@sissa.it> References: <200407231800.22326.a.soon@auckland.ac.nz> <41077976.000001.12391@tide.yandex.ru> <41078C33.5020009@sissa.it> <4107982C.00000C.19094@tide.yandex.ru> <41079934.5030309@sissa.it> <4107A1AE.000004.21579@tide.yandex.ru> <4107A421.3050902@sissa.it> Message-ID: <410A16E3.000001.27373@ariel.yandex.ru> Dear PWscf users and authors, I compiled the pw code (I think so). But I have a problems with running code (serial mode). I got segmentation fault: ..... Cartesian axes site n. atom positions (a_0 units) 1 Al tau( 1) = ( 0.5000000 0.5000000 -2.1213200 ) 2 Al tau( 2) = ( 0.0000000 0.0000000 -1.4142130 ) 3 Al tau( 3) = ( 0.5000000 0.5000000 -0.7071070 ) 4 Al tau( 4) = ( 0.0000000 0.0000000 0.0000000 ) 5 Al tau( 5) = ( 0.5000000 0.5000000 0.7071070 ) 6 Al tau( 6) = ( 0.0000000 0.0000000 1.4142130 ) 7 Al tau( 7) = ( 0.5000000 0.5000000 2.1213200 ) number of k points= 3 gaussian broad. (ryd)= 0.0500 ngauss = 1 cart. coord. in units 2pi/a_0 k( 1) = ( 0.1250000 0.1250000 0.0000000), wk = 0.5000000 k( 2) = ( 0.1250000 0.3750000 0.0000000), wk = 1.0000000 k( 3) = ( 0.3750000 0.3750000 0.0000000), wk = 0.5000000 G cutoff = 34.1959 ( 6689 G-vectors) FFT grid: ( 12, 12, 96) nbndx = 60 nbnd = 15 natomwfc = 28 npwx = 860 nelec = 21.00 nkb = 28 ngl = 351 Segmentation fault It is example3. I compiled the code with FFTW, local fftw and ESSL fftw. The result is the same. I used ESSL and lapack libraries. Any suggestions? Thanks, Sergey From lvtieyu1975 at yahoo.com.cn Fri Jul 30 14:21:47 2004 From: lvtieyu1975 at yahoo.com.cn (=?gb2312?q?=BB=AA=20=C0=EE?=) Date: Fri, 30 Jul 2004 20:21:47 +0800 (CST) Subject: [Pw_forum] bandgap of germanium Message-ID: <20040730122147.14543.qmail@web15803.mail.cnb.yahoo.com> dear all: I'm a new user of pwscf code. I successed with calculations for Si in example1. When I tried to calculate Ge, there is a problem. Using other code, we will get nearly 0 ev, even negative. Using pwscf, I get 0.2959 ev. that is ok? ************************************************************************************* I adjust input file of Si to the one of Ge celldm(1)= 10.6769 ATOMIC_SPECIES Ge 72.59 Ge.pz-bhs.UPF ATOMIC_POSITIONS Ge 0.00 0.00 0.00 Ge 0.25 0.25 0.25 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% I get the bands : k = 0.0000 0.0000 0.0000 ( 387 PWs) bands (ev): -6.8856 5.7367 5.7367 5.7367 6.3235 8.2033 8.2033 8.2033 k = 0.1000 0.1000 0.1000 ( 381 PWs) bands (ev): -6.7709 3.9714 5.5040 5.5040 7.0825 8.5187 8.5187 8.9337 k = 0.2000 0.2000 0.2000 ( 386 PWs) bands (ev): -6.4324 1.9466 5.0670 5.0670 6.7871 9.0770 9.0770 10.7253 k = 0.3000 0.3000 0.3000 ( 395 PWs) bands (ev): -5.8950 0.1370 4.6838 4.6838 6.3897 9.3538 9.3538 12.4674 k = 0.4000 0.4000 0.4000 ( 389 PWs) bands (ev): -5.2555 -1.2853 4.4397 4.4397 6.1241 9.3304 9.3304 12.8071 k = 0.5000 0.5000 0.5000 ( 392 PWs) bands (ev): -4.8902 -1.9150 4.3561 4.3561 6.0326 9.2850 9.2850 12.6981 --------------------------------- Do You Yahoo!? ?????????"??"?????????? 100????????????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20040730/6da6e142/attachment.htm From bbouhafs at mail.com Fri Jul 30 18:05:34 2004 From: bbouhafs at mail.com (Bachir Bouhafs) Date: Fri, 30 Jul 2004 17:05:34 +0100 Subject: [Pw_forum] bandgap of germanium Message-ID: <20040730160534.13143164038@ws1-4.us4.outblaze.com> An HTML attachment was scrubbed... URL: /pipermail/attachments/20040730/5ef4b46f/attachment.htm From lvtieyu1975 at yahoo.com.cn Sat Jul 31 00:33:29 2004 From: lvtieyu1975 at yahoo.com.cn (=?gb2312?q?=BB=AA=20=C0=EE?=) Date: Sat, 31 Jul 2004 06:33:29 +0800 (CST) Subject: [Pw_forum] bandgap of germanium In-Reply-To: <20040730160534.13143164038@ws1-4.us4.outblaze.com> Message-ID: <20040730223329.67069.qmail@web15803.mail.cnb.yahoo.com> Dr. B. Bouhafs: 6.3235 is the value at band number 5! Not at band number 4! So the bandgap is 0.2959 ev, not -0.2909 ev. Sincerely your, tylv Bachir Bouhafs wrote: Dear User, When I take a look at your output data, you find that the calculated band gap is negative and equal to -0,2909 eV. This value is the difference between the first conduction band at L(0.5,0.5,0.5) with the corresponding value of 6.0326 (band number 5) and the top of the valence band at gamma (0.,0.,0.0) with the corresponding value of 6.3235 (band number 4). So, this value obtained by the PWSCF code is acceptable, since the ab-initio band gap using other methods for Ge was found negative using both LDA or GGA. Sincerely your, Dr. B. Bouhafs ----- Original Message ----- From: ? ? Date: Fri, 30 Jul 2004 20:21:47 +0800 (CST) To: pw_forum at pwscf.org Subject: [Pw_forum] bandgap of germanium dear all: I'm a new user of pwscf code. I successed with calculations for Si in example1. When I tried to calculate Ge, there is a problem. Using other code, we will get nearly 0 ev, even negative. Using pwscf, I get 0.2959 ev. that is ok? ************************************************************************************* I adjust input file of Si to the one of Ge celldm(1)= 10.6769 ATOMIC_SPECIES Ge 72.59 Ge.pz-bhs.UPF ATOMIC_POSITIONS Ge 0.00 0.00 0.00 Ge 0.25 0.25 0.25 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% I get the bands : k = 0.0000 0.0000 0.0000 ( 387 PWs) bands (ev): -6.8856 5.7367 5.7367 5.7367 6.3235 8.2033 8.2033 8.2033 k = 0.1000 0.1000 0.1000 ( 381 PWs) bands (ev): -6.7709 3.9714 5.5040 5.5040 7.0825 8.5187 8.5187 8.9337 k = 0.2000 0.2000 0.2000 ( 386 PWs) bands (ev): -6.4324 1.9466 5.0670 5.0670 6.7871 9.0770 9.0770 10.7253 k = 0.3000 0.3000 0.3000 ( 395 PWs) bands (ev): -5.8950 0.1370 4.6838 4.6838 6.3897 9.3538 9.3538 12.4674 k = 0.4000 0.4000 0.4000 ( 389 PWs) bands (ev): -5.2555 -1.2853 4.4397 4.4397 6.1241 9.3304 9.3304 12.8071 k = 0.5000 0.5000 0.5000 ( 392 PWs) bands (ev): -4.8902 -1.9150 4.3561 4.3561 6.0326 9.2850 9.2850 12.6981 --------------------------------- Do You Yahoo!? ?????????"??"?????????? 100????????????????? Dr. Bachir BouhafsPhysics DepartmentUniversity of Sidi Bel Abbes22000 Sidi Bel Abbes, AlgeriaMobil:+213 (0)71 18 21 47Fax:+213 48 54 11 52 E-mail: bbouhafs at mail.com -- ___________________________________________________________ Sign-up for Ads Free at Mail.com http://www.mail.com/?sr=signup _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum --------------------------------- Do You Yahoo!? ?????????"??"?????????? 100????????????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20040731/417f8fd5/attachment.htm From bbouhafs at mail.com Sat Jul 31 11:04:23 2004 From: bbouhafs at mail.com (Bachir Bouhafs) Date: Sat, 31 Jul 2004 10:04:23 +0100 Subject: [Pw_forum] bandgap of germanium Message-ID: <20040731090423.652EA6EEE9@ws1-5.us4.outblaze.com> An HTML attachment was scrubbed... URL: /pipermail/attachments/20040731/ec21d92f/attachment.htm