From dxs571 at mail.ustc.edu.cn Thu Sep 1 09:37:07 2005 From: dxs571 at mail.ustc.edu.cn (XiaoSong Du) Date: Thu, 01 Sep 2005 15:37:07 +0800 Subject: [Pw_forum] sorry for the above bug letter Message-ID: <325560227.06588@ustc.edu.cn> In a test of GaAs phonon frequency caculation, at a k-point e.g. (-0.25000 -0.25000 0.25000 ),the outfile has such result as follows, what i can't understand is the meaning of the 4 q-points in the star. Number of q in the star = 4 List of q in the star: 1 -0.250000000 -0.250000000 0.250000000 2 0.250000000 0.250000000 0.250000000 3 0.250000000 -0.250000000 -0.250000000 4 -0.250000000 0.250000000 -0.250000000 In addition there is the -q list: 1 0.250000000 0.250000000-0.250000000 2-0.250000000-0.250000000-0.250000000 3-0.250000000 0.250000000 0.250000000 4 0.250000000-0.250000000 0.250000000 By the way in the example i find the inputfile of non-scf and phonon caculation of GaAs describing k-space sampling is wrong ,which should be a 444 M_P sampling instead of the 2 point.Am I right? ????????????????Xiaosong Du ????????????????dxs571 at mail.ustc.edu.cn ????????????????????2005-08-31 From zfhou at fudan.edu.cn Thu Sep 1 09:42:19 2005 From: zfhou at fudan.edu.cn (Zhufeng Hou) Date: Thu, 01 Sep 2005 15:42:19 +0800 Subject: [Pw_forum] Problem on paralle ph.x Message-ID: <1125560539.4121.16.camel@venus> Dear pwscf users: Recently I have used pwscf2.1.3, which is complied on redhat 8.0 and by PGI fortran 5.1 and mpich, to calculate the phonon of a compound. The paralle pw.x works very well under PBS job queue system. But the paralle ph.x exits after reading the *.save file, and in the output file the following information are printed: -------------BEGIN-------------------------- Program PHONON v.2.1.3 starts ... Today is 1Sep2005 at 14:28:56 Parallel version (MPI) Number of processors in use: 2 R & G space division: nprocp = 2 Ultrasoft (Vanderbilt) Pseudopotentials Reading file tt.save ... read complete Reading file tt.save ... read complete Planes per process (thick) : nr3 = 36 npp = 18 ncplane = 1296 Proc/ planes cols G planes cols G columns G Pool (dense grid) (smooth grid) (wavefct grid) 1 18 378 6902 18 378 6902 127 1319 2 18 379 6901 18 379 6901 126 1318 0 36 757 13803 36 757 13803 253 2637 nbndx = 30 nbnd = 30 natomwfc = 43 npwx = 970 nelec = 38.00 nkb = 60 ngl = 188 ---------------END---------------------- With the same inputs, the serial pw.x and ph.x works very well. What is wrong with the paralle ph.x? How can I fixed it? Your helps will be gratelly appricated! Best wishes! Zhufeng From giannozz at nest.sns.it Thu Sep 1 10:50:10 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Sep 2005 10:50:10 +0200 Subject: [Pw_forum] Problem on paralle ph.x In-Reply-To: <1125560539.4121.16.camel@venus> References: <1125560539.4121.16.camel@venus> Message-ID: <200509011050.10529.giannozz@nest.sns.it> On Thursday 01 September 2005 09:42, Zhufeng Hou wrote: > With the same inputs, the serial pw.x and ph.x works very well. > What is wrong with the parallel ph.x? nothing is wrong, to the best of our knowledge. Try the same input on different parallel machines or with different compilers or different versions of the mpi libraries 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 Sep 1 11:00:36 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Sep 2005 11:00:36 +0200 Subject: [Pw_forum] sorry for the above bug letter In-Reply-To: <325560227.06588@ustc.edu.cn> References: <325560227.06588@ustc.edu.cn> Message-ID: <200509011100.36522.giannozz@nest.sns.it> On Thursday 01 September 2005 09:37, XiaoSong Du wrote: > what i can't understand is the meaning of the 4 q-points in the star. > Number of q in the star = 4 > List of q in the star: > 1 -0.250000000 -0.250000000 0.250000000 > 2 0.250000000 0.250000000 0.250000000 > 3 0.250000000 -0.250000000 -0.250000000 > 4 -0.250000000 0.250000000 -0.250000000 > In addition there is the -q list: > 1 0.250000000 0.250000000-0.250000000 > 2-0.250000000-0.250000000-0.250000000 > 3-0.250000000 0.250000000 0.250000000 > 4 0.250000000-0.250000000 0.250000000 the code uses a rather complex symmetrization, involving the response to several displacement patterns that transform into each other. As a result one gets the dynamical matrix at all the q-vectors that belong to the same star. There are 8 vectors in the star of <1/4, 1/4, 1/4> for cubic materials, and the results for -q and q are the same because of time reversal symmetry. > By the way in the example I find the inputfile of non-scf and phonon > calculation of GaAs describing k-space sampling is wrong, which > should be a 444 M_P sampling instead of the 2 point. Am I right? it is a 444 Monkhorst-Pack sampling, without displacement (the grid has to include q=0) 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 Sep 1 12:13:22 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Sep 2005 12:13:22 +0200 Subject: [Pw_forum] Questions on supercell calculation of MgB2 In-Reply-To: References: Message-ID: <200509011213.22133.giannozz@nest.sns.it> On Monday 29 August 2005 11:40, ? ?? wrote: > I have done a calculation to check the convergence of supercell > method by calculating a 3*3*1 supercell without atom loss, but > the band energy is very different with those of original unit cell you mean "the band energies" ? and what does it mean that "they are different" ? of course they are: the Gamma point of the supercell contains several k-points of the original cell that are refolded into the Gamma point of the supercell. In order to establish an exact mapping of the band energies for the supercell into those of the original cell, you need to compare calculations that use the same grid to sample the respective Brillouin zones. Do this and you will find that the supercell has exactly 9 times the total energy of the original cell and that all eigenvalues at Gamma of the supercell can be mapped into eigenvalues at Gamma + refolded k-points of the original cell. > Further more, when I want to plot the band structure using > bands.x to get the .dat file, there is an error like: > [...] > from punch_band : error # 5 > increase maxdeg the degeneracy of energy bands in a supercell may be quite high. If you really want to plot bands in a supercell, do what it says and recompile > And at the beggining of the scf calculation procedure, there are > "additional transition found" good: this means that the identity plus a fractional translation is a symmetry operation, i.e. that your cell is a supercell. In a supercell all translation vectors that are lattice vectors of the original cell but not of the supercell will appear as fractional translations. 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 shaposh at isp.nsc.ru Thu Sep 1 14:11:49 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Thu, 01 Sep 2005 19:11:49 +0700 Subject: [Pw_forum] CVS version -how to get? Message-ID: <1125576709.3135.4.camel@m00> Dear authors, I would like to test wonders of cvs version, but i am unable to login using instructions in README.cvs file (2.1.5). It never gets past passwd: line. Is the information in README.cvs valid? Thank you. Best Regards, Alexander Shaposhnikov From giannozz at nest.sns.it Thu Sep 1 14:21:42 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Sep 2005 14:21:42 +0200 Subject: [Pw_forum] CVS version -how to get? In-Reply-To: <1125576709.3135.4.camel@m00> References: <1125576709.3135.4.camel@m00> Message-ID: <200509011421.42875.giannozz@nest.sns.it> On Thursday 01 September 2005 14:11, Alexander Shaposhnikov wrote: > I would like to test wonders of cvs version, but i am unable to > login using instructions in README.cvs file (2.1.5). It never gets > past passwd: line. > Is the information in README.cvs valid? Thank you. it is, with one exception: "cvs co fpmd" does not work any longer. Did you define "CVS_RSH" to be "ssh" ? P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From shaposh at isp.nsc.ru Thu Sep 1 14:43:54 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Thu, 1 Sep 2005 19:43:54 +0700 Subject: [Pw_forum] CVS version -how to get? In-Reply-To: <200509011421.42875.giannozz@nest.sns.it> Message-ID: <200509011243.j81ChnUl019182@ns.isp.nsc.ru> Yes I did everything exactly as described in the file. Still cannot login. After asking the passwd (cvsanon) it waits forever. Is there any other way to obtain cvs version? Maybe you could create so-called "nightly builds" on regular basis? Best Regards, Alexander Shaposhnikov -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Thursday, September 01, 2005 7:22 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] CVS version -how to get? On Thursday 01 September 2005 14:11, Alexander Shaposhnikov wrote: > I would like to test wonders of cvs version, but i am unable to > login using instructions in README.cvs file (2.1.5). It never gets > past passwd: line. > Is the information in README.cvs valid? Thank you. it is, with one exception: "cvs co fpmd" does not work any longer. Did you define "CVS_RSH" to be "ssh" ? P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From dxs571 at mail.ustc.edu.cn Thu Sep 1 15:43:26 2005 From: dxs571 at mail.ustc.edu.cn (XiaoSong Du) Date: Thu, 01 Sep 2005 21:43:26 +0800 Subject: [Pw_forum] electron-phonon output explanation Message-ID: <325582206.08885@ustc.edu.cn> In the test of electron-phonon coefficient, the lambda was caculated using several values of gaussian broadening from 0.01-0.1, but the inputfile of non-scf caculation has already gotten a degauss value of 0.022, what's the matter?And how i can control gaussian broadening range? And the output file contains such result Gaussian Broadening: 0.030 Ry, ngauss= 1 DOS = 2.225329 states/spin/Ry/Unit Cell at Ef= 8.330064 eV double delta at Ef = 16.641185 lambda( 1)= 0.192566 gamma= 14.766974 GHz lambda( 2)= 0.192566 gamma= 14.766974 GHz lambda( 3)= 0.205463 gamma= 46.832268 GHz What's the meaning of three gammas in parallel to lambda?Does it have something with the phonon frequency omega( 1) = 6.007176 [THz] = 200.379142 [cm-1] omega( 2) = 6.007176 [THz] = 200.379142 [cm-1] omega( 3) = 10.356665 [THz] = 345.463465 [cm-1] ? Thanks to your reply.Best wishs From dxs571 at mail.ustc.edu.cn Thu Sep 1 16:28:44 2005 From: dxs571 at mail.ustc.edu.cn (XiaoSong Du) Date: Thu, 01 Sep 2005 22:28:44 +0800 Subject: [Pw_forum] question about pdos and dos Message-ID: <325584925.24072@ustc.edu.cn> The output file containing dos and pdos bepuzzles me, according to my knowledge, dos contains all states including shell and valence states but pdos contains only the valence states, but in Pwscf dos equals pdos in the pdos output file.And the dos caculation's column writes like this E (eV) dosup(E) dosdw(E) Int dos(E) 5.749 0.0000E+00 0.0000E+00 0.0000E+00 5.849 0.9326E-03 0.4531E-03 0.1386E-03 What's the meaning of Int dos?And what's the unit of dos, or it's just in a.u. From giannozz at nest.sns.it Thu Sep 1 16:50:18 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Sep 2005 16:50:18 +0200 Subject: [Pw_forum] electron-phonon output explanation In-Reply-To: <325582206.08885@ustc.edu.cn> References: <325582206.08885@ustc.edu.cn> Message-ID: <200509011650.18312.giannozz@nest.sns.it> On Thursday 01 September 2005 15:43, XiaoSong Du wrote: > In the test of electron-phonon coefficient, the lambda was > calculated using several values of gaussian broadening from > 0.01-0.1, but the input file of non-scf calculation has already > gotten a degauss value of 0.022 they serve different purposes and so they are different > And how i can control gaussian broadening range? see file PH/elphon.f90, routine elphsum, parameters 'nsig' and 'degauss1' > What's the meaning of three gammas in parallel to lambda? lambda = gamma/(pi N(Ef) omega^2) There is some documentation on electron-phonon calculations in the manual and in example 7. The answer to many questions can ofter be found in the archives of pw_forum, that are accessible (and searchable) via the www.pwscf.org web site 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 Sep 1 16:55:01 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 1 Sep 2005 16:55:01 +0200 Subject: [Pw_forum] CVS version -how to get? In-Reply-To: <200509011243.j81ChnUl019182@ns.isp.nsc.ru> References: <200509011243.j81ChnUl019182@ns.isp.nsc.ru> Message-ID: <200509011655.02000.giannozz@nest.sns.it> On Thursday 01 September 2005 14:43, Alexander Shaposhnikov wrote: > Yes I did everything exactly as described in the file. > Still cannot login. After asking the passwd (cvsanon) > it waits forever. it's looks like a network problem: if your machine is properly configured, there is little that can be done, except by a knowledgeable network engineer (if you manage to find one) 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 dxs571 at mail.ustc.edu.cn Thu Sep 1 17:11:01 2005 From: dxs571 at mail.ustc.edu.cn (XiaoSong Du) Date: Thu, 01 Sep 2005 23:11:01 +0800 Subject: [Pw_forum] (no subject) Message-ID: <325587461.03922@ustc.edu.cn> sorry for the repeating stupid question about pdos and dos , and the question of lamda and gamma in el-ph calculation.I have gotten useful information from latest archive. Thank you for your patience and replys. From leoant21 at hotmail.com Fri Sep 2 10:07:10 2005 From: leoant21 at hotmail.com (=?gb2312?B?1cUguumx8g==?=) Date: Fri, 02 Sep 2005 08:07:10 +0000 Subject: [Pw_forum] Questions on supercell calculation of MgB2 In-Reply-To: <20050902053648.25079.17480.Mailman@democritos.sissa.it> Message-ID: Dear Paolo: Thank you very much for your reply. However, I am not familiar with supercell calculation, especially how "several k-points of the original cell" refold into the Gamma point of the supercell. I can't find any documents related to this topic and I cannot image it at all.Further more, on how to "compare calculations that use the same grid to sample the respective Brillouin zones", would you give me some further advice? What does "the same grid" meaning? With the same deltak? Or something else? Please help me. With my kindest regards. Hongbin Zhang >Message: 5 >From: Paolo Giannozzi >Organization: Scuola Normale Superiore di Pisa >To: pw_forum at pwscf.org >Subject: Re: [Pw_forum] Questions on supercell calculation of MgB2 >Date: Thu, 1 Sep 2005 12:13:22 +0200 >Reply-To: pw_forum at pwscf.org > >On Monday 29 August 2005 11:40, =D5=C5 =BA=E9=B1=F2 wrote: > > > I have done a calculation to check the convergence of supercell > > method by calculating a 3*3*1 supercell without atom loss, but=20 > > the band energy is very different with those of original unit cell > >you mean "the band energies" ? and what does it mean that=20 >"they are different" ? of course they are: the Gamma point of=20 >the supercell contains several k-points of the original cell that=20 >are refolded into the Gamma point of the supercell. >In order to establish an exact mapping of the band energies for=20 >the supercell into those of the original cell, you need to compare >calculations that use the same grid to sample the respective=20 >Brillouin zones. Do this and you will find that the supercell has=20 >exactly 9 times the total energy of the original cell and that all >eigenvalues at Gamma of the supercell can be mapped into >eigenvalues at Gamma + refolded k-points of the original cell. > > > Further more, when I want to plot the band structure using > > bands.x to get the .dat file, there is an error like: > > [...] > > from punch_band : error # 5 > > increase maxdeg > >the degeneracy of energy bands in a supercell may be quite high.=20 >If you really want to plot bands in a supercell, do what it says and >recompile > > > And at the beggining of the scf calculation procedure, there are > > "additional transition found" > >good: this means that the identity plus a fractional translation=20 >is a symmetry operation, i.e. that your cell is a supercell. In a >supercell all translation vectors that are lattice vectors of the=20 >original cell but not of the supercell will appear as fractional >translations.=20 > >Paolo > >=2D-=20 >Paolo Giannozzi e-mail: giannozz at nest.sns.it >Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513=20 >Piazza dei Cavalieri 7 I-56126 Pisa, Italy > >--__--__-- > _________________________________________________________________ ???? MSN Explorer: http://explorer.msn.com/lccn/ From g.ballabio at cineca.it Fri Sep 2 10:38:09 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Fri, 2 Sep 2005 10:38:09 +0200 (MEST) Subject: [Pw_forum] Questions on supercell calculation of MgB2 In-Reply-To: (from leoant21@hotmail.com on Fri Sep 2 10:07:10 2005) References: Message-ID: <1125650287l.4219l.0l@nb-ballabio.cineca.it> On 09/02/2005 10:07:10 AM, ? ?? wrote: > Dear Paolo: > Thank you very much for your reply. However, I am not familiar > with supercell calculation, especially how "several k-points of the > original cell" refold into the Gamma point of the supercell. This is standard solid state theory. Basically, a Bravais lattice in direct space is identified by its three generating vectors, let's call them a_1, a_2, a_3. The reciprocal lattice in k-space is then generated by the three vectors b_1, b_2, b_3 such that: a_i * b_j = 2 pi delta_ij where "*" means the dot product (a*b = ax bx + ay by + az bz) and delta_ij is Kronecker delta, that is, 1 if i=j and 0 otherwise. You can easily see that as the "a" grow larger, the "b" become smaller, and in particular if you take a supercell, the reciprocal space cell is a SUBcell of the original one. Now, all k-points that only differ by a reciprocal lattice vector (that is, a linear combination of b_1, b_2, b_3 with integer coefficients) are equivalent. This means that you can "refold" k-points, that is, replace each of them with its equivalent point *that lies in the unit cell*. If the supercell is N times larger than the original cell, there is N-to-1 correspondence, that is, N k-points that were distinct in the original reciprocal cell refold into THE SAME k-point in the reciprocal of the supercell. And since each of those N points had different energy levels, every band in the original cell splits into N bands in the supercell. Gerardo From yuwen_66 at yahoo.com Mon Sep 5 11:34:43 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Mon, 5 Sep 2005 02:34:43 -0700 (PDT) Subject: [Pw_forum] FPMD segmentation fault Message-ID: <20050905093443.55394.qmail@web51011.mail.yahoo.com> Dear CPMD users, I tried example21 with CPMD for H2O. The complilation with ifort gives no erro. But run_example terminated with the errer message quoted below: Running the calculation with 32 H2O molecules...forrtl:severe(174):SIGSEGV,segmentation fault occurred. When I tried the same example with the FPMD 2.04 version, the calculation stopped normally although with somewhat different contents in the output file. Can anybody help me with this problem? Thanks! W. YU ______________________________________________________ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From chaohao2002 at 163.com Tue Sep 6 11:46:41 2005 From: chaohao2002 at 163.com (=?gb2312?B?Q2hhb2hhbyBIdQ==?=) Date: Tue, 6 Sep 2005 17:46:41 +0800 (CST) Subject: [Pw_forum] =?gb2312?B?YSBxdWVzdGlvbiBhYm91dCB0aGUgcmVsYXRpb25zaGlwIGJldHdlZW4=?= =?gb2312?B?IEZlcm1pIGVuZXJneSBhbmQgc3RhYmlsaXR5?= Message-ID: <431D6581.0000BC.14774@bj163app55.163.com> Dear pwscf users, During my calculations, if considering different crystal structure for a certain material, I obtained different total energies and fermi energies. For example, for structure 1, its equilibrium total energy is 119.5 Ry and fermi energy is 8.0 Ry; for structure 2, its total energy is 119.1 Ry, but its fermi energy is close to 0.0 Ry. Which is the more stable structure? Now I am confused to this. Who can explain more details about this problem? Thanks in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050906/dc81e0f6/attachment.htm From chaohao2002 at 163.com Tue Sep 6 12:09:49 2005 From: chaohao2002 at 163.com (=?gb2312?B?Q2hhb2hhbyBIdQ==?=) Date: Tue, 6 Sep 2005 18:09:49 +0800 (CST) Subject: [Pw_forum] =?gb2312?B?UmU6IFtQd19mb3J1bV1hIHF1ZXN0aW9uIGFib3V0IHRoZSByZWxhdGk=?= =?gb2312?B?b25zaGlwIGJldHdlZW4gRmVybWkgZW5lcmd5IGFuZCBzdGFiaWxpdHk=?= Message-ID: <431D6AED.0000DF.14774@bj163app55.163.com> Sorry, the unit of Fermi energy is eV. Dear pwscf users, During my calculations, if considering different crystal structure for a certain material, I obtained different total energies and fermi energies. For example, for structure 1, its equilibrium total energy is 119.5 Ry and fermi energy is 8.0 Ry; for structure 2, its total energy is 119.1 Ry, but its fermi energy is close to 0.0 Ry. Which is the more stable structure? Now I am confused to this. Who can explain more details about this problem? Thanks in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050906/6107a37c/attachment.htm From eyvaz_isaev at yahoo.com Tue Sep 6 13:12:02 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Tue, 6 Sep 2005 04:12:02 -0700 (PDT) Subject: [Pw_forum] Re: [Pw_forum]a question about the relationship between Fermi energy and stability In-Reply-To: <431D6AED.0000DF.14774@bj163app55.163.com> Message-ID: <20050906111203.29282.qmail@web60323.mail.yahoo.com> Hi, The Fermi energy you calculate is not directly related to the crystal stability until you study its position on the density of states. According to the Hume-Rothery rule the Fermi level is located in the (pseudo)minimum of the DOS, reducing the total energy. I assume it is difficult to say which structure is more stable based on only total energies, because there is no details of structures you studied. Besides, hopefully, you have not 119.5Ry, but -119.5Ry. Bests, Eyvaz. --- Chaohao Hu wrote: > Sorry, the unit of Fermi energy is eV. > > Dear pwscf users, > During my calculations, if considering different > crystal structure for a certain material, I obtained > different total energies and fermi energies. > For example, for structure 1, its equilibrium total > energy is 119.5 Ry and fermi energy is 8.0 Ry; for > structure 2, its total energy is 119.1 Ry, but its > fermi energy is close to 0.0 Ry. > Which is the more stable structure? > Now I am confused to this. Who can explain more > details about this problem? > Thanks in advance. > ______________________________________________________ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From suriano at sissa.it Tue Sep 6 19:32:43 2005 From: suriano at sissa.it (Antonio Suriano) Date: Tue, 6 Sep 2005 19:32:43 +0200 Subject: [Pw_forum] fftw3 In-Reply-To: <325584925.24072@ustc.edu.cn> References: <325584925.24072@ustc.edu.cn> Message-ID: <8DE6D158-AFBC-4983-B8C2-C13741317DD6@sissa.it> Hi Pw is still written using fftw 2.1.5 or something like that .. is there a way to link it to fftw3 or is in project to switch to fftw3? This can be interesting because usually fftw3 are best optimized for sse2 sse3 altivec and other multi-core (or something like that ... i am not an expert) architecture and this can be a speed improvement (50%?) for espresso that makes a lot of fft. Antonio Suriano Phd student in Condensed Matter to Sissa/ISAS Trieste suriano at sissa.it antoniosuriano at yahoo.it Room 205 new building +39 347 8577081 (Vodafone - main number and MSN contact) +39 338 5248106 (TIM - send only SMS at this number please) From yuwen_66 at yahoo.com Wed Sep 7 09:29:22 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Wed, 7 Sep 2005 00:29:22 -0700 (PDT) Subject: [Pw_forum] phonon spectra Brillouin-zone integration In-Reply-To: <20050903053644.29490.65553.Mailman@democritos.sissa.it> Message-ID: <20050907072922.95528.qmail@web51001.mail.yahoo.com> Dear PW users, When calculating thermodynamics properties under the name of quasiharmonic approximation, in principle,one needs only to sum over all the Brillouin zone wave vectors and the band index of the phonon dispersions. But in practice, the phonon dispersion curves along high symmetry lines are not adequate. In this case, should we integrate using the phonon density of state or we should find phonon frequencies and the appropriate weights for each k point at a three dimensional regular grid and make a sum? Any suggestions would be much appreciated. Thanks, W. YU __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From giannozz at nest.sns.it Wed Sep 7 09:52:57 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 7 Sep 2005 09:52:57 +0200 Subject: [Pw_forum] fftw3 In-Reply-To: <8DE6D158-AFBC-4983-B8C2-C13741317DD6@sissa.it> References: <325584925.24072@ustc.edu.cn> <8DE6D158-AFBC-4983-B8C2-C13741317DD6@sissa.it> Message-ID: <200509070952.57211.giannozz@nest.sns.it> On Tuesday 06 September 2005 19:32, Antonio Suriano wrote: > Pw is still written using fftw 2.1.5 or something like that ... > is there a way to link it to fftw3 no, because v.3 has different calling sequences from v.2 > or is in project to switch to fftw3? there is a modified fft_scalar module, written by Pascal Thibaudeau et al., that works with fftw3 on Alpha and SGI machines, but not on Intel PCs, for unknown reasons. Apparently fftw3 is slower than fftw2 on SGI machines, faster on Alpha but only if memory is allocated in such a way to be aligned with pages. Anybody interested in using fftw3 can start from there, if the author(s) of the modified module agree in distributing it. 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 shaposh at isp.nsc.ru Wed Sep 7 09:58:26 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Wed, 7 Sep 2005 14:58:26 +0700 Subject: [Pw_forum] fftw3 In-Reply-To: <200509070952.57211.giannozz@nest.sns.it> Message-ID: <200509070758.j877wQOr023419@ns.isp.nsc.ru> Fftw project is not being developed anymore. I am having problems with compiling both fftw2 and fftw3 on my machines. Maybe it would be better to implement support for vendor-specific fft implementations, for example in mkl, acml and so on. They should be faster anyway and there will be continuing development. Best Regards, Alexander Shaposhnikov -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Paolo Giannozzi Sent: Wednesday, September 07, 2005 2:53 PM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] fftw3 On Tuesday 06 September 2005 19:32, Antonio Suriano wrote: > Pw is still written using fftw 2.1.5 or something like that ... > is there a way to link it to fftw3 no, because v.3 has different calling sequences from v.2 > or is in project to switch to fftw3? there is a modified fft_scalar module, written by Pascal Thibaudeau et al., that works with fftw3 on Alpha and SGI machines, but not on Intel PCs, for unknown reasons. Apparently fftw3 is slower than fftw2 on SGI machines, faster on Alpha but only if memory is allocated in such a way to be aligned with pages. Anybody interested in using fftw3 can start from there, if the author(s) of the modified module agree in distributing it. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From hushujun at mail.sdu.edu.cn Wed Sep 7 11:27:22 2005 From: hushujun at mail.sdu.edu.cn (Shujun Hu) Date: Wed, 07 Sep 2005 17:27:22 +0800 Subject: [Pw_forum] How to plot charge density for different orbitals? Message-ID: <326085242.15792@mail.sdu.edu.cn> Dear Drs, I want to know how to plot the charge or spin density for different orbitals, like 3d, 4s, even T2g in crystal field. As the projwfc.x can only manage the total charge or spin density. Best wishes Shujun Hu ------------------------------ -- Shujun Hu e-mail: hushujun at 163.com Shandong university Phone: +86/0531-88375097 Jinan, Shandong Province, China, 250100 From hushujun at mail.sdu.edu.cn Wed Sep 7 11:32:31 2005 From: hushujun at mail.sdu.edu.cn (Shujun Hu) Date: Wed, 07 Sep 2005 17:32:31 +0800 Subject: [Pw_forum] How to plot charge density for different orbitals? Message-ID: <326085551.16264@mail.sdu.edu.cn> Dear all, I am sorry. It should be chdens.x, no projwfc.x in the upper letter. Shujun Hu ------------------------------ -- Shujun Hu e-mail: hushujun at 163.com Shandong university Phone: +86/0531-88375097 Jinan, Shandong Province, China, 250100 From chaohao2002 at 163.com Wed Sep 7 12:15:35 2005 From: chaohao2002 at 163.com (Chaohao Hu) Date: Wed, 7 Sep 2005 18:15:35 +0800 Subject: [Pw_forum] How to solve this problem Message-ID: <20050907101223.9C5751127D0@democritos.sissa.it> Dear Sir, During my cacluations about TiH2, the SCF calculations is OK, But when calculating DOS and PDOS, even I modified my makefile many times after reading manul file, some wrong information about segmentation fault described below still appears in my bash terminal. My system is Redhat Linux9.0, the complier is intel-fortran8.1 and numerical lib is intel-mkl721. In addition, the unit of TiH2 only include three atoms, a primitive cell. In my previous calculations about other systems, a larger cell is also considered and the program runs very well. How to solve this problem? ============================== ....... running the scf calculation for TiH2... done running DOS calculation for TiH2...forrtl: severe (174): SIGSEGV, segmentation fault occurred Image PC Routine Line Source pw.x 084A953B Unknown Unknown Unknown pw.x 084BC364 Unknown Unknown Unknown pw.x 080613D5 Unknown Unknown Unknown pw.x 08060619 Unknown Unknown Unknown pw.x 0804B2B9 Unknown Unknown Unknown pw.x 0804B048 Unknown Unknown Unknown libc.so.6 42015574 Unknown Unknown Unknown pw.x 0804AF01 Unknown Unknown Unknown forrtl: severe (174): SIGSEGV, segmentation fault occurred Image PC Routine Line Source libc.so.6 4207A42B Unknown Unknown Unknown libc.so.6 4206432C Unknown Unknown Unknown libc.so.6 4204F18D Unknown Unknown Unknown pw.x 084A8DF3 Unknown Unknown Unknown pw.x 084A8B33 Unknown Unknown Unknown pw.x 084AA393 Unknown Unknown Unknown pw.x 084A9486 Unknown Unknown Unknown pw.x 084AA998 Unknown Unknown Unknown libpthread.so.0 40060610 Unknown Unknown Unknown pw.x 084BC364 Unknown Unknown Unknown pw.x 080613D5 Unknown Unknown Unknown ...... =========================== From chesrunners at yahoo.com Wed Sep 7 14:01:05 2005 From: chesrunners at yahoo.com (Matt Small) Date: Wed, 7 Sep 2005 05:01:05 -0700 (PDT) Subject: [Pw_forum] Grid Computing In-Reply-To: <20050907053634.17959.51196.Mailman@democritos.sissa.it> Message-ID: <20050907120105.69966.qmail@web30509.mail.mud.yahoo.com> Dear pwscf users, Do any of the programs in espresso happen to be compatible with grid computing (e.g. pw.x, fpmd.x, etc.)? Some of the IT people at our university are planning to set up a grid and are looking for programs that are capable of this type of dispersment. Thank you in advance, Matt Small __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From dxl at ustc.edu.cn Thu Sep 8 11:23:58 2005 From: dxl at ustc.edu.cn (XunLei Ding) Date: Thu, 08 Sep 2005 17:23:58 +0800 Subject: [Pw_forum] scan the lattice constant Message-ID: <326171438.22436@ustc.edu.cn> Dear all, I do a calculation to get the lattice constant of Ni crystal. I scan the lattice constant from 6.60 to 6.70 a.u. using the bash file below. The experimental value is 6.66 and is consistent with my result. But it is clear that the curve departs into two parts at 6.66. (see the picture at http://www.bsc.ustc.edu.cn/~dxl/download/Graph1.JPG) Would you tell my why and how to solve this problem. Thank you! Ding Xunlei for a0 in 6.60 6.61 6.62 6.63 6.64 6.65 6.66 6.67 6.68 6.69 6.70 do cat >test.in<test.in<test.in< Message-ID: <20050908100357.39996.qmail@web60325.mail.yahoo.com> Hi, Presumably, you should use more k-points, and may be larger cutof energy. More accurate calculations (conv_thr) also might help. Bests, Eyvaz --- XunLei Ding wrote: > Dear all, > I do a calculation to get the lattice constant of Ni > crystal. > I scan the lattice constant from 6.60 to 6.70 a.u. > using the bash file below. > The experimental value is 6.66 and is consistent > with my result. > But it is clear that the curve departs into two > parts at 6.66. > (see the picture at > http://www.bsc.ustc.edu.cn/~dxl/download/Graph1.JPG) > Would you tell my why and how to solve this problem. > > Thank you! > > Ding Xunlei > > for a0 in 6.60 6.61 6.62 6.63 6.64 6.65 6.66 6.67 > 6.68 6.69 6.70 > do > cat >test.in< &control > calculation='scf' > restart_mode='from_scratch', > pseudo_dir = '/home/bsc/dxl/pseudo/', > outdir='./' > prefix='zz' > tprnfor = .true., > for a0 in 6.64 > do > cat >test.in< &control > calculation='scf' > for a0 in 6.65 > do > cat >test.in< &control > calculation='scf' > restart_mode='from_scratch', > pseudo_dir = '/home/bsc/dxl/pseudo/', > outdir='./' > prefix='zz' > tprnfor = .true., > tstress = .true. > / > &system > ibrav=2, celldm(1) =$a0 > nat=1,ntyp=1, > nspin = 2, starting_magnetization(1)=0.7, > ecutwfc = 24.0, ecutrho = 288.0, > occupations='smearing', > smearing='methfessel-paxton', degauss=0.01 > / > &electrons > electron_maxstep= 200 > diagonalization='' > conv_thr = 1.0e-8 > mixing_beta = 0.2 > &cell > cell_dynamics = 'none' > cell_factor = 1.2 > / > &ions > ion_dynamics = 'damp' > > / > ATOMIC_SPECIES > Ni 58.69 Ni.pbe-nd-rrkjus.UPF > ATOMIC_POSITIONS > Ni 0.00000000 0.000000000 0.000000000 1 > 1 1 > K_POINTS (automatic) > 13 13 13 1 1 1 > ! > > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > ______________________________________________________ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From fornari at phy.cmich.edu Thu Sep 8 13:35:27 2005 From: fornari at phy.cmich.edu (Marco Fornari) Date: Thu, 8 Sep 2005 07:35:27 -0400 (EDT) Subject: [Pw_forum] scan the lattice constant In-Reply-To: <326171438.22436@ustc.edu.cn> References: <326171438.22436@ustc.edu.cn> Message-ID: <33394.24.231.203.142.1126179327.squirrel@phy.cmich.edu> Hello, it may be a problem with the augmented charges of the pseudopotential. Try to increase ecutrho to explore this possibilities. Best, Marco > Dear all, > I do a calculation to get the lattice constant of Ni crystal. > I scan the lattice constant from 6.60 to 6.70 a.u. using the bash file > below. The experimental value is 6.66 and is consistent with my result. > But it is clear that the curve departs into two parts at 6.66. > (see the picture at http://www.bsc.ustc.edu.cn/~dxl/download/Graph1.JPG) > Would you tell my why and how to solve this problem. > > Thank you! > > Ding Xunlei > > for a0 in 6.60 6.61 6.62 6.63 6.64 6.65 6.66 6.67 6.68 6.69 6.70 > do > cat >test.in< &control > calculation='scf' > restart_mode='from_scratch', > pseudo_dir = '/home/bsc/dxl/pseudo/', > outdir='./' > prefix='zz' > tprnfor = .true., > for a0 in 6.64 > do > cat >test.in< &control > calculation='scf' > for a0 in 6.65 > do > cat >test.in< &control > calculation='scf' > restart_mode='from_scratch', > pseudo_dir = '/home/bsc/dxl/pseudo/', > outdir='./' > prefix='zz' > tprnfor = .true., > tstress = .true. > / > &system > ibrav=2, celldm(1) =$a0 > nat=1,ntyp=1, > nspin = 2, starting_magnetization(1)=0.7, > ecutwfc = 24.0, ecutrho = 288.0, > occupations='smearing', smearing='methfessel-paxton', degauss=0.01 > / > &electrons > electron_maxstep= 200 > diagonalization='' > conv_thr = 1.0e-8 > mixing_beta = 0.2 > &cell > cell_dynamics = 'none' > cell_factor = 1.2 > / > &ions > ion_dynamics = 'damp' > > / > ATOMIC_SPECIES > Ni 58.69 Ni.pbe-nd-rrkjus.UPF > ATOMIC_POSITIONS > Ni 0.00000000 0.000000000 0.000000000 1 1 1 > K_POINTS (automatic) > 13 13 13 1 1 1 > ! > > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum From marzari at MIT.EDU Thu Sep 8 13:57:45 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Thu, 08 Sep 2005 07:57:45 -0400 Subject: [Pw_forum] scan the lattice constant In-Reply-To: <326171438.22436@ustc.edu.cn> References: <326171438.22436@ustc.edu.cn> Message-ID: <43202739.8090601@mit.edu> XunLei, most likely it's a cutoff issue. Years ago struggling for cutoff convergence was quite routine - there is even a paper by MCPayne and coworkers, in the early 90s on JPhysCondMatt, discussing how to correct approximately for too-small-a-cutoff. Basically, as you enalarge your unit cell, your brillouin zone shrinks, and at constant cutoff it means that more lattice vectors (i.e. plane waves) enter into the fixed cutoff sphere. More plane waves means a systematically larger basis set (this is one of the good things of plane waves, it's easy to make the basis set more and more complete - a nightmare in Gaussian), and, variationally, a lower energy. You see in fact your energy drop going to the left. Check then what is the number of plane waves for each of your calculations, but most likely the wfc cutoff is the culprit. Your ecutrho corresponds to a dual of 12 (ecutrho=dual*ecutwfc), that is a very safe value, although higher values are occasionally needed for accurate magnetic properties (and lower values, down to 6, can be sometimes used). My rule of thumb is to start from 8, and repeat calculations with 6, 10, and 12, to figure out what works. (See e.g. the tests at: http://www.pwscf.org/pseudo/upfdetails.php?upf=C.pbe-rrkjus.UPF ). Note that if you do not have ultrasoft pseudopotentials in your calculations, a dual of 4 is the most you need (and, for super-large scale calculations, we used to play around with a dual of 3, again ages ago). This is due to the fact that if the wavefunctions are expanded up to gmax, the charge density (that is the square of the wavefunctions, in real space) will have in reciprocal space components up to 2 gmax (rho(r)=sum g=1,gmax, g'=1,gmax psi_g^start psi_g' exp i(g-g')r , and g-g' can then be anywhere between -2 gmax and +2 gmax) . The problem you see is particularly relevant in variable cell calculations (you do not want discontiuities while your cell breathes), and Sandro Scandolo (one of the fpmd authors, in quantum-espresso) has worked out a smearing scheme a few years ago (published, I think prb) that smooths away those disconituities. Happy computing, nicola XunLei Ding wrote: > Dear all, > I do a calculation to get the lattice constant of Ni crystal. > I scan the lattice constant from 6.60 to 6.70 a.u. using the bash file below. > The experimental value is 6.66 and is consistent with my result. > But it is clear that the curve departs into two parts at 6.66. > (see the picture at http://www.bsc.ustc.edu.cn/~dxl/download/Graph1.JPG) > Would you tell my why and how to solve this problem. > > Thank you! > > Ding Xunlei > > for a0 in 6.60 6.61 6.62 6.63 6.64 6.65 6.66 6.67 6.68 6.69 6.70 > do > cat >test.in< &control > calculation='scf' > restart_mode='from_scratch', > pseudo_dir = '/home/bsc/dxl/pseudo/', > outdir='./' > prefix='zz' > tprnfor = .true., > for a0 in 6.64 > do > cat >test.in< &control > calculation='scf' > for a0 in 6.65 > do > cat >test.in< &control > calculation='scf' > restart_mode='from_scratch', > pseudo_dir = '/home/bsc/dxl/pseudo/', > outdir='./' > prefix='zz' > tprnfor = .true., > tstress = .true. > / > &system > ibrav=2, celldm(1) =$a0 > nat=1,ntyp=1, > nspin = 2, starting_magnetization(1)=0.7, > ecutwfc = 24.0, ecutrho = 288.0, > occupations='smearing', smearing='methfessel-paxton', degauss=0.01 > / > &electrons > electron_maxstep= 200 > diagonalization='' > conv_thr = 1.0e-8 > mixing_beta = 0.2 > &cell > cell_dynamics = 'none' > cell_factor = 1.2 > / > &ions > ion_dynamics = 'damp' > > / > ATOMIC_SPECIES > Ni 58.69 Ni.pbe-nd-rrkjus.UPF > ATOMIC_POSITIONS > Ni 0.00000000 0.000000000 0.000000000 1 1 1 > K_POINTS (automatic) > 13 13 13 1 1 1 > ! > > > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From giannozz at nest.sns.it Thu Sep 8 14:20:48 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 8 Sep 2005 14:20:48 +0200 Subject: [Pw_forum] How to plot charge density for different orbitals? In-Reply-To: <326085242.15792@mail.sdu.edu.cn> References: <326085242.15792@mail.sdu.edu.cn> Message-ID: <200509081420.48933.giannozz@nest.sns.it> On Wednesday 07 September 2005 11:27, Shujun Hu wrote: > I want to know how to plot the charge or spin density for different > orbitals, like 3d, 4s, even T2g in crystal field. As the projwfc.x can > only manage the total charge or spin density. it is not presently implemented. One should modify projwfc.x to calculate and write the projected charge density in the format of pp.x, then plot it in the usual way 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 Sep 8 14:24:53 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 8 Sep 2005 14:24:53 +0200 Subject: [Pw_forum] How to solve this problem In-Reply-To: <20050907101223.9C5751127D0@democritos.sissa.it> References: <20050907101223.9C5751127D0@democritos.sissa.it> Message-ID: <200509081424.54009.giannozz@nest.sns.it> On Wednesday 07 September 2005 12:15, Chaohao Hu wrote: > running DOS calculation for TiH2...forrtl: severe (174): SIGSEGV, > segmentation fault occurred try to understand where the error occurs; to unset limits for the stack ("limits stack unlimited", or "ulimit -s unlimited"); check if you are close to the maximum available memory; try a different compiler (for instance g95) 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 eyvaz_isaev at yahoo.com Thu Sep 8 14:44:34 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Thu, 8 Sep 2005 05:44:34 -0700 (PDT) Subject: [Pw_forum] scan the lattice constant In-Reply-To: <33394.24.231.203.142.1126179327.squirrel@phy.cmich.edu> Message-ID: <20050908124434.79247.qmail@web60311.mail.yahoo.com> Hi again, I assume total energy calculations are more accurate using the tetrahedra method rather than MP k-points. For "smearing" option one has to check also "degauss". Bests, Eyvaz. --- Marco Fornari wrote: > Hello, > > it may be a problem with the augmented charges of > the pseudopotential. > Try to increase ecutrho to explore this > possibilities. > Best, > Marco > > > Dear all, > > I do a calculation to get the lattice constant of > Ni crystal. > > I scan the lattice constant from 6.60 to 6.70 a.u. > using the bash file > > below. The experimental value is 6.66 and is > consistent with my result. > > But it is clear that the curve departs into two > parts at 6.66. > > (see the picture at > http://www.bsc.ustc.edu.cn/~dxl/download/Graph1.JPG) > > Would you tell my why and how to solve this > problem. > > > > Thank you! > > > > Ding Xunlei > > > > for a0 in 6.60 6.61 6.62 6.63 6.64 6.65 6.66 6.67 > 6.68 6.69 6.70 > > do > > cat >test.in< > &control > > calculation='scf' > > restart_mode='from_scratch', > > pseudo_dir = '/home/bsc/dxl/pseudo/', > > outdir='./' > > prefix='zz' > > tprnfor = .true., > > for a0 in 6.64 > > do > > cat >test.in< > &control > > calculation='scf' > > for a0 in 6.65 > > do > > cat >test.in< > &control > > calculation='scf' > > restart_mode='from_scratch', > > pseudo_dir = '/home/bsc/dxl/pseudo/', > > outdir='./' > > prefix='zz' > > tprnfor = .true., > > tstress = .true. > > / > > &system > > ibrav=2, celldm(1) =$a0 > > nat=1,ntyp=1, > > nspin = 2, starting_magnetization(1)=0.7, > > ecutwfc = 24.0, ecutrho = 288.0, > > occupations='smearing', > smearing='methfessel-paxton', degauss=0.01 > > / > > &electrons > > electron_maxstep= 200 > > diagonalization='' > > conv_thr = 1.0e-8 > > mixing_beta = 0.2 > > &cell > > cell_dynamics = 'none' > > cell_factor = 1.2 > > / > > &ions > > ion_dynamics = 'damp' > > > > / > > ATOMIC_SPECIES > > Ni 58.69 Ni.pbe-nd-rrkjus.UPF > > ATOMIC_POSITIONS > > Ni 0.00000000 0.000000000 0.000000000 > 1 1 1 > > K_POINTS (automatic) > > 13 13 13 1 1 1 > > ! > > > > > > > > > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > ______________________________________________________ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From konstantin_kudin at yahoo.com Thu Sep 8 21:05:10 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Thu, 8 Sep 2005 12:05:10 -0700 (PDT) Subject: [Pw_forum] version 2.1.5 - strange crashes Message-ID: <20050908190511.76802.qmail@web52009.mail.yahoo.com> Hi there, I got a report that compiling Espresso 2.1.5 under Linux with either IFC 8.0 or 9.0 (initial free 32-bit versions, no MPI) gives some random crashes at the end of nlset_base in pseudo_base.f line 264. This happens when running the test suite supplied with the package. Has anyone else encountered something like that? On one hand, initial Intel compiler versions are not known to be that reliable, however, it is a bit suspicious if 2 different compiler versions crash the code in a similar fashion. Any ideas/suggestions? Kostya __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From konstantin_kudin at yahoo.com Thu Sep 8 21:17:27 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Thu, 8 Sep 2005 12:17:27 -0700 (PDT) Subject: [Pw_forum] scan the lattice constant In-Reply-To: <43202739.8090601@mit.edu> Message-ID: <20050908191728.27590.qmail@web52003.mail.yahoo.com> > Basically, as you enalarge your unit cell, your brillouin zone > shrinks, > and at constant cutoff it means that more lattice vectors (i.e. plane > waves) enter into the fixed cutoff sphere. More plane waves means a > systematically larger basis set (this is one of the good things > of plane waves, it's easy to make the basis set more and more > complete - > a nightmare in Gaussian), and, variationally, a lower energy. You see > in > fact your energy drop going to the left. I'd like to chime in here. While indeed reaching the basis set limit is sort of difficult with Gaussian basis sets, there is this powerful principle called "cancellation of errors". This allows one to chase not the full convergence, but convergence with the basis set that is good enough. As it turns out, such convergence is actually rather easy to achieve. Then one has to care that the energy is fully variational and smooth (i.e. differentiable) with respect to changes in atomic coordinates and lattice vectors. While indeed PWs do reach the basis set limits easier, the smoothness of the potential energy surfaces on the other hand leaves something to be desired. The smoothness issue is usually much better taken care of with Gaussians. So there you go ... :-) Kostya ______________________________________________________ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From baroni at sissa.it Thu Sep 8 21:44:33 2005 From: baroni at sissa.it (Stefano Baroni) Date: Thu, 8 Sep 2005 21:44:33 +0200 Subject: [Pw_forum] scan the lattice constant In-Reply-To: <20050908191728.27590.qmail@web52003.mail.yahoo.com> References: <20050908191728.27590.qmail@web52003.mail.yahoo.com> Message-ID: Kostya: I beg to differ here. I do not want to open a thread "PW's vs. Gaussians" which would have the same chances to be settled as "coca vs. pepsi", but what you say, while perfectly (kind of) clear to me, may be misleading to the less experienced. PW potential energy surfaces (PES) are as smooth as they can be (i.e. *perfectly smooth*, or of class C(\infty), if you prefer) *with fixed boundary conditions* (i.e. for a fixed shape/size of the unit cell which determines the periodic boundary conditions). This smoothness has strictly nothing to do with any cancellation of errors. With a fixed number of PW's, PES's are perfectly smooth, but the (variational) accuracy would depend on the size/shape of the unit cell. Discontinuities come in when varying the boundary conditions because PW basis sets depend on boundary conditions (conceptually) for exactly the same reasons why localized basis sets depend on nuclear positions. With Gaussians you have "Pulay forces" (which require a special treatment going beyond Helmann Feynman), with PW's you have "Pulay stresses" (which require an equally special ad hoc treatment). It is indeed the ease with which accuracy can be assessed and improved with PW's which make the problem appear. As said, if one was ready to use a fixed basis set for different geometries (as one does with localized basis sets without really caring about the dependence of the accuracy on the geometry), the potential energy surface would be perfectly smooth, though less accurate. As simple as that. No misterious powerful cancellation principles here. Stefano On Sep 8, 2005, at 9:17 PM, Konstantin Kudin wrote: > > >> Basically, as you enalarge your unit cell, your brillouin zone >> shrinks, >> and at constant cutoff it means that more lattice vectors (i.e. plane >> waves) enter into the fixed cutoff sphere. More plane waves means a >> systematically larger basis set (this is one of the good things >> of plane waves, it's easy to make the basis set more and more >> complete - >> a nightmare in Gaussian), and, variationally, a lower energy. You see >> in >> fact your energy drop going to the left. >> > > I'd like to chime in here. While indeed reaching the basis set limit > is sort of difficult with Gaussian basis sets, there is this powerful > principle called "cancellation of errors". This allows one to chase > not > the full convergence, but convergence with the basis set that is good > enough. As it turns out, such convergence is actually rather easy to > achieve. > > Then one has to care that the energy is fully variational and smooth > (i.e. differentiable) with respect to changes in atomic coordinates > and > lattice vectors. While indeed PWs do reach the basis set limits > easier, > the smoothness of the potential energy surfaces on the other hand > leaves something to be desired. The smoothness issue is usually much > better taken care of with Gaussians. So there you go ... :-) > > Kostya > > > > > ______________________________________________________ > Click here to donate to the Hurricane Katrina relief effort. > http://store.yahoo.com/redcross-donate3/ > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050908/65ddd152/attachment.htm From diegovmorenor at gmail.com Fri Sep 9 03:10:09 2005 From: diegovmorenor at gmail.com (Diego Moreno) Date: Thu, 8 Sep 2005 20:10:09 -0500 Subject: [Pw_forum] can you help me? Message-ID: <3af87fff05090818108252efe@mail.gmail.com> Dear espresso users: I compiled espresso v2.1.4 with ifc 8.0 and mkl721 and all examples is "done", During my phonons calculations for tungsten bcc crash with message: en bash q=0.00,1.00,0.00...forrtl: severe (24): end-of-file during read, unit -5, file Internal List-Directed Read Image PC Routine Line Source pw.x 08572B88 Unknown Unknown Unknown pw.x 085723EC Unknown Unknown Unknown pw.x 0853BE4D Unknown Unknown Unknown pw.x 08500388 Unknown Unknown Unknown pw.x 0850082B Unknown Unknown Unknown pw.x 08513F68 Unknown Unknown Unknown pw.x 0818AFA3 Unknown Unknown Unknown pw.x 0818AE94 Unknown Unknown Unknown pw.x 080C7AF2 Unknown Unknown Unknown pw.x 080C4CF0 Unknown Unknown Unknown pw.x 0804EE46 Unknown Unknown Unknown pw.x 0804ECB8 Unknown Unknown Unknown libc.so.6 40495974 Unknown Unknown Unknown pw.x 0804EB71 Unknown Unknown Unknown and my input is : &control calculation='phonon' restart_mode='from_scratch', prefix='W', pseudo_dir = '/home/diego/ESPRESSO/Software/Pseudopotenciales/', outdir='/scratch/' / &system ibrav= 0, celldm(1) =5.4691, nat= 2, ntyp= 2, ecutwfc =18.0, occupations='smearing', smearing='methfessel-paxton', degauss=0.05 / &electrons mixing_beta = 0.7, conv_thr = 1.0d-8 / &phonon xqq(1) = 0.00, xqq(2) = 1.00, xqq(3) = 0.00 / ATOMIC_SPECIES W 183.84 W.pz-bhs.UPF CELL_PARAMETERS 1.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 1.0 ATOMIC_POSITIONS W 0.0 0.0 0.0 W 0.5 0.5 0.5 K_POINTS 28 0.0 0.0 0.0 1.0 0.0 0.0 0.1 1.0 0.0 0.0 0.2 1.0 0.0 0.0 0.3 1.0 0.0 0.0 0.4 1.0 0.0 0.0 0.5 1.0 0.0 0.0 0.6 1.0 0.0 0.0 0.7 1.0 0.0 0.0 0.8 1.0 0.0 0.0 0.9 1.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 1.0 0.0 0.1 0.1 1.0 0.0 0.2 0.2 1.0 0.0 0.3 0.3 1.0 0.0 0.4 0.4 1.0 0.0 0.5 0.5 1.0 0.0 0.6 0.6 1.0 0.0 0.7 0.7 1.0 0.0 0.8 0.8 1.0 0.0 0.9 0.9 1.0 0.0 1.0 1.0 1.0 0.0 0.0 0.0 1.0 0.1 0.1 0.1 1.0 0.2 0.2 0.2 1.0 0.3 0.3 0.3 1.0 0.4 0.4 0.4 1.0 0.5 0.5 0.5 1.0 and for ph.x is fonones para W &inputph tr2_ph=1.0d-12, prefix='W', amass(1)=183.84, outdir='/scratch', fildyn='dyn.0.00,1.00,0.00' / thanks for help me Diego Moreno -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050908/e3bcab5f/attachment.htm From konstantin_kudin at yahoo.com Fri Sep 9 04:12:27 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Thu, 8 Sep 2005 19:12:27 -0700 (PDT) Subject: [Pw_forum] version 2.1.5 - strange crashes In-Reply-To: <20050908190511.76802.qmail@web52009.mail.yahoo.com> Message-ID: <20050909021227.30096.qmail@web52008.mail.yahoo.com> I am replying to myself here, anyway, the crashes happen in fpmd.x in 2.1.5 Here is an example: running the calculation with fixed ions...forrtl: severe (174): SIGSEGV, segmentation fault occurred Image PC Routine Line Source fpmd.x 080CB7A7 Unknown Unknown Unknown Stack trace terminated abnormally. Kostya --- Konstantin Kudin wrote: > Hi there, > > I got a report that compiling Espresso 2.1.5 under Linux with either > IFC 8.0 or 9.0 (initial free 32-bit versions, no MPI) gives some > random > crashes at the end of nlset_base in pseudo_base.f line 264. This > happens when running the test suite supplied with the package. Has > anyone else encountered something like that? > > On one hand, initial Intel compiler versions are not known to be > that > reliable, however, it is a bit suspicious if 2 different compiler > versions crash the code in a similar fashion. > > Any ideas/suggestions? > > Kostya > > > > > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From shaposh at isp.nsc.ru Fri Sep 9 06:10:20 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Fri, 9 Sep 2005 11:10:20 +0700 Subject: [Pw_forum] version 2.1.5 - strange crashes In-Reply-To: <20050909021227.30096.qmail@web52008.mail.yahoo.com> Message-ID: <200509090410.j894AFkW031093@ns.isp.nsc.ru> I have already reported this problem some time ago. It happens with 32bit 2.1.5 cpmd.x executable and ifort compiler. No problem with 64bit on Amd64 with ifort. Unfortunately, I do not longer have access to 32 bit Intel machine (dual Xeon) I have only Amd64 systems. As 64bit executables are significantly faster than 32bit on Amd64 (some 20-40% faster) I do not use 32bit systems and don't know if the problem remains for 32bit on Amd64. -----Original Message----- From: pw_forum-admin at pwscf.org [mailto:pw_forum-admin at pwscf.org] On Behalf Of Konstantin Kudin Sent: Friday, September 09, 2005 9:12 AM To: pw_forum at pwscf.org Subject: Re: [Pw_forum] version 2.1.5 - strange crashes I am replying to myself here, anyway, the crashes happen in fpmd.x in 2.1.5 Here is an example: running the calculation with fixed ions...forrtl: severe (174): SIGSEGV, segmentation fault occurred Image PC Routine Line Source fpmd.x 080CB7A7 Unknown Unknown Unknown Stack trace terminated abnormally. Kostya --- Konstantin Kudin wrote: > Hi there, > > I got a report that compiling Espresso 2.1.5 under Linux with either > IFC 8.0 or 9.0 (initial free 32-bit versions, no MPI) gives some > random > crashes at the end of nlset_base in pseudo_base.f line 264. This > happens when running the test suite supplied with the package. Has > anyone else encountered something like that? > > On one hand, initial Intel compiler versions are not known to be > that > reliable, however, it is a bit suspicious if 2 different compiler > versions crash the code in a similar fashion. > > Any ideas/suggestions? > > Kostya > > > > > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com _______________________________________________ Pw_forum mailing list Pw_forum at pwscf.org http://www.democritos.it/mailman/listinfo/pw_forum From jaita at jncasr.ac.in Fri Sep 9 08:53:03 2005 From: jaita at jncasr.ac.in (Jaita Paul) Date: Fri, 9 Sep 2005 12:23:03 +0530 (IST) Subject: [Pw_forum] xcrysden-cannot open pwscf output file In-Reply-To: <200509090410.j894AFkW031093@ns.isp.nsc.ru> References: <20050909021227.30096.qmail@web52008.mail.yahoo.com> <200509090410.j894AFkW031093@ns.isp.nsc.ru> Message-ID: <43396.202.41.111.151.1126248783.squirrel@202.41.111.151> hello everyone. when i try to open a pwscf output file to see the optimized structure i get the following error msgs: error while executing "pwo2xsf.sh" program the error info: /home/jaita/XCrySDen-B1.0bin-static/scripts/pwo2xsf.sh: line 1: bc: command not found /home/jaita/XCrySDen-B1.0bin-static/scripts/pwo2xsf.sh: line 88: test: -eq: unary operator expected can anyone help? best regards, Jaita. From virginie.quequet at polytechnique.fr Fri Sep 9 09:14:57 2005 From: virginie.quequet at polytechnique.fr (Virginie Quequet) Date: 09 Sep 2005 09:14:57 +0200 Subject: [Pw_forum] scan the lattice constant Message-ID: <1126250097.3963.23.camel@gurosai.polytechnique.fr> Dear Stefano Are you sure, that, even for metal, the potential energy surfaces are smooth when we use fixed number of PW's? Because some time ago, I have problem of discontinuities with Ti and I try to impose fixed number of PW's, but the discontinnuities did not disappeared. I have thought that not good enough convergence in k-point was the reason, and that the potential energy surface is perfectly smooth only in the limit of an infinit number of k-point for metal. Is it wrong? -- Virginie Quequet From: Stefano Baroni Subject: Re: [Pw_forum] scan the lattice constant Date: Thu, 8 Sep 2005 21:44:33 +0200 To: pw_forum at pwscf.org Reply-To: pw_forum at pwscf.org --Apple-Mail-1--274898337 Content-Transfer-Encoding: 7bit Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed Kostya: I beg to differ here. I do not want to open a thread "PW's vs. Gaussians" which would have the same chances to be settled as "coca vs. pepsi", but what you say, while perfectly (kind of) clear to me, may be misleading to the less experienced. PW potential energy surfaces (PES) are as smooth as they can be (i.e. *perfectly smooth*, or of class C(\infty), if you prefer) *with fixed boundary conditions* (i.e. for a fixed shape/size of the unit cell which determines the periodic boundary conditions). This smoothness has strictly nothing to do with any cancellation of errors. With a fixed number of PW's, PES's are perfectly smooth, but the (variational) accuracy would depend on the size/shape of the unit cell. Discontinuities come in when varying the boundary conditions because PW basis sets depend on boundary conditions (conceptually) for exactly the same reasons why localized basis sets depend on nuclear positions. With Gaussians you have "Pulay forces" (which require a special treatment going beyond Helmann Feynman), with PW's you have "Pulay stresses" (which require an equally special ad hoc treatment). It is indeed the ease with which accuracy can be assessed and improved with PW's which make the problem appear. As said, if one was ready to use a fixed basis set for different geometries (as one does with localized basis sets without really caring about the dependence of the accuracy on the geometry), the potential energy surface would be perfectly smooth, though less accurate. As simple as that. No misterious powerful cancellation principles here. Stefano From eyvaz_isaev at yahoo.com Fri Sep 9 09:23:00 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 9 Sep 2005 00:23:00 -0700 (PDT) Subject: [Pw_forum] can you help me? In-Reply-To: <3af87fff05090818108252efe@mail.gmail.com> Message-ID: <20050909072300.36502.qmail@web60314.mail.yahoo.com> Hi, Did you start your phonon calculation after scf-step? Bests, Eyvaz. --- Diego Moreno wrote: > Dear espresso users: > I compiled espresso v2.1.4 with ifc 8.0 and mkl721 > and all examples is > "done", During my phonons calculations for tungsten > bcc crash with message: > en bash > q=0.00,1.00,0.00...forrtl: severe (24): end-of-file > during read, unit -5, > file Internal List-Directed Read > Image PC Routine Line Source > pw.x 08572B88 Unknown Unknown Unknown > pw.x 085723EC Unknown Unknown Unknown > pw.x 0853BE4D Unknown Unknown Unknown > pw.x 08500388 Unknown Unknown Unknown > pw.x 0850082B Unknown Unknown Unknown > pw.x 08513F68 Unknown Unknown Unknown > pw.x 0818AFA3 Unknown Unknown Unknown > pw.x 0818AE94 Unknown Unknown Unknown > pw.x 080C7AF2 Unknown Unknown Unknown > pw.x 080C4CF0 Unknown Unknown Unknown > pw.x 0804EE46 Unknown Unknown Unknown > pw.x 0804ECB8 Unknown Unknown Unknown > libc.so.6 40495974 Unknown Unknown Unknown > pw.x 0804EB71 Unknown Unknown Unknown > > and my input is : > > &control > calculation='phonon' > restart_mode='from_scratch', > prefix='W', > pseudo_dir = > '/home/diego/ESPRESSO/Software/Pseudopotenciales/', > outdir='/scratch/' > / > &system > ibrav= 0, celldm(1) =5.4691, nat= 2, ntyp= 2, > ecutwfc =18.0, occupations='smearing', > smearing='methfessel-paxton', > degauss=0.05 > / > &electrons > mixing_beta = 0.7, > conv_thr = 1.0d-8 > / > &phonon > xqq(1) = 0.00, xqq(2) = 1.00, xqq(3) = 0.00 > / > ATOMIC_SPECIES > W 183.84 W.pz-bhs.UPF > CELL_PARAMETERS > 1.0 0.0 0.0 > 0.0 1.0 0.0 > 0.0 0.0 1.0 > ATOMIC_POSITIONS > W 0.0 0.0 0.0 > W 0.5 0.5 0.5 > K_POINTS > 28 > 0.0 0.0 0.0 1.0 > 0.0 0.0 0.1 1.0 > 0.0 0.0 0.2 1.0 > 0.0 0.0 0.3 1.0 > 0.0 0.0 0.4 1.0 > 0.0 0.0 0.5 1.0 > 0.0 0.0 0.6 1.0 > 0.0 0.0 0.7 1.0 > 0.0 0.0 0.8 1.0 > 0.0 0.0 0.9 1.0 > 0.0 0.0 1.0 1.0 > 0.0 0.0 0.0 1.0 > 0.0 0.1 0.1 1.0 > 0.0 0.2 0.2 1.0 > 0.0 0.3 0.3 1.0 > 0.0 0.4 0.4 1.0 > 0.0 0.5 0.5 1.0 > 0.0 0.6 0.6 1.0 > 0.0 0.7 0.7 1.0 > 0.0 0.8 0.8 1.0 > 0.0 0.9 0.9 1.0 > 0.0 1.0 1.0 1.0 > 0.0 0.0 0.0 1.0 > 0.1 0.1 0.1 1.0 > 0.2 0.2 0.2 1.0 > 0.3 0.3 0.3 1.0 > 0.4 0.4 0.4 1.0 > 0.5 0.5 0.5 1.0 > > and for ph.x is > fonones para W > &inputph > tr2_ph=1.0d-12, > prefix='W', > amass(1)=183.84, > outdir='/scratch', > fildyn='dyn.0.00,1.00,0.00' > / > > thanks for help me > > Diego Moreno > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From baroni at sissa.it Fri Sep 9 09:51:34 2005 From: baroni at sissa.it (Stefano Baroni) Date: Fri, 9 Sep 2005 09:51:34 +0200 Subject: [Pw_forum] scan the lattice constant In-Reply-To: <1126250097.3963.23.camel@gurosai.polytechnique.fr> References: <1126250097.3963.23.camel@gurosai.polytechnique.fr> Message-ID: Dear Virginie: the only reason I can see why E(V) may be discontinuous in a metal while keeping the number of PW's fixed is the finite number of calculated energy bands above the Fermi energy. It is perfectly possible that this number varies with volume (if bands are calculated up to some prescribed cutoff energy), or else that the character of some of the calculated bands changes with volume. In either cases, I would expect discontinuities in E(V) if the range of bands above Ef is not large enough with respect to the (Gaussian) smearing. I do not think that these discontinuities have much to do with the number of k points. This can be easily checked by alternatively increasing the number of k points or the number of extra bands above Ef. Stefano On Sep 9, 2005, at 9:14 AM, Virginie Quequet wrote: > Dear Stefano > > Are you sure, that, even for metal, the potential energy surfaces are > smooth when we use fixed number of PW's? Because some time ago, I have > problem of discontinuities with Ti and I try to impose fixed number of > PW's, but the discontinnuities did not disappeared. I have thought > that > not good enough convergence in k-point was the reason, and that the > potential energy surface is perfectly smooth only in the limit of an > infinit number of k-point for metal. Is it wrong? > > > -- > Virginie Quequet > > > From: Stefano Baroni > Subject: Re: [Pw_forum] scan the lattice constant > Date: Thu, 8 Sep 2005 21:44:33 +0200 > To: pw_forum at pwscf.org > Reply-To: pw_forum at pwscf.org > > > --Apple-Mail-1--274898337 > Content-Transfer-Encoding: 7bit > Content-Type: text/plain; > charset=US-ASCII; > delsp=yes; > format=flowed > > Kostya: I beg to differ here. I do not want to open a thread "PW's > vs. Gaussians" which would have the same chances to be settled as > "coca vs. pepsi", but what you say, while perfectly (kind of) clear > to me, may be misleading to the less experienced. > > PW potential energy surfaces (PES) are as smooth as they can be (i.e. > *perfectly smooth*, or of class C(\infty), if you prefer) *with fixed > boundary conditions* (i.e. for a fixed shape/size of the unit cell > which determines the periodic boundary conditions). This smoothness > has strictly nothing to do with any cancellation of errors. With a > fixed number of PW's, PES's are perfectly smooth, but the > (variational) accuracy would depend on the size/shape of the unit > cell. Discontinuities come in when varying the boundary conditions > because PW basis sets depend on boundary conditions (conceptually) > for exactly the same reasons why localized basis sets depend on > nuclear positions. With Gaussians you have "Pulay forces" (which > require a special treatment going beyond Helmann Feynman), with PW's > you have "Pulay stresses" (which require an equally special ad hoc > treatment). > > It is indeed the ease with which accuracy can be assessed and > improved with PW's which make the problem appear. As said, if one was > ready to use a fixed basis set for different geometries (as one does > with localized basis sets without really caring about the dependence > of the accuracy on the geometry), the potential energy surface would > be perfectly smooth, though less accurate. > > As simple as that. No misterious powerful cancellation principles > here. > > Stefano > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050909/6164c341/attachment.htm From g.ballabio at cineca.it Fri Sep 9 09:57:41 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Fri, 9 Sep 2005 09:57:41 +0200 (MEST) Subject: [Pw_forum] xcrysden-cannot open pwscf output file In-Reply-To: <43396.202.41.111.151.1126248783.squirrel@202.41.111.151> (from jaita@jncasr.ac.in on Fri Sep 9 08:53:03 2005) References: <20050909021227.30096.qmail@web52008.mail.yahoo.com> <200509090410.j894AFkW031093@ns.isp.nsc.ru> <43396.202.41.111.151.1126248783.squirrel@202.41.111.151> Message-ID: <1126252660l.4221l.0l@nb-ballabio.cineca.it> On 09/09/2005 08:53:03 AM, Jaita Paul wrote: > /home/jaita/XCrySDen-B1.0bin-static/scripts/pwo2xsf.sh: line 1: bc: > command not found > /home/jaita/XCrySDen-B1.0bin-static/scripts/pwo2xsf.sh: line 88: > test: -eq: unary operator expected The first error means that the "bc" program can't be found on your system. The second error is a consequence. "bc" is a command-line calculator that comes sort of standard on Unix/Linux systems. You may ask your system administrator to install it. If you are yourself the administrator of your machine, if it's a Linux machine, I'm sure packages for bc exist for practically all Linux distributions. Gerardo From yuwen_66 at yahoo.com Fri Sep 9 11:10:16 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Fri, 9 Sep 2005 02:10:16 -0700 (PDT) Subject: [Pw_forum] phonon dos normalization Message-ID: <20050909091016.91699.qmail@web51006.mail.yahoo.com> Dear PW users, After several testing, I found that the phonon density of state from the PWscf calculations are normalized to six times the number of atoms in a unit cell. That is twice of the degrees of freedom. I wonder why don't we just normalize it to the degrees of freedom or just to one? I would appreciate it very much if someone could give me an explanation. W. YU ______________________________________________________ Click here to donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From mousumi at jncasr.ac.in Fri Sep 9 12:42:25 2005 From: mousumi at jncasr.ac.in (Mousumi Upadhyay Kahaly) Date: Fri, 9 Sep 2005 16:12:25 +0530 (IST) Subject: [Pw_forum] phonon dos normalization In-Reply-To: <20050909091016.91699.qmail@web51006.mail.yahoo.com> References: <20050909091016.91699.qmail@web51006.mail.yahoo.com> Message-ID: <45571.202.41.111.151.1126262545.squirrel@202.41.111.151> Hi, In order to cancel first order errors in d/dx form, and to get better results, each atom is once displaced by +delta and then by -delta. Then Dynamical matrices are calculated. Hence matrix size= 2 times the total degrees of freedom, i.e. twice what you expect. regards. > Dear PW users, > > After several testing, I found that the phonon density > of state from the PWscf calculations are normalized to > six times the number of atoms in a unit cell. That is > twice of the degrees of freedom. I wonder why don't we > just normalize it to the degrees of freedom or just to > one? > > I would appreciate it very much if someone could give > me an explanation. > > W. YU > > > > > ______________________________________________________ > Click here to donate to the Hurricane Katrina relief effort. > http://store.yahoo.com/redcross-donate3/ > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From giannozz at nest.sns.it Fri Sep 9 12:47:31 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 9 Sep 2005 12:47:31 +0200 Subject: [Pw_forum] phonon dos normalization In-Reply-To: <20050909091016.91699.qmail@web51006.mail.yahoo.com> References: <20050909091016.91699.qmail@web51006.mail.yahoo.com> Message-ID: <200509091247.31347.giannozz@nest.sns.it> On Friday 09 September 2005 11:10, W. YU wrote: > After several testing, I found that the phonon density of > states from the PWscf calculations are normalized to six > times the number of atoms in a unit cell there is no deep reason. The phonon DOS is calculated by the same routine that calculates the electron DOS. This is normalised to the number of electrons, i.e., in the spin-unpolarised case, to two times the number of states. In the phonon case, this means 2 * (3 * # atoms). You can easily change this to whatever you like. In the next version the normalisation of the phonon dos will be explicitly written. Thank you for bringing this to our attention 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 baroni at sissa.it Fri Sep 9 14:52:48 2005 From: baroni at sissa.it (Stefano Baroni) Date: Fri, 9 Sep 2005 14:52:48 +0200 Subject: [Pw_forum] phonon dos normalization In-Reply-To: <45571.202.41.111.151.1126262545.squirrel@202.41.111.151> References: <20050909091016.91699.qmail@web51006.mail.yahoo.com> <45571.202.41.111.151.1126262545.squirrel@202.41.111.151> Message-ID: On Sep 9, 2005, at 12:42 PM, Mousumi Upadhyay Kahaly wrote: > Hi, > > In order to cancel first order errors in d/dx form, > and to get better results, each atom is once displaced by > +delta and then by -delta. Then Dynamical matrices are > calculated. Hence matrix size= 2 times the total degrees > of freedom, i.e. twice what you expect. ??? May I suggest that only people who i) unrerstand questions and ii) know sensible answers provide the latter? SB > > regards. > > > >> Dear PW users, >> >> After several testing, I found that the phonon density >> of state from the PWscf calculations are normalized to >> six times the number of atoms in a unit cell. That is >> twice of the degrees of freedom. I wonder why don't we >> just normalize it to the degrees of freedom or just to >> one? >> >> I would appreciate it very much if someone could give >> me an explanation. >> >> W. YU >> >> >> >> >> ______________________________________________________ >> Click here to donate to the Hurricane Katrina relief effort. >> http://store.yahoo.com/redcross-donate3/ >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> >> > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > --- Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - Trieste [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) Please, if possible, don't send me MS Word or PowerPoint attachments Why? See: http://www.gnu.org/philosophy/no-word-attachments.html -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050909/6285b192/attachment.htm From mousumi at jncasr.ac.in Fri Sep 9 14:59:14 2005 From: mousumi at jncasr.ac.in (Mousumi Upadhyay Kahaly) Date: Fri, 9 Sep 2005 18:29:14 +0530 (IST) Subject: [Pw_forum] phonon dos normalization In-Reply-To: References: <20050909091016.91699.qmail@web51006.mail.yahoo.com> <45571.202.41.111.151.1126262545.squirrel@202.41.111.151> Message-ID: <54621.202.41.111.151.1126270754.squirrel@202.41.111.151> Hi, I am extremely sorry for my wrong mail. I mistook PWSCF with "Siesta" and in Siesta, following is the cause for getting phonon FC matrix size = 2 times the total degrees of freedom. Am really sorry, mousumi. > On Sep 9, 2005, at 12:42 PM, Mousumi Upadhyay Kahaly wrote: > >> Hi, >> >> In order to cancel first order errors in d/dx form, >> and to get better results, each atom is once displaced by >> +delta and then by -delta. Then Dynamical matrices are >> calculated. Hence matrix size= 2 times the total degrees >> of freedom, i.e. twice what you expect. > > ??? > > May I suggest that only people who i) unrerstand questions and ii) > know sensible answers provide the latter? > > SB > > >> >> regards. >> >> >> >>> Dear PW users, >>> >>> After several testing, I found that the phonon density >>> of state from the PWscf calculations are normalized to >>> six times the number of atoms in a unit cell. That is >>> twice of the degrees of freedom. I wonder why don't we >>> just normalize it to the degrees of freedom or just to >>> one? >>> >>> I would appreciate it very much if someone could give >>> me an explanation. >>> >>> W. YU >>> >>> >>> >>> >>> ______________________________________________________ >>> Click here to donate to the Hurricane Katrina relief effort. >>> http://store.yahoo.com/redcross-donate3/ >>> _______________________________________________ >>> Pw_forum mailing list >>> Pw_forum at pwscf.org >>> http://www.democritos.it/mailman/listinfo/pw_forum >>> >>> >> >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> > > --- > Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - > Trieste > [+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) > > Please, if possible, don't send me MS Word or PowerPoint attachments > Why? See: http://www.gnu.org/philosophy/no-word-attachments.html > > > > From marzari at MIT.EDU Fri Sep 9 15:03:18 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Fri, 09 Sep 2005 09:03:18 -0400 Subject: [Pw_forum] phonon dos normalization In-Reply-To: <54621.202.41.111.151.1126270754.squirrel@202.41.111.151> References: <20050909091016.91699.qmail@web51006.mail.yahoo.com> <45571.202.41.111.151.1126262545.squirrel@202.41.111.151> <54621.202.41.111.151.1126270754.squirrel@202.41.111.151> Message-ID: <43218816.5070907@mit.edu> Hi Mousumi, no problems ! It's very important to know what to have first, the siesta, or the espresso. (Sorry, I couldn't resist). nicola Mousumi Upadhyay Kahaly wrote: > Hi, I am extremely sorry for my wrong mail. > > I mistook PWSCF with "Siesta" and in Siesta, following is the cause for > getting phonon FC matrix size = 2 times the total degrees of freedom. > > Am really sorry, mousumi. > > > >>On Sep 9, 2005, at 12:42 PM, Mousumi Upadhyay Kahaly wrote: >> >> >>>Hi, >>> >>>In order to cancel first order errors in d/dx form, >>>and to get better results, each atom is once displaced by >>>+delta and then by -delta. Then Dynamical matrices are >>>calculated. Hence matrix size= 2 times the total degrees >>>of freedom, i.e. twice what you expect. >> >>??? >> >>May I suggest that only people who i) unrerstand questions and ii) >>know sensible answers provide the latter? >> >>SB >> >> >> >>>regards. >>> >>> >>> >>> >>>>Dear PW users, >>>> >>>>After several testing, I found that the phonon density >>>>of state from the PWscf calculations are normalized to >>>>six times the number of atoms in a unit cell. That is >>>>twice of the degrees of freedom. I wonder why don't we >>>>just normalize it to the degrees of freedom or just to >>>>one? >>>> >>>>I would appreciate it very much if someone could give >>>>me an explanation. >>>> >>>>W. YU >>>> >>>> >>>> >>>> >>>>______________________________________________________ >>>>Click here to donate to the Hurricane Katrina relief effort. >>>>http://store.yahoo.com/redcross-donate3/ >>>>_______________________________________________ >>>>Pw_forum mailing list >>>>Pw_forum at pwscf.org >>>>http://www.democritos.it/mailman/listinfo/pw_forum >>>> >>>> >>> >>>_______________________________________________ >>>Pw_forum mailing list >>>Pw_forum at pwscf.org >>>http://www.democritos.it/mailman/listinfo/pw_forum >>> >> >>--- >>Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - >>Trieste >>[+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) >> >>Please, if possible, don't send me MS Word or PowerPoint attachments >>Why? See: http://www.gnu.org/philosophy/no-word-attachments.html >> >> >> >> > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From eyvaz_isaev at yahoo.com Fri Sep 9 15:14:06 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 9 Sep 2005 06:14:06 -0700 (PDT) Subject: [Pw_forum] phonon dos normalization In-Reply-To: <43218816.5070907@mit.edu> Message-ID: <20050909131406.62959.qmail@web60325.mail.yahoo.com> Hi Nicola, I enjoyed very much your excellent jock!!! Thank you!!1 Bests, Eyvaz. --- Nicola Marzari wrote: > > > Hi Mousumi, > > no problems ! It's very important to know what to > have first, > the siesta, or the espresso. > > (Sorry, I couldn't resist). > > > nicola > > > Mousumi Upadhyay Kahaly wrote: > > > Hi, I am extremely sorry for my wrong mail. > > > > I mistook PWSCF with "Siesta" and in Siesta, > following is the cause for > > getting phonon FC matrix size = 2 times the total > degrees of freedom. > > > > Am really sorry, mousumi. > > > > > > > >>On Sep 9, 2005, at 12:42 PM, Mousumi Upadhyay > Kahaly wrote: > >> > >> > >>>Hi, > >>> > >>>In order to cancel first order errors in d/dx > form, > >>>and to get better results, each atom is once > displaced by > >>>+delta and then by -delta. Then Dynamical > matrices are > >>>calculated. Hence matrix size= 2 times the total > degrees > >>>of freedom, i.e. twice what you expect. > >> > >>??? > >> > >>May I suggest that only people who i) unrerstand > questions and ii) > >>know sensible answers provide the latter? > >> > >>SB > >> > >> > >> > >>>regards. > >>> > >>> > >>> > >>> > >>>>Dear PW users, > >>>> > >>>>After several testing, I found that the phonon > density > >>>>of state from the PWscf calculations are > normalized to > >>>>six times the number of atoms in a unit cell. > That is > >>>>twice of the degrees of freedom. I wonder why > don't we > >>>>just normalize it to the degrees of freedom or > just to > >>>>one? > >>>> > >>>>I would appreciate it very much if someone could > give > >>>>me an explanation. > >>>> > >>>>W. YU > >>>> > >>>> > >>>> > >>>> > >>>>______________________________________________________ > >>>>Click here to donate to the Hurricane Katrina > relief effort. > >>>>http://store.yahoo.com/redcross-donate3/ > >>>>_______________________________________________ > >>>>Pw_forum mailing list > >>>>Pw_forum at pwscf.org > >>>>http://www.democritos.it/mailman/listinfo/pw_forum > >>>> > >>>> > >>> > >>>_______________________________________________ > >>>Pw_forum mailing list > >>>Pw_forum at pwscf.org > >>>http://www.democritos.it/mailman/listinfo/pw_forum > >>> > >> > >>--- > >>Stefano Baroni - SISSA & DEMOCRITOS National > Simulation Center - > >>Trieste > >>[+39] 040 3787 406 (tel) -528 (fax) / > stefanobaroni (skype) > >> > >>Please, if possible, don't send me MS Word or > PowerPoint attachments > >>Why? See: > http://www.gnu.org/philosophy/no-word-attachments.html > >> > >> > >> > >> > > > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > -- > --------------------------------------------------------------------- > Prof Nicola Marzari Department of Materials > Science and Engineering > 13-5066 MIT 77 Massachusetts Avenue Cambridge > MA 02139-4307 USA > tel 617.4522758 fax 617.2586534 marzari at mit.edu > http://nnn.mit.edu > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From mousumi at jncasr.ac.in Fri Sep 9 15:16:13 2005 From: mousumi at jncasr.ac.in (Mousumi Upadhyay Kahaly) Date: Fri, 9 Sep 2005 18:46:13 +0530 (IST) Subject: [Pw_forum] phonon dos normalization In-Reply-To: <43218816.5070907@mit.edu> References: <20050909091016.91699.qmail@web51006.mail.yahoo.com> <45571.202.41.111.151.1126262545.squirrel@202.41.111.151> <54621.202.41.111.151.1126270754.squirrel@202.41.111.151> <43218816.5070907@mit.edu> Message-ID: <59160.202.41.111.151.1126271773.squirrel@202.41.111.151> Dear Sir & all, I am in "Siesta" mailing list too, for calculations of "large no of atoms". And some time back when the mail came, by some bad chance, I mistook the pw forum mail with siesta mail.... and thus created all these problems. I am sorry again. mousumi. > > > Hi Mousumi, > > no problems ! It's very important to know what to have first, > the siesta, or the espresso. > > (Sorry, I couldn't resist). > > > nicola > > > Mousumi Upadhyay Kahaly wrote: > >> Hi, I am extremely sorry for my wrong mail. >> >> I mistook PWSCF with "Siesta" and in Siesta, following is the cause for >> getting phonon FC matrix size = 2 times the total degrees of freedom. >> >> Am really sorry, mousumi. >> >> >> >>>On Sep 9, 2005, at 12:42 PM, Mousumi Upadhyay Kahaly wrote: >>> >>> >>>>Hi, >>>> >>>>In order to cancel first order errors in d/dx form, >>>>and to get better results, each atom is once displaced by >>>>+delta and then by -delta. Then Dynamical matrices are >>>>calculated. Hence matrix size= 2 times the total degrees >>>>of freedom, i.e. twice what you expect. >>> >>>??? >>> >>>May I suggest that only people who i) unrerstand questions and ii) >>>know sensible answers provide the latter? >>> >>>SB >>> >>> >>> >>>>regards. >>>> >>>> >>>> >>>> >>>>>Dear PW users, >>>>> >>>>>After several testing, I found that the phonon density >>>>>of state from the PWscf calculations are normalized to >>>>>six times the number of atoms in a unit cell. That is >>>>>twice of the degrees of freedom. I wonder why don't we >>>>>just normalize it to the degrees of freedom or just to >>>>>one? >>>>> >>>>>I would appreciate it very much if someone could give >>>>>me an explanation. >>>>> >>>>>W. YU >>>>> >>>>> >>>>> >>>>> >>>>>______________________________________________________ >>>>>Click here to donate to the Hurricane Katrina relief effort. >>>>>http://store.yahoo.com/redcross-donate3/ >>>>>_______________________________________________ >>>>>Pw_forum mailing list >>>>>Pw_forum at pwscf.org >>>>>http://www.democritos.it/mailman/listinfo/pw_forum >>>>> >>>>> >>>> >>>>_______________________________________________ >>>>Pw_forum mailing list >>>>Pw_forum at pwscf.org >>>>http://www.democritos.it/mailman/listinfo/pw_forum >>>> >>> >>>--- >>>Stefano Baroni - SISSA & DEMOCRITOS National Simulation Center - >>>Trieste >>>[+39] 040 3787 406 (tel) -528 (fax) / stefanobaroni (skype) >>> >>>Please, if possible, don't send me MS Word or PowerPoint attachments >>>Why? See: http://www.gnu.org/philosophy/no-word-attachments.html >>> >>> >>> >>> >> >> >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum > > -- > --------------------------------------------------------------------- > Prof Nicola Marzari Department of Materials Science and Engineering > 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA > tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From diegovmorenor at gmail.com Fri Sep 9 20:35:49 2005 From: diegovmorenor at gmail.com (Diego Moreno) Date: Fri, 9 Sep 2005 13:35:49 -0500 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #741 - 4 msgs In-Reply-To: <20050909075201.26754.46351.Mailman@democritos.sissa.it> References: <20050909075201.26754.46351.Mailman@democritos.sissa.it> Message-ID: <3af87fff05090911357d8f7e46@mail.gmail.com> Hi Eyvaz, thanks, for you response. and, yes scf and phonon gamma calculation Hi, > > Did you start your phonon calculation after scf-step? > > Bests, > Eyvaz. > > --- Diego Moreno wrote: > > > Dear espresso users: > > I compiled espresso v2.1.4 with ifc 8.0 and mkl721 > > and all examples is > > "done", During my phonons calculations for tungsten > > bcc crash with message: > > en bash > > q=0.00,1.00,0.00...forrtl: severe (24): end-of-file > > during read, unit -5, > > file Internal List-Directed Read > > Image PC Routine Line Source > > pw.x 08572B88 Unknown Unknown Unknown > > pw.x 085723EC Unknown Unknown Unknown > -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050909/3abda00b/attachment.htm From eyvaz_isaev at yahoo.com Fri Sep 9 20:55:47 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 9 Sep 2005 11:55:47 -0700 (PDT) Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #741 - 4 msgs In-Reply-To: <3af87fff05090911357d8f7e46@mail.gmail.com> Message-ID: <20050909185547.3305.qmail@web60322.mail.yahoo.com> Hi again, I looked at more carefully your input file and found an error. You specify ntyp=2, assuming that you have 2 different W atoms. But this is inconsistent with ATOMIC_SPECIES where you show just one atomic type and PsP. So, you should either set up ntyp=1 or specify the second W atom. Bests, Eyvaz. --- Diego Moreno wrote: > Hi Eyvaz, > thanks, for you response. > > and, yes scf and phonon gamma calculation > > > > Hi, > > > > Did you start your phonon calculation after > scf-step? > > > > Bests, > > Eyvaz. > > > > --- Diego Moreno wrote: > > > > > Dear espresso users: > > > I compiled espresso v2.1.4 with ifc 8.0 and > mkl721 > > > and all examples is > > > "done", During my phonons calculations for > tungsten > > > bcc crash with message: > > > en bash > > > q=0.00,1.00,0.00...forrtl: severe (24): > end-of-file > > > during read, unit -5, > > > file Internal List-Directed Read > > > Image PC Routine Line Source > > > pw.x 08572B88 Unknown Unknown Unknown > > > pw.x 085723EC Unknown Unknown Unknown > > > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From naromero at gmail.com Sat Sep 10 02:41:55 2005 From: naromero at gmail.com (Nichols A. Romero) Date: Fri, 9 Sep 2005 20:41:55 -0400 Subject: [Pw_forum] pseudopotential for studying high-pressure elements Message-ID: <6ac064b60509091741413bca77@mail.gmail.com> Hi, I am interesting in doing DFT calculations of some elements at high-pressure, e.g. carbon. Some pseudopotential (PP) generating schemes achieve superior total energy convergence versus planewave cutoff but at the "cost" of increasing r_c. The paper by Troullier and J. L. Martins PRB 43, 1993 (1991) states that if one matches the transferrabiliy of the pseudopotentials (I assuming through the logarithmic derivatives) then the radii used in the TM scheme are double that in the HSC scheme. The result is that the PP generated in the TM scheme is softer than the HSC. My question is: Will the PP give meaning results once the atomic seperation is less than r_c. In general, I would think not. So the next question is that for a fixed r_c what PP will be softess: TM, RRKJ, Ultrasoft-PP, etc.? I would appreciate comments and any relevant citations that I may have overlooked. 1613 Denise Dr. Apt. D Forest Hill, MD 21050 (217) 417-5210 From zfhou at fudan.edu.cn Sat Sep 10 16:32:02 2005 From: zfhou at fudan.edu.cn (Zhufeng Hou) Date: Sat, 10 Sep 2005 22:32:02 +0800 Subject: [Pw_forum] error in Message Passing (mp) module from parallel ph.x Message-ID: <0IML0060HUIYXC@mail.fudan.edu.cn> Dear everyone, I rencently run parellel ph.x (espresso 2.1.5) a computer cluster with Redhat 8.0, which is complied by pgf90 wih mpich-gm. The job exists after getting the Effective charges U-E in cartesian axis and occurs the follwoing error: ---------------- Effective charges U-E in cartesian axis atom 1 ( 1.88305 0.00000 0.00000 ) ( 0.00000 1.88305 0.00000 ) ( 0.00000 0.00000 1.88305 ) atom 2 ( -3.23812 0.00000 0.00000 ) ( 0.00000 -3.23812 0.00000 ) ( 0.00000 0.00000 -3.23812 ) [1] MPI Abort by user Aborting program ! [1] Aborting program! *** error in Message Passing (mp) module *** *** error code: 8911 [0] MPI Abort by user Aborting program ! [0] Aborting program! ----------------- Btw: the input file for ph.x ------------------------- phonons of AlAs at Gamma &inputph tr2_ph=1.0d-12, prefix='alas', epsil=.true., iverbosity=1, amass(1)=26.98, amass(2)=74.92, outdir='/people/zfhou/espresso/tmp/', fildyn='alas.dynG', / 0.0 0.0 0.0 ------------------------ How to fix the problem? Your helps will be greately appreciated. Best wish. Zhufeng From zfhou at fudan.edu.cn Sat Sep 10 17:20:25 2005 From: zfhou at fudan.edu.cn (Zhufeng Hou) Date: Sat, 10 Sep 2005 23:20:25 +0800 Subject: [Pw_forum] MPI Abort by user Aborting program from parellel ph.x Message-ID: <0IML0063KWRKSS@mail.fudan.edu.cn> Dear everyone, I rencently run parellel ph.x (espresso 2.1.5) a computer cluster with Redhat 8.0, which is complied by pgf90 wih mpich-gm. The job exists after getting the Effective charges U-E in cartesian axis and occurs the follwoing error: ---------------- Effective charges U-E in cartesian axis atom 1 ( 1.88305 0.00000 0.00000 ) ( 0.00000 1.88305 0.00000 ) ( 0.00000 0.00000 1.88305 ) atom 2 ( -3.23812 0.00000 0.00000 ) ( 0.00000 -3.23812 0.00000 ) ( 0.00000 0.00000 -3.23812 ) [1] MPI Abort by user Aborting program ! [1] Aborting program! *** error in Message Passing (mp) module *** *** error code: 8911 [0] MPI Abort by user Aborting program ! [0] Aborting program! ----------------- Btw: the input file for ph.x ------------------------- phonons of AlAs at Gamma &inputph tr2_ph=1.0d-12, prefix='alas', epsil=.true., iverbosity=1, amass(1)=26.98, amass(2)=74.92, outdir='/people/zfhou/espresso/tmp/', fildyn='alas.dynG', / 0.0 0.0 0.0 ------------------------ How to fix these problems? Your helps will be greately appreciated. Best wish. Zhufeng From giannozz at nest.sns.it Mon Sep 12 09:49:59 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 12 Sep 2005 09:49:59 +0200 Subject: [Pw_forum] error in Message Passing (mp) module from parallel ph.x In-Reply-To: <0IML0060HUIYXC@mail.fudan.edu.cn> References: <0IML0060HUIYXC@mail.fudan.edu.cn> Message-ID: <200509120949.59601.giannozz@nest.sns.it> On Saturday 10 September 2005 16:32, Zhufeng Hou wrote: > *** error in Message Passing (mp) module *** > *** error code: 8911 look for "8911" in the code and you will find that an array size exceeds the maximum allowed dimension. Print "msglen" and "mp_msgsiz_max" just before the check is performed (line 478 in Modules/mp.f90). If "msglen" has a large but still reasonable value, increase "mp_msgsiz_max". If "msglen" is something unreasonable, try a different compiler. 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 yuwen_66 at yahoo.com Mon Sep 12 15:23:22 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Mon, 12 Sep 2005 06:23:22 -0700 (PDT) Subject: [Pw_forum] scan the lattice constant-seems not a closed question In-Reply-To: <20050911053633.25657.21340.Mailman@democritos.sissa.it> Message-ID: <20050912132322.7694.qmail@web51009.mail.yahoo.com> Dear PW users, I have read the topic about scanning lattice constant and found some explanations about the problem met by XunLei. The answer by Nicola is the most detailed one. But new question arise from here. In the answer, Nicola said: == You see in fact your energy drop going to the left.== But it seems to me the energy drop going to the right. And furthermore, the example figures given by Nicola about the convergence test of Diamond and Graphite a bit puzzled me. As is put by Nicola and basically I believe so, more complet basis set will give lower energy. But in the example figures, especially the Diamond case, the energies obtained at the same k points seem to go higher with the increasing cutoffs. Also in my personal experience of convergence test. There are situations when I fixed a certain cutoff energy and keep increasing the number of k points, the energy some times will go down first and go up later. I couldn't figure out why. I hope I am not misunderstanding something here and if I'm right, could anybody give some further comments. Thanks a lot. W. YU __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From marzari at MIT.EDU Mon Sep 12 17:11:19 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Mon, 12 Sep 2005 11:11:19 -0400 Subject: [Pw_forum] scan the lattice constant-seems not a closed question In-Reply-To: <20050912132322.7694.qmail@web51009.mail.yahoo.com> References: <20050912132322.7694.qmail@web51009.mail.yahoo.com> Message-ID: <43259A97.5020604@mit.edu> > Nicola said: > > == You see in fact your energy drop going to the > left.== > > But it seems to me the energy drop going to the right. Right (on two counts), I actually meant saying "going to the right", i.e. towards larger lattice parameters, i.e. more plane waves in the set. > And furthermore, the example figures given by Nicola > about the convergence test of Diamond and Graphite a > bit puzzled me. As is put by Nicola and basically I > believe so, more complet basis set will give lower > energy. But in the example figures, especially the > Diamond case, the energies obtained at the same k > points seem to go higher with the increasing cutoffs. Wait ! The energy drops very significantly, and the curves become smoother and smoother, as the plane-wave cutoff increases from 24 to 26 to 28 to 30 Ry - i.e. going from black to blue. This is the clear result of enlarging your basis set. The very subtle differences in going from left to right, for a given colour (i.e. given plane wave cutoff) have to do with finer and finer charge denisty grids, i.e. more accurate representations of the ultrasoft augmentation charges. nicola --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From yuwen_66 at yahoo.com Tue Sep 13 03:53:14 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Mon, 12 Sep 2005 18:53:14 -0700 (PDT) Subject: [Pw_forum] scan the lattice constant-seems not a closed question In-Reply-To: <20050911053633.25657.21340.Mailman@democritos.sissa.it> Message-ID: <20050913015314.73704.qmail@web51012.mail.yahoo.com> Thanks to Nicola, I really took the charge density cutoff as ecutoff in the former question. Sorry! But I still have two questions: First, in the diamond case, the lowest energy at an ecutoff seems closer to the left side of the red curve(ecutoff=26Ryd)(which is on the small lattice parameter side) and this seems to be contradictory to the more plane wave basis fact at small lattice constants. (but there is no such a problem for the ecutoff=24Ryd case, why) Second, still comes from the diamond case. As can be seen from the curves at fixed ecutoffs, the lowest energy seems to rise with increasing ecutrho from 160 to 200 Ryd and drops from 200 to 240 Ryd. But the energy at 240 Ryd is still higher than at 160 Ryd. I also met this fact when checking the ecutrho at a fixed ecut in my former calculations. Here the question is: What is the relationship between ecutrho and the total energy (if monotonic lowering of the energy with increasing ecutrho is not always the case) Best regards, W. YU __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From leoant21 at hotmail.com Tue Sep 13 08:18:09 2005 From: leoant21 at hotmail.com (=?gb2312?B?1cUguumx8g==?=) Date: Tue, 13 Sep 2005 06:18:09 +0000 Subject: [Pw_forum] Question on the calculation of MgB2 with distorted lattice. Message-ID: Dear all: I am now trying to estimate the e-p coupling by the deformation potentials, especially the E2g mode. However, I artificially add a change to both the B atoms coordinates following the moving behaviour of the E2g mode, and did a scf calculation with no problem, but when I want to get the band structure by a nscf calculation, there was an error like this: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from Efermi: error #1 unexpected error %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Besides, using ABINIT I can get almost the same band structure with those reported. Please tell me why? Thank you in advance. _________________________________________________________________ ?????????????? MSN Messenger: http://messenger.msn.com/cn From giannozz at nest.sns.it Tue Sep 13 09:48:34 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Tue, 13 Sep 2005 09:48:34 +0200 Subject: [Pw_forum] Question on the calculation of MgB2 with distorted lattice. In-Reply-To: References: Message-ID: <200509130948.34188.giannozz@nest.sns.it> On Tuesday 13 September 2005 08:18, ? ?? wrote: > from Efermi: error #1 > unexpected error please provide a test job > Besides, using ABINIT I can get almost the same band structure > with those reported. Please tell me why? why should they be different ??? -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From stewart at cnf.cornell.edu Tue Sep 13 17:23:31 2005 From: stewart at cnf.cornell.edu (stewart at cnf.cornell.edu) Date: 13 Sep 2005 11:23:31 -0400 Subject: [Pw_forum] Re: Question on the calculation of MgB2 with distorted lattice. In-Reply-To: References: Message-ID: <20050913112331.17833.qmail@xuxa.iecc.com> I have run into that error in the past when doing nscf calculations. You need to make sure the system section of your nscf calculation does not include the smearing and degauss keywords. Once these are removed, the nscf calculation is able to run without problems. Best regards, Derek ############################## Derek Stewart, Ph. D. Scientific Computation Associate 250 Duffield Hall Cornell Nanoscale Facility (CNF) Ithaca, NY 14853 stewart (at) cnf.cornell.edu [gb2312] ?? ??? writes: > Dear all: > I am now trying to estimate the e-p coupling by the deformation > potentials, especially the E2g mode. However, I artificially add a change > to both the B atoms coordinates following the moving behaviour of the E2g > mode, and did a scf calculation with no problem, but when I want to get > the band structure by a nscf calculation, there was an error like this: > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%%%%%%%% > > from Efermi: error #1 > unexpected error > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > %%%%%%%%%%%%%% > > Besides, using ABINIT I can get almost the same band structure with those > reported. Please tell me why? > Thank you in advance. > > _________________________________________________________________ > ????????????????????? MSN Messenger: http://messenger.msn.com/cn > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum From marzari at MIT.EDU Tue Sep 13 17:28:16 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Tue, 13 Sep 2005 11:28:16 -0400 Subject: [Pw_forum] scan the lattice constant-seems not a closed question In-Reply-To: <20050913015314.73704.qmail@web51012.mail.yahoo.com> References: <20050913015314.73704.qmail@web51012.mail.yahoo.com> Message-ID: <4326F010.604@mit.edu> > But I still have two questions: First, in the diamond > case, the lowest energy at an ecutoff seems closer to > the left side of the red curve(ecutoff=26Ryd)(which is > on the small lattice parameter side) and this seems to > be contradictory to the more plane wave basis fact at > small lattice constants. (but there is no such a > problem for the ecutoff=24Ryd case, why) The exact E(V) curve will be a smooth curve closely resembling the mathematical form of a Murnaghan or Birch equation of state. If you have a low cutoff, there will be discontinuities every time an infinitesimal increase in your lattice constant introduces one more plane wave in your basis set, lowering by a discrete amount the total energy. In order to answer your questions, you might want to try an E(V) curve for very many lattice parameters (e.g. 100), at a low cutoff, and try to find out the discontinuities mentioned. > What is the relationship between ecutrho > and the total energy (if monotonic lowering of the > energy with increasing ecutrho is not always the case) None relevant to your case - mostly it is related to describing properly the augmentation charges of the Vanderbilt ultrasoft pseudopotentials. This can be very important to predict correctly magnetic multiplets, or empty states (for what they are worth). So it's an important parameter to check, but we can't say if it will lower or raise the energy. Last - different codes treat differently the augmentation charges. PWSCF represents them on the fine real-space mesh for which ecutrho is provided. CP represents them with a localized real-space box-grid (technically called a "scatoletta"), of a resolution given by ecutrho, but identical to zero outside this box-grid. While infinite ecutrho will always work, I believe that the CP-way is a bit faster in reaching convergence (and, on top, very very efficient due to the small size of the box-grid - this is all due to Alfredo Pasquarello's work, discussed in detail in the 1993 PRB). Best, nicola --------------------------------------------------------------------- Prof Nicola Marzari Department of Materials Science and Engineering 13-5066 MIT 77 Massachusetts Avenue Cambridge MA 02139-4307 USA tel 617.4522758 fax 617.2586534 marzari at mit.edu http://nnn.mit.edu From jasonsun98 at hotmail.com Wed Sep 14 05:30:55 2005 From: jasonsun98 at hotmail.com (sun jason) Date: Wed, 14 Sep 2005 03:30:55 +0000 Subject: [Pw_forum] about phonon calculation In-Reply-To: <20050913053631.4204.63946.Mailman@democritos.sissa.it> Message-ID: Dear all, I encounter a problem when I calculate the phonon frequency of an orthorhombic lattice, I use 4*4*4 to test, the scf,nscf and phonon calculations are ok but when q2r.x begin to run, it crashed, and said "q not allow..." what's wrong? any suggestions would be appreciated. the lattice of my system is like this: -------------------------------- &SYSTEM ibrav = 8, A = 2.5383, B = 2.5126 , C = 3.6069 , cosAB = 0 , cosAC = 0 , cosBC = 0 , / K_POINTS AUTOMATIC 4 4 4 0 0 0 ------------------------------ I used k-points of a (444) uniform grid of the primertive orthorhombic lattice to test qpoints="0.2500000,0.0000000,0.0000000\ 0.5000000,0.0000000,0.0000000\ 0.0000000,0.2525508,0.0000000\ 0.2500000,0.2525508,0.0000000\ 0.5000000,0.2525508,0.0000000\ 0.0000000,0.5051015,0.0000000\ 0.2500000,0.5051015,0.0000000\ 0.5000000,0.5051015,0.0000000\ 0.0000000,0.0000000,0.1759324\ 0.2500000,0.0000000,0.1759324\ 0.5000000,0.0000000,0.1759324\ 0.0000000,0.2525508,0.1759324\ 0.2500000,0.2525508,0.1759324\ 0.5000000,0.2525508,0.1759324\ 0.0000000,0.5051015,0.1759324\ 0.2500000,0.5051015,0.1759324\ 0.5000000,0.5051015,0.1759324\ 0.0000000,0.0000000,0.3518649\ 0.2500000,0.0000000,0.3518649\ 0.5000000,0.0000000,0.3518649\ 0.0000000,0.2525508,0.3518649\ 0.2500000,0.2525508,0.3518649\ 0.5000000,0.2525508,0.3518649\ 0.0000000,0.5051015,0.3518649\ 0.2500000,0.5051015,0.3518649\ 0.5000000,0.5051015,0.3518649" cat > q2r.in < C(R)...\c" $Q2R_COMMAND < q2r.in > q2r.out $ECHO " done" and the output file of q2r.x is like this: ---------- reading dyn.mat. from file test.dynG macroscopic fields = T 6.77491 0.00000 0.00000 0.00000 5.88300 0.00000 0.00000 0.00000 5.92033 na= 1 1.05118 0.00000 0.00000 0.00000 1.32053 0.00000 0.00000 0.00000 1.56316 na= 2 1.76349 0.00000 0.00000 0.00000 0.84508 0.00000 0.00000 0.00000 1.18992 na= 3 -1.88952 0.00000 0.00000 0.00000 -1.55326 0.00000 0.00000 0.00000 -1.58880 na= 4 -1.63859 0.00000 0.00000 0.00000 -1.20512 0.00000 0.00000 0.00000 -1.21948 nqs= 1 q= 0.00000000 0.00000000 0.00000000 reading dyn.mat. from file dyn.0.2500000,0.0000000,0.0000000 nqs= 2 q= 0.25000000 0.00000000 0.00000000 q= -0.25000000 0.00000000 0.00000000 reading dyn.mat. from file dyn.0.5000000,0.0000000,0.0000000 nqs= 1 q= 0.50000000 0.00000000 0.00000000 reading dyn.mat. from file dyn.0.0000000,0.2525508,0.0000000 nqs= 2 q= 0.00000000 0.25255080 0.00000000 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from init : error # 1 q not allowed %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... ------------------------------------------- best regards, ============================================== Jian Sun Physics Dept. of Nanjing University National Lab. of Solid State Microstructures 22 Hankou Road, Gulou District Nanjing, Jiangsu Province 210093 China ============================================== From giannozz at nest.sns.it Wed Sep 14 10:15:21 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 14 Sep 2005 10:15:21 +0200 Subject: [Pw_forum] about phonon calculation In-Reply-To: References: Message-ID: <200509141015.21705.giannozz@nest.sns.it> On Wednesday 14 September 2005 05:30, sun jason wrote: > when q2r.x begin to run, it crashed, and said "q not allow..." > > what's wrong? your grid of q points should be something like: (2pi/a)*(1/4)*( I, J*a/b, K*a/c ) with I, J,K integers running from 0 to 3. Is it? I don't think so P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From eyvaz_isaev at yahoo.com Wed Sep 14 12:02:49 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Wed, 14 Sep 2005 03:02:49 -0700 (PDT) Subject: [Pw_forum] about phonon calculation In-Reply-To: Message-ID: <20050914100250.13876.qmail@web60314.mail.yahoo.com> Hi Jason, Please be patient and do not spam. Your problem, more likely is connected to q-points you used. More precisely, based on information you provided I could not reproduce the same q-points. As example, your particular q-point for which you have got the error is q= 0.00000000 0.25255080 0.00000000 But I have q= 0.0000000 0.2525571 0.0000000 using both pw.x and kpoints.x. There is a difference for other q-points, too. I do not think that the difference is due to the use of different compilers/CPU/OS. So, please check your q-points. By the way, you can reduce eps in q2r.f90 down to 1.0d-6, recompile tools and look what happens. Bests, Eyvaz. --- sun jason < --- sun jason wrote: > Dear all, > > I encounter a problem when I calculate the phonon > frequency of an > orthorhombic lattice, I use 4*4*4 to test, the > scf,nscf and phonon > calculations are ok > but when q2r.x begin to run, it crashed, and said "q > not allow..." > > what's wrong? any suggestions would be appreciated. > > the lattice of my system is like this: > > -------------------------------- > &SYSTEM > ibrav = 8, > A = 2.5383, > B = 2.5126 , > C = 3.6069 , > cosAB = 0 , > cosAC = 0 , > cosBC = 0 , > > / > K_POINTS AUTOMATIC > 4 4 4 0 0 0 > ------------------------------ > > I used k-points of a (444) uniform grid of the > primertive orthorhombic > lattice to test > > qpoints="0.2500000,0.0000000,0.0000000\ > 0.5000000,0.0000000,0.0000000\ > 0.0000000,0.2525508,0.0000000\ > 0.2500000,0.2525508,0.0000000\ > 0.5000000,0.2525508,0.0000000\ > 0.0000000,0.5051015,0.0000000\ > 0.2500000,0.5051015,0.0000000\ > 0.5000000,0.5051015,0.0000000\ > 0.0000000,0.0000000,0.1759324\ > 0.2500000,0.0000000,0.1759324\ > 0.5000000,0.0000000,0.1759324\ > 0.0000000,0.2525508,0.1759324\ > 0.2500000,0.2525508,0.1759324\ > 0.5000000,0.2525508,0.1759324\ > 0.0000000,0.5051015,0.1759324\ > 0.2500000,0.5051015,0.1759324\ > 0.5000000,0.5051015,0.1759324\ > 0.0000000,0.0000000,0.3518649\ > 0.2500000,0.0000000,0.3518649\ > 0.5000000,0.0000000,0.3518649\ > 0.0000000,0.2525508,0.3518649\ > 0.2500000,0.2525508,0.3518649\ > 0.5000000,0.2525508,0.3518649\ > 0.0000000,0.5051015,0.3518649\ > 0.2500000,0.5051015,0.3518649\ > 0.5000000,0.5051015,0.3518649" > > > cat > q2r.in < &input > nr1=4, nr2=4, nr3=4, zasr=.true., > fild='test444.fc' > / > 27 > test.dynG > dyn.0.2500000,0.0000000,0.0000000 > dyn.0.5000000,0.0000000,0.0000000 > dyn.0.0000000,0.2525508,0.0000000 > dyn.0.2500000,0.2525508,0.0000000 > dyn.0.5000000,0.2525508,0.0000000 > dyn.0.0000000,0.5051015,0.0000000 > dyn.0.2500000,0.5051015,0.0000000 > dyn.0.5000000,0.5051015,0.0000000 > dyn.0.0000000,0.0000000,0.1759324 > dyn.0.2500000,0.0000000,0.1759324 > dyn.0.5000000,0.0000000,0.1759324 > dyn.0.0000000,0.2525508,0.1759324 > dyn.0.2500000,0.2525508,0.1759324 > dyn.0.5000000,0.2525508,0.1759324 > dyn.0.0000000,0.5051015,0.1759324 > dyn.0.2500000,0.5051015,0.1759324 > dyn.0.5000000,0.5051015,0.1759324 > dyn.0.0000000,0.0000000,0.3518649 > dyn.0.2500000,0.0000000,0.3518649 > dyn.0.5000000,0.0000000,0.3518649 > dyn.0.0000000,0.2525508,0.3518649 > dyn.0.2500000,0.2525508,0.3518649 > dyn.0.5000000,0.2525508,0.3518649 > dyn.0.0000000,0.5051015,0.3518649 > dyn.0.2500000,0.5051015,0.3518649 > dyn.0.5000000,0.5051015,0.3518649 > EOF > > $ECHO " transforming C(q) => C(R)...\c" > $Q2R_COMMAND < q2r.in > q2r.out > $ECHO " done" > > and the output file of q2r.x is like this: > ---------- > reading dyn.mat. from file > test.dynG > > > > macroscopic fields = T > 6.77491 0.00000 0.00000 > 0.00000 5.88300 0.00000 > 0.00000 0.00000 5.92033 > na= 1 > 1.05118 0.00000 0.00000 > 0.00000 1.32053 0.00000 > 0.00000 0.00000 1.56316 > na= 2 > 1.76349 0.00000 0.00000 > 0.00000 0.84508 0.00000 > 0.00000 0.00000 1.18992 > na= 3 > -1.88952 0.00000 0.00000 > 0.00000 -1.55326 0.00000 > 0.00000 0.00000 -1.58880 > na= 4 > -1.63859 0.00000 0.00000 > 0.00000 -1.20512 0.00000 > 0.00000 0.00000 -1.21948 > nqs= 1 > q= 0.00000000 0.00000000 0.00000000 > reading dyn.mat. from file > dyn.0.2500000,0.0000000,0.0000000 > > > > nqs= 2 > q= 0.25000000 0.00000000 0.00000000 > q= -0.25000000 0.00000000 0.00000000 > reading dyn.mat. from file > dyn.0.5000000,0.0000000,0.0000000 > > > > nqs= 1 > q= 0.50000000 0.00000000 0.00000000 > reading dyn.mat. from file > dyn.0.0000000,0.2525508,0.0000000 > > > > nqs= 2 > q= 0.00000000 0.25255080 0.00000000 > > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > from init : error # 1 > q not allowed > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > > stopping ... > ------------------------------------------- > > > best regards, > ============================================== > Jian Sun > Physics Dept. of Nanjing University > National Lab. of Solid State Microstructures > 22 Hankou Road, Gulou District > Nanjing, Jiangsu Province > 210093 > China > ============================================== > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From yi.kong at gmail.com Wed Sep 14 15:46:07 2005 From: yi.kong at gmail.com (Yi Kong) Date: Wed, 14 Sep 2005 21:46:07 +0800 Subject: [Pw_forum] problem in phonon DOS calculation of HCP structure Message-ID: <8c9dc66c050914064620db0e7c@mail.gmail.com> Dear all, I encounter a problem when I do phonon DOS calculation of HCP structure. In my calculation, I use 4*4*4 grid, that is to say, 12 special points to perform PH.x calculation, than I use Q2r.x to get the force constants in real space. These calculations are all ended correctly. When I want to use matdyn.x to get the phonon DOS, an error message appeared: ********************** error message **************** 63.3333333333333333333 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from frc_blk : error # wrong total_weight %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping ... *************************** end **************** Who can help me to solve this problem? Or any tips about doing phonon calculation of HCP structure are welcome. Thanks a lot! Sincerely, Yi Kong -- ============================= Yi Kong Department of MSE Tsinghua University PR China, 100084 ============================= -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050914/3ae40042/attachment.htm From giannozz at nest.sns.it Wed Sep 14 15:59:21 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 14 Sep 2005 15:59:21 +0200 Subject: [Pw_forum] problem in phonon DOS calculation of HCP structure In-Reply-To: <8c9dc66c050914064620db0e7c@mail.gmail.com> References: <8c9dc66c050914064620db0e7c@mail.gmail.com> Message-ID: <200509141559.21267.giannozz@nest.sns.it> On Wednesday 14 September 2005 15:46, Yi Kong wrote: > I encounter a problem when I do phonon DOS calculation of HCP structure. > In my calculation, I use 4*4*4 grid, that is to say, 12 special points to > perform PH.x calculation, than I use Q2r.x to get the force constants in > real space. These calculations are all ended correctly. When I want to use > matdyn.x to get the phonon DOS, an error message appeared [...] > > Who can help me to solve this problem? please provide a job showing this error -- 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 yi.kong at gmail.com Wed Sep 14 16:25:07 2005 From: yi.kong at gmail.com (Yi Kong) Date: Wed, 14 Sep 2005 22:25:07 +0800 Subject: [Pw_forum] problem in phonon DOS calculation of HCP structure (Job details) Message-ID: <8c9dc66c050914072517237d20@mail.gmail.com> Dear all, The following are the input files in my calculations. Thanks a lot! Sincerely, Yi Kong **************** to get the dynamical matrix ************* # how to run execubles PW_COMMAND="mpirun -np 8 $BIN_DIR/pw.x -npool 4" PH_COMMAND="mpirun -np 8 $BIN_DIR/ph.x -npool 4" $ECHO $ECHO " running pw.x as: $PW_COMMAND" $ECHO " running ph.x as: $PH_COMMAND" $ECHO # clean TMP_DIR $ECHO " cleaning $TMP_DIR...\c" rm -rf $TMP_DIR/* $ECHO " done" # self-consistent calculation cat > PdRu3.scf.in << EOF &control calculation='scf' restart_mode='from_scratch', tprnfor= .true. prefix='PdRu3', pseudo_dir = '$PSEUDO_DIR/', outdir='$TMP_DIR/' / &system ibrav= 4, celldm(1) = 10.44, celldm(3)= 0.793 , nat= 8, ntyp= 2, ecutwfc =30.0, ecutrho = 300.0, occupations='smearing',degauss=0.02, smearing='marzari-vanderbilt' / &electrons conv_thr = 1.0d-8 mixing_beta = 0.3 mixing_mode = 'plain' diagonalization='david' mixing_ndim=8 / ATOMIC_SPECIES Ru 101.07 Ru.pbe-n-van.UPF Pd 106.4 Pd.pbe-nd-rrkjus.UPF ATOMIC_POSITIONS (crystal) Ru 0.50 0.0 0.00 0 0 0 Ru 0.50 0.50 0.00 0 0 0 Ru 0.00 0.50 0.00 0 0 0 Ru 0.333333 0.1666667 0.50 0 0 0 Ru 0.833333 0.1666667 0.50 0 0 0 Ru 0.833333 0.6666667 0.50 0 0 0 Pd 0.00 0.00 0.00 0 0 0 Pd 0.333333 0.6666667 0.50 0 0 0 K_POINTS AUTOMATIC 8 8 8 0 0 0 EOF $ECHO " running the scf calculation...\c" $PW_COMMAND < PdRu3.scf.in > PdRu3.scf.out $ECHO " done" # phonons cat > PdRu3.phG.in << EOF phonons of PdRu3 at Gamma &inputph tr2_ph=1.0d-12, prefix='PdRu3', amass(2)=106.4, amass(1)=101.07, outdir='$TMP_DIR/', fildyn='PdRu3.dynG', /0.0 0.0 0.0 EOF $ECHO " running the phonon calculation at Gamma...\c" $PH_COMMAND < PdRu3.phG.in > PdRu3.phG.out $ECHO " done" # These are k-points of a (444) uniform grid in the irreducible # Brillouin Zone of the bcc lattice (Gamma is treated separately) /bin/rm PdRu3.nscf.out PdRu3.ph.out qpoints="0.0000000,0.0000000,0.3152585\ 0.0000000,0.0000000,-0.6305170\ 0.0000000,0.2886751,0.0000000\ 0.0000000,0.2886751,0.3152585\ 0.0000000,0.2886751,-0.6305170\ 0.0000000,-0.5773503,0.0000000\ 0.0000000,-0.5773503,0.3152585\ 0.0000000,-0.5773503,-0.6305170\ 0.2500000,0.4330127,0.0000000\ 0.2500000,0.4330127,0.3152585\ 0.2500000,0.4330127,-0.6305170" for qpoint in $qpoints ; do # non self-consistent calculation qx=`$ECHO $qpoint | cut -d, -f1` qy=`$ECHO $qpoint | cut -d, -f2` qz=`$ECHO $qpoint | cut -d, -f3` cat > PdRu3.nscf.in << EOF &control calculation='phonon' restart_mode='from_scratch', prefix='PdRu3', pseudo_dir = '$PSEUDO_DIR/', outdir='$TMP_DIR/' / &system ibrav= 4, celldm(1) = 10.44, celldm(3)= 0.793 , nat= 8, ntyp= 2, ecutwfc =30.0, ecutrho = 300.0, occupations='smearing',degauss=0.02, smearing='marzari-vanderbilt' / &electrons conv_thr = 1.0d-8 mixing_beta = 0.3 mixing_mode = 'plain' diagonalization='david' mixing_ndim=8 / &phonon xqq(1) = $qx, xqq(2) = $qy, xqq(3) = $qz / ATOMIC_SPECIES Ru 101.07 Ru.pbe-n-van.UPF Pd 106.4 Pd.pbe-nd-rrkjus.UPF ATOMIC_POSITIONS (crystal) Ru 0.50 0.0 0.00 0 0 0 Ru 0.50 0.50 0.00 0 0 0 Ru 0.00 0.50 0.00 0 0 0 Ru 0.333333 0.1666667 0.50 0 0 0 Ru 0.833333 0.1666667 0.50 0 0 0 Ru 0.833333 0.6666667 0.50 0 0 0 Pd 0.00 0.00 0.00 0 0 0 Pd 0.333333 0.6666667 0.50 0 0 0 K_POINTS AUTOMATIC 8 8 8 0 0 0 EOF $ECHO " running the nscf calculation at q=$qpoint...\c" $PW_COMMAND < PdRu3.nscf.in >> PdRu3.nscf.out $ECHO " done" # phonon calculation cat > PdRu3.ph.in << EOF phonons of PdRu3 at $qpoint &inputph tr2_ph=1.0d-12, prefix='PdRu3', amass(2)=106.4, amass(1)=101.07, outdir='$TMP_DIR/', fildyn='dyn.$qpoint' alpha_mix(1)=0.3 / $qpoint EOF $ECHO " running the phonon calculation at q=$qpoint...\c" $PH_COMMAND < PdRu3.ph.in >> PdRu3.ph.out $ECHO " done" done $ECHO $ECHO "$CURRENT_DIR: done" ******************** to get the force constants ******** # how to run execubles Q2R_COMMAND="mpirun -np 8 $BIN_DIR/q2r.x " $ECHO $ECHO " running q2r.x as: $Q2R_COMMAND" $ECHO cat > q2r.in << EOF &input nr1=4, nr2=4, nr3=4, fild='pd3ru444.fc' / 12 pd3ru.dynGdyn.0.0000000,0.0000000,0.3152585dyn.0.0000000,0.0000000,-0.6305170dyn.0.0000000,0.2886751,0.0000000dyn.0.0000000,0.2886751,0.3152585dyn.0.0000000,0.2886751,-0.6305170dyn.0.0000000,-0.5773503,0.0000000dyn.0.0000000,-0.5773503,0.3152585dyn.0.0000000,-0.5773503,-0.6305170dyn.0.2500000,0.4330127,0.0000000dyn.0.2500000,0.4330127,0.3152585dyn.0.2500000,0.4330127,-0.6305170 EOF $ECHO "transforming C(q)=>C(R)...\c" $Q2R_COMMAND < q2r.in > q2r.out $ECHO "done" $ECHO $ECHO "$CURRENT_DIR: done" ***************** to get the phonon DOS ********* # how to run execubles MATDYN_COMMAND="$PARA_PREFIX $BIN_DIR/matdyn.x $PARA_POSTFIX" $ECHO $ECHO " running matdyn.x as: $MATDYN_COMMAND" $ECHO cat >phdos.in < > phdos.out $ECHO " done " $ECHO $ECHO "$CURRENT_DIR: done" -- ============================= Yi Kong Department of MSE Tsinghua University PR China, 100084 ============================= -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050914/a4fe793b/attachment.htm From giannozz at nest.sns.it Wed Sep 14 16:55:50 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 14 Sep 2005 16:55:50 +0200 Subject: [Pw_forum] problem in phonon DOS calculation of HCP structure (Job details) In-Reply-To: <8c9dc66c050914072517237d20@mail.gmail.com> References: <8c9dc66c050914072517237d20@mail.gmail.com> Message-ID: <200509141655.50636.giannozz@nest.sns.it> All programs in the pwtool/ directory should be run on just one processor. I am not sure that this is relevant to your problem, 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 asen at iitk.ac.in Wed Sep 14 17:25:31 2005 From: asen at iitk.ac.in (A Sen) Date: Wed, 14 Sep 2005 20:55:31 +0530 (IST) Subject: [Pw_forum] Compilation error Message-ID: <46602.172.26.113.66.1126711531.squirrel@webmail.iitk.ac.in> Dear all, While compiling espresso-2.1.5, I encounter a problem of the following form: --------------- IPO link: can not find "(" ifort: error: problem during multi-file optimization compilation (code 1) make[1]: *** [memory.x] Error 1 make[1]: Leaving directory `/home/arimita/soft/espresso-2.1.5/PW' make: *** [pw] Error 2 --------------- Any suggestions would be appreciated. Sincerely, Arijit From giannozz at nest.sns.it Wed Sep 14 18:28:27 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Wed, 14 Sep 2005 18:28:27 +0200 Subject: [Pw_forum] Compilation error In-Reply-To: <46602.172.26.113.66.1126711531.squirrel@webmail.iitk.ac.in> References: <46602.172.26.113.66.1126711531.squirrel@webmail.iitk.ac.in> Message-ID: <200509141828.27081.giannozz@nest.sns.it> On Wednesday 14 September 2005 17:25, A Sen wrote: > While compiling espresso-2.1.5, I encounter a problem of the > following form: > --------------- > IPO link: can not find "(" > ifort: error: problem during multi-file optimization compilation (code 1) > make[1]: *** [memory.x] Error 1 > make[1]: Leaving directory `/home/arimita/soft/espresso-2.1.5/PW' > make: *** [pw] Error 2 > --------------- > Any suggestions would be appreciated. lower the optimization level, or upgrade to a less buggy version of the compiler, or both P. -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From yi.kong at gmail.com Thu Sep 15 02:41:15 2005 From: yi.kong at gmail.com (Yi Kong) Date: Thu, 15 Sep 2005 08:41:15 +0800 Subject: [Pw_forum] problem in phonon DOS calculation of HCP structure (Job details) Message-ID: <8c9dc66c050914174123eeab45@mail.gmail.com> Dear Paolo, The matdyn.x is for sure running on a single processor. The Ru-Pd alloy system I studied is an immiscible system, that is to say, the alloy has positive heat of formation and is artificially created nonequilibrium alloy. I don't know if this feature can lead to this error, but I indeed obtained the phonon spectra of this alloy system in cubic structures. Sincerely, Yi Kong -- ============================= Yi Kong Department of MSE Tsinghua University PR China, 100084 ============================= -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050915/ea94bc8c/attachment.htm From shaposh at isp.nsc.ru Thu Sep 15 09:26:20 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Thu, 15 Sep 2005 14:26:20 +0700 Subject: [Pw_forum] Compilation error In-Reply-To: <200509141828.27081.giannozz@nest.sns.it> References: <46602.172.26.113.66.1126711531.squirrel@webmail.iitk.ac.in> <200509141828.27081.giannozz@nest.sns.it> Message-ID: <1126769180.6316.2.camel@m00> With Intel icc and ifort 9.0 compiler use -no-ipo flag Just compile mpich (or mpich2) with "icc -no-ipo" and "ifort -no-ipo" as c and fortran compilers. On Wed, 2005-09-14 at 18:28 +0200, Paolo Giannozzi wrote: > On Wednesday 14 September 2005 17:25, A Sen wrote: > > > While compiling espresso-2.1.5, I encounter a problem of the > > following form: > > --------------- > > IPO link: can not find "(" > > ifort: error: problem during multi-file optimization compilation (code 1) > > make[1]: *** [memory.x] Error 1 > > make[1]: Leaving directory `/home/arimita/soft/espresso-2.1.5/PW' > > make: *** [pw] Error 2 > > --------------- > > Any suggestions would be appreciated. > > lower the optimization level, or upgrade to a less buggy version > of the compiler, or both > > P. From giannozz at nest.sns.it Thu Sep 15 09:38:12 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Thu, 15 Sep 2005 09:38:12 +0200 Subject: [Pw_forum] problem in phonon DOS calculation of HCP structure (Job details) In-Reply-To: <8c9dc66c050914174123eeab45@mail.gmail.com> References: <8c9dc66c050914174123eeab45@mail.gmail.com> Message-ID: <200509150938.12806.giannozz@nest.sns.it> On Thursday 15 September 2005 02:41, Yi Kong wrote: > The matdyn.x is for sure running on a single processor. your script runs q2r.x and matdyn.x on multiple processors > The Ru-Pd alloy system I studied is an immiscible system, that is > to say, the alloy has positive heat of formation and is artificially > created nonequilibrium alloy. I don't know if this feature can lead > to this error no -- 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 chaohao2002 at 163.com Fri Sep 16 07:01:20 2005 From: chaohao2002 at 163.com (=?gb2312?B?Q2hhb2hhbyBIdQ==?=) Date: Fri, 16 Sep 2005 13:01:20 +0800 (CST) Subject: [Pw_forum] =?gb2312?B?aG93IHRvIHNvbHZlIHRoaXMgcHJvYmxlbQ==?= Message-ID: <432A51A0.000017.02388@bj163app67.163.com> Dear users, I want to know whether stress considering tetragonal distion can be put on the cell in pwscf program. if it can, how do it? Who knows? please help me. Thanks in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050916/a140e17d/attachment.htm From liyanpcl at yahoo.com.cn Fri Sep 16 09:56:45 2005 From: liyanpcl at yahoo.com.cn (li yan) Date: Fri, 16 Sep 2005 15:56:45 +0800 (CST) Subject: [Pw_forum] where can i get the input file about raman? Message-ID: <20050916075645.72177.qmail@web15602.mail.cnb.yahoo.com> i use the version 2.1.3 of pwscf and want to calculate the raman intensity, but i can't understand the input data in the exampe15 about raman. would you be so kind to help me? --------------------------------- DO YOU YAHOO!? ????G??????????????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050916/5e1465cb/attachment.htm From liyanpcl at yahoo.com.cn Fri Sep 16 10:29:45 2005 From: liyanpcl at yahoo.com.cn (li yan) Date: Fri, 16 Sep 2005 16:29:45 +0800 (CST) Subject: [Pw_forum] Could you please tell me where to find the input data format of the ram.x? Message-ID: <20050916082945.83095.qmail@web15604.mail.cnb.yahoo.com> Could you please tell me where to find the input data format of the ram.x program ? --------------------------------- DO YOU YAHOO!? ????G??????????????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050916/95469331/attachment.htm From giannozz at nest.sns.it Fri Sep 16 11:03:23 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 16 Sep 2005 11:03:23 +0200 Subject: [Pw_forum] where can i get the input file about raman? In-Reply-To: <20050916075645.72177.qmail@web15602.mail.cnb.yahoo.com> References: <20050916075645.72177.qmail@web15602.mail.cnb.yahoo.com> Message-ID: <200509161103.23581.giannozz@nest.sns.it> On Friday 16 September 2005 09:56, li yan wrote: > i use the version 2.1.3 of pwscf and want to calculate the raman > intensity, but i can't understand the input data in the exampe15 > about raman. would you be so kind to help me? see file Doc/INPUT_RAM . See also http://www.democritos.it/pipermail/pw_forum/2005-March/002223.html http://www.democritos.it/pipermail/pw_forum/2005-April/002380.html -- 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 yi.kong at gmail.com Fri Sep 16 15:45:51 2005 From: yi.kong at gmail.com (Yi Kong) Date: Fri, 16 Sep 2005 21:45:51 +0800 Subject: [Pw_forum] problem in phonon DOS calculation of HCP structure (Job details) Message-ID: <8c9dc66c050916064590b9d68@mail.gmail.com> Dear Paolo, I run the q2r.x and matdyn.x in my single CPU personal computer again. The error is just the same. Sincerely Yi Kong -- ============================= Yi Kong Department of MSE Tsinghua University PR China, 100084 ============================= -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050916/1f0c1a45/attachment.htm From diegovmorenor at gmail.com Fri Sep 16 23:07:10 2005 From: diegovmorenor at gmail.com (Diego Moreno) Date: Fri, 16 Sep 2005 16:07:10 -0500 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #743 - 4 msgs In-Reply-To: <20050910053631.25684.50879.Mailman@democritos.sissa.it> References: <20050910053631.25684.50879.Mailman@democritos.sissa.it> Message-ID: <3af87fff05091614071244ab48@mail.gmail.com> Thanks Eyvaz. > I looked at more carefully your input file and found > an error. > You specify ntyp=2, assuming that you have 2 it's a stupid error, sorry. I interesting in used fqha.x, there is a manual for this? or input example Thanks, again. Diego From giannozz at nest.sns.it Mon Sep 19 10:56:27 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Mon, 19 Sep 2005 10:56:27 +0200 Subject: [Pw_forum] Re: Pw_forum digest, Vol 1 #743 - 4 msgs In-Reply-To: <3af87fff05091614071244ab48@mail.gmail.com> References: <20050910053631.25684.50879.Mailman@democritos.sissa.it> <3af87fff05091614071244ab48@mail.gmail.com> Message-ID: <200509191056.27827.giannozz@nest.sns.it> On Friday 16 September 2005 23:07, Diego Moreno wrote: > I interesting in used fqha.x, there is a manual for this? or input example none of the above, but the code (pwtools/fqhja,.f90) is really very simple 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 liyanpcl at yahoo.com.cn Tue Sep 20 08:10:34 2005 From: liyanpcl at yahoo.com.cn (li yan) Date: Tue, 20 Sep 2005 14:10:34 +0800 (CST) Subject: [Pw_forum] Could you please tell me where to find the input data format of the ram.x? Message-ID: <20050920061034.48085.qmail@web15604.mail.cnb.yahoo.com> Could you please tell me where to find the input data format of the ram.x program ? I cannot understand some input datas . Thank you! --------------------------------- ????G??????????????????? ?????????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050920/e53d90a6/attachment.htm From yuwen_66 at yahoo.com Tue Sep 20 09:44:50 2005 From: yuwen_66 at yahoo.com (W. YU) Date: Tue, 20 Sep 2005 00:44:50 -0700 (PDT) Subject: [Pw_forum] accustic sum rule on soft phonon mode In-Reply-To: <20050917053645.15825.87787.Mailman@democritos.sissa.it> Message-ID: <20050920074450.4629.qmail@web51006.mail.yahoo.com> Dear PW users, I met a problelm during phonon calculation. When I use accoustic sum rule (ASR) with dynmat.x on a system with soft mode, the positive frequencies were only slightly affected, but the soft mode frequency changed a lot. In this case, can I still use ASR and trust the soft mode frequency given by dynmat.x? Bye the way, I know it is possible to apply accoustic sum rule to effective charge manually. If I only do phonon calculation at a single q point, say X point for silicon, can Is there a way (manually if possible)to ensure accoustic sum rule as done by dynmat.x with phonon dispersion curves? Any comments would be much appreciated! W. YU ______________________________________________________ Yahoo! for Good Donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From arne.klaveness at kjemi.uio.no Tue Sep 20 14:20:34 2005 From: arne.klaveness at kjemi.uio.no (Arne Klaveness) Date: Tue, 20 Sep 2005 14:20:34 +0200 Subject: [Pw_forum] XML output Message-ID: <432FFE92.90108@kjemi.uio.no> Hello :) I'm trying to get the output of pwscf in the for of xml, but I have no idea of how to use the pw_export.x tool... Does anyone have an example? thanks :) Arne -------------- next part -------------- A non-text attachment was scrubbed... Name: arne.klaveness.vcf Type: text/x-vcard Size: 303 bytes Desc: not available Url : /pipermail/attachments/20050920/5f91e3c9/attachment.vcf From ferretti.andrea at unimore.it Tue Sep 20 15:06:05 2005 From: ferretti.andrea at unimore.it (Andrea Ferretti) Date: Tue, 20 Sep 2005 15:06:05 +0200 (CEST) Subject: [Pw_forum] XML output In-Reply-To: <432FFE92.90108@kjemi.uio.no> References: <432FFE92.90108@kjemi.uio.no> Message-ID: Hi Arne, here I report an example of input for pw_export.x ... ===================================== &INPUTPP prefix='silicon' outdir='./' pseudo_dir='./' psfile(1)='Si.vbc.UPF' single_file= .FALSE. ascii = .FALSE. / ===================================== prefix, outdir as in pw.x pseudo_dir if set, write pseudo_dir inside the xml output (at the moment there is no way to get this dir directly from the pwscf datafile, since it is only read from input and not stored). If you do not need pseudopot forget about this variable psfile(:) the names of the pseudopotential files (the reason for the variable is the same as before...) single_file if .TRUE. a single xml output file is produced, otherwise (which is the DEFAULT) a directory name prefix.export is created and output files are put there. ascii if .TRUE. all the files in prefix.export are textual (including wfc and grids). This is useful for debug purposes of to take a look at the format of the files... few further notes: pw_export uses the IOTK library (by Giovanni Bussi, see http://nanofs.unimo.it/~software/iotk/ or the iotkpatch from the pwscf.org website to get a manual) The library writes both textual files (xml fmt) and binary files with the same xml "logical structure"... in a sense you can think of textual xml (which actually is xml) and binary "xml-like" files.... textual and binary fmts are trated in the same way and it is possbile to convert one into the other by using the iotk_copy.x utility in ESPRESSO_HOME/iotk/bin cheers Andrea On Tue, 20 Sep 2005, Arne Klaveness wrote: > Hello :) > > I'm trying to get the output of pwscf in the for of xml, but I have no > idea of how to use the pw_export.x tool... > > Does anyone have an example? > > thanks :) > Arne > > From sbraccia at sissa.it Tue Sep 20 15:24:53 2005 From: sbraccia at sissa.it (carlo sbraccia) Date: Tue, 20 Sep 2005 09:24:53 -0400 Subject: [Pw_forum] XML output In-Reply-To: References: <432FFE92.90108@kjemi.uio.no> Message-ID: <43300DA5.1090409@sissa.it> Hi All, an alternative possibility is downloading the CVS version of the code and compiling it with -D__NEWPUNCH. The resulting executable will produce a punch file ( outdir/prefix.save-new/restart.xml, but this name will change in the forthcoming stable version ) that is already in the XML format. Moreover this format is already kind of stable and is the one that will be used the next stable release. Note that the CVS version of the CP code already uses this format for its restart file. carlo Andrea Ferretti wrote: > > Hi Arne, > here I report an example of input for pw_export.x ... > > ===================================== > > &INPUTPP > prefix='silicon' > outdir='./' > pseudo_dir='./' > psfile(1)='Si.vbc.UPF' > single_file= .FALSE. > ascii = .FALSE. > / > > ===================================== > > prefix, outdir as in pw.x > pseudo_dir if set, write pseudo_dir inside the xml output > (at the moment there is no way to get this dir > directly from the pwscf datafile, since it is only read > from input and not stored). > If you do not need pseudopot forget about this variable > psfile(:) the names of the pseudopotential files (the reason for the > variable is the same as before...) > single_file if .TRUE. a single xml output file is produced, otherwise > (which is the DEFAULT) a directory name prefix.export is created > and output files are put there. > ascii if .TRUE. all the files in prefix.export are textual > (including wfc and grids). This is useful for debug purposes > of to take a look at the format of the files... > > few further notes: > pw_export uses the IOTK library (by Giovanni Bussi, > see http://nanofs.unimo.it/~software/iotk/ or the iotkpatch from the > pwscf.org website to get a manual) > > The library writes both textual files (xml fmt) and binary files with the same > xml "logical structure"... in a sense you can think of textual xml (which actually > is xml) and binary "xml-like" files.... > textual and binary fmts are trated in the same way and it is possbile to convert > one into the other by using the iotk_copy.x utility in ESPRESSO_HOME/iotk/bin > > cheers > Andrea > > > > On Tue, 20 Sep 2005, Arne Klaveness wrote: > > >>Hello :) >> >>I'm trying to get the output of pwscf in the for of xml, but I have no >>idea of how to use the pw_export.x tool... >> >>Does anyone have an example? >> >>thanks :) >>Arne >> >> > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > From arne.klaveness at kjemi.uio.no Wed Sep 21 09:27:46 2005 From: arne.klaveness at kjemi.uio.no (Arne Klaveness) Date: Wed, 21 Sep 2005 09:27:46 +0200 Subject: [Pw_forum] XML output In-Reply-To: <43300DA5.1090409@sissa.it> References: <432FFE92.90108@kjemi.uio.no> <43300DA5.1090409@sissa.it> Message-ID: <43310B72.1080607@kjemi.uio.no> Thanks :). Compiled the CVS version with the extra tag, and the xml output looks good. A couple of questions: -Is there a schema somewhere? -Is the design of the binary (.rho) files available anywhere? Thanks again :) Arne On 20.09.2005 15:24, carlo sbraccia wrote: > Hi All, > > an alternative possibility is downloading the CVS version of the code > and compiling it with -D__NEWPUNCH. The resulting executable will > produce a punch file ( outdir/prefix.save-new/restart.xml, but this name > will change in the forthcoming stable version ) that is already in the > XML format. Moreover this format is already kind of stable and is the > one that will be used the next stable release. Note that the CVS version > of the CP code already uses this format for its restart file. > > carlo > > Andrea Ferretti wrote: > >> >> Hi Arne, >> here I report an example of input for pw_export.x ... >> >> ===================================== >> >> &INPUTPP >> prefix='silicon' >> outdir='./' >> pseudo_dir='./' >> psfile(1)='Si.vbc.UPF' >> single_file= .FALSE. >> ascii = .FALSE. >> / >> >> ===================================== >> >> prefix, outdir as in pw.x >> pseudo_dir if set, write pseudo_dir inside the xml output >> (at the moment there is no way to get this dir >> directly from the pwscf datafile, since it is only read >> from input and not stored). >> If you do not need pseudopot forget about this variable >> psfile(:) the names of the pseudopotential files (the reason >> for the >> variable is the same as before...) >> single_file if .TRUE. a single xml output file is produced, >> otherwise >> (which is the DEFAULT) a directory name >> prefix.export is created >> and output files are put there. >> ascii if .TRUE. all the files in prefix.export are textual >> (including wfc and grids). This is useful for debug >> purposes of to take a look at the format of the >> files... >> >> few further notes: >> pw_export uses the IOTK library (by Giovanni Bussi, see >> http://nanofs.unimo.it/~software/iotk/ or the iotkpatch from the >> pwscf.org website to get a manual) >> The library writes both textual files (xml fmt) and >> binary files with the same >> xml "logical structure"... in a sense you can think of textual xml >> (which actually is xml) and binary "xml-like" files.... >> textual and binary fmts are trated in the same way and it is possbile >> to convert >> one into the other by using the iotk_copy.x utility in >> ESPRESSO_HOME/iotk/bin >> >> cheers >> Andrea >> >> >> >> On Tue, 20 Sep 2005, Arne Klaveness wrote: >> >> >>> Hello :) >>> >>> I'm trying to get the output of pwscf in the for of xml, but I have no >>> idea of how to use the pw_export.x tool... >>> >>> Does anyone have an example? >>> >>> thanks :) >>> Arne >>> >>> >> >> _______________________________________________ >> Pw_forum mailing list >> Pw_forum at pwscf.org >> http://www.democritos.it/mailman/listinfo/pw_forum >> > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -------------- next part -------------- A non-text attachment was scrubbed... Name: arne.klaveness.vcf Type: text/x-vcard Size: 303 bytes Desc: not available Url : /pipermail/attachments/20050921/36ca0f6a/attachment.vcf From sbraccia at sissa.it Wed Sep 21 15:24:00 2005 From: sbraccia at sissa.it (carlo sbraccia) Date: Wed, 21 Sep 2005 09:24:00 -0400 Subject: [Pw_forum] XML output In-Reply-To: <43310B72.1080607@kjemi.uio.no> References: <432FFE92.90108@kjemi.uio.no> <43300DA5.1090409@sissa.it> <43310B72.1080607@kjemi.uio.no> Message-ID: <43315EF0.6030805@sissa.it> Unfortunately, not yet. At the moment the only source of information consists in the files Modules/xml_io_base.f90, PW/pw_restart, and CPV/cp_restart.f90. carlo Arne Klaveness wrote: > Thanks :). Compiled the CVS version with the extra tag, and the xml > output looks good. > > A couple of questions: > -Is there a schema somewhere? > -Is the design of the binary (.rho) files available anywhere? > > Thanks again :) > Arne From konstantin_kudin at yahoo.com Wed Sep 21 19:48:32 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Wed, 21 Sep 2005 10:48:32 -0700 (PDT) Subject: [Pw_forum] abysmal parallel performance of the CP code Message-ID: <20050921174832.8605.qmail@web52006.mail.yahoo.com> Hi, I've done some parallel benchmarks for the CP code so I thought I'd share them with the rest of the group. The system we have is a cluster of dual Opterons 2.0 Ghz with 1Gbit ethernet. I looked at 2 different measures of time, CPU time, and wall time computed as the difference between "This run was started" and "This run was terminated". By the way, such wall time could probably be printed by the code directly to be readily available. The system is a reasonably sized simulation cell with 20 CP (electronic+ionic) steps total. The compiler is IFC 9.0, GOTO library is for BLAS, and mpich 1.2.6 used for the MPI. The CP version is the CVS from Aug. 20, 2005. What is crazy is that even for 2 cpus sitting in the same box there is lots of cpu time just lost somewhere. The strange thing is that the quad we have at 2.2 Ghz seems to lose just as much wall time as 2 duals talking across the network. And note how 4 cpus are barely better than 2x compared to single cpu performance if the wall clock time is considered. I know Nicola Marzari has done some parallel benchmarks, but I do not think that wall times were being paid attention to ... Kostya P.S. Any suggestions what might be going on here? Ncpu CPU time Wall time 1 1h22m 1h24m 2 45m33.41s 57m13s 4 27m30.80s 44m21s 6 18m22.71s 43m18s 8 14m53.91s 45m56s 4(quad) 37m18.56s 45m32s __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From silviu at Princeton.EDU Wed Sep 21 23:19:22 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Wed, 21 Sep 2005 23:19:22 +0200 Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <20050921174832.8605.qmail@web52006.mail.yahoo.com> References: <20050921174832.8605.qmail@web52006.mail.yahoo.com> Message-ID: <4331CE5A.8040400@Princeton.EDU> Hi Kostya, I am not sure if that's the case but I also noticed similar problems in the past. My impression then was that some of the difference is due to the fact that the reported wall time includes the time it takes to read the initial and write the final restart files to the disk, in contrast to the reported CPU time. In some cases, if the cluster network is very loaded, it may take several minutes (!) to write big files (hundreds of MB). In CP the restart file is not partitioned as in PW, so there is a lot of traffic in collecting data from all the nodes and then actually writing it to the disk. For long runs you don't see the effect of the additional last disk write, but when having only 20 md steps, it may become dominant. Were you also writing intermediate restart files during the 20 steps of the benchmark? Silviu. Konstantin Kudin wrote: > Hi, > > I've done some parallel benchmarks for the CP code so I thought I'd >share them with the rest of the group. The system we have is a cluster >of dual Opterons 2.0 Ghz with 1Gbit ethernet. > > I looked at 2 different measures of time, CPU time, and wall time >computed as the difference between "This run was started" and "This run >was terminated". By the way, such wall time could probably be printed >by the code directly to be readily available. > > The system is a reasonably sized simulation cell with 20 CP >(electronic+ionic) steps total. > > The compiler is IFC 9.0, GOTO library is for BLAS, and mpich 1.2.6 >used for the MPI. The CP version is the CVS from Aug. 20, 2005. > > What is crazy is that even for 2 cpus sitting in the same box there is >lots of cpu time just lost somewhere. The strange thing is that the >quad we have at 2.2 Ghz seems to lose just as much wall time as 2 duals >talking across the network. And note how 4 cpus are barely better than >2x compared to single cpu performance if the wall clock time is >considered. > > I know Nicola Marzari has done some parallel benchmarks, but I do not >think that wall times were being paid attention to ... > > Kostya > >P.S. Any suggestions what might be going on here? > > >Ncpu CPU time Wall time >1 1h22m 1h24m >2 45m33.41s 57m13s >4 27m30.80s 44m21s >6 18m22.71s 43m18s >8 14m53.91s 45m56s > >4(quad) 37m18.56s 45m32s > > > >__________________________________________________ >Do You Yahoo!? >Tired of spam? Yahoo! Mail has the best spam protection around >http://mail.yahoo.com >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Zilberman Silviu 213 Frick Laboratory, Department of Chemistry Princeton University Princeton, NJ 08544 phone: 609-258-1834 fax: 609-258-6746 silviu at Princeton.EDU %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -------------- next part -------------- A non-text attachment was scrubbed... Name: silviu.vcf Type: text/x-vcard Size: 272 bytes Desc: not available Url : /pipermail/attachments/20050921/9c03c57c/attachment.vcf From konstantin_kudin at yahoo.com Thu Sep 22 00:51:55 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Wed, 21 Sep 2005 15:51:55 -0700 (PDT) Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <4331CE5A.8040400@Princeton.EDU> Message-ID: <20050921225156.9501.qmail@web52007.mail.yahoo.com> Silviu, Indeed, with a busy network writing restart files does take a while, even on a single cpu there are couple extra minutes in the wall times. However, this is nothing like the difference on 2 and more cpus. We tracked down the slow speed on the quad to the NUMA being turned off in contrast to duals, so those differences are explainable by the memory accesses. Still, an interesting question is why CP is such a dog already on 2 cpus as seeing in the huge difference between the wall and cpu times. 2 cpus in a box never talk across the network except for the occasional restart files. I am curious if other clusters behave in the same way as far as wall vs. cpu times are concerned. Kostya --- Silviu Zilberman wrote: > Hi Kostya, > > I am not sure if that's the case but I also noticed similar problems > in > the past. My impression then was that some of the difference is due > to > the fact that the reported wall time includes the time it takes to > read > the initial and write the final restart files to the disk, in > contrast > to the reported CPU time. In some cases, if the cluster network is > very > loaded, it may take several minutes (!) to write big files (hundreds > of > MB). In CP the restart file is not partitioned as in PW, so there is > a > lot of traffic in collecting data from all the nodes and then > actually > writing it to the disk. For long runs you don't see the effect of the > > additional last disk write, but when having only 20 md steps, it may > become dominant. Were you also writing intermediate restart files > during > the 20 steps of the benchmark? > > Silviu. > > Konstantin Kudin wrote: > > > Hi, > > > > I've done some parallel benchmarks for the CP code so I thought I'd > >share them with the rest of the group. The system we have is a > cluster > >of dual Opterons 2.0 Ghz with 1Gbit ethernet. > > > > I looked at 2 different measures of time, CPU time, and wall time > >computed as the difference between "This run was started" and "This > run > >was terminated". By the way, such wall time could probably be > printed > >by the code directly to be readily available. > > > > The system is a reasonably sized simulation cell with 20 CP > >(electronic+ionic) steps total. > > > > The compiler is IFC 9.0, GOTO library is for BLAS, and mpich 1.2.6 > >used for the MPI. The CP version is the CVS from Aug. 20, 2005. > > > > What is crazy is that even for 2 cpus sitting in the same box there > is > >lots of cpu time just lost somewhere. The strange thing is that the > >quad we have at 2.2 Ghz seems to lose just as much wall time as 2 > duals > >talking across the network. And note how 4 cpus are barely better > than > >2x compared to single cpu performance if the wall clock time is > >considered. > > > > I know Nicola Marzari has done some parallel benchmarks, but I do > not > >think that wall times were being paid attention to ... > > > > Kostya > > > >P.S. Any suggestions what might be going on here? > > > > > >Ncpu CPU time Wall time > >1 1h22m 1h24m > >2 45m33.41s 57m13s > >4 27m30.80s 44m21s > >6 18m22.71s 43m18s > >8 14m53.91s 45m56s > > > >4(quad) 37m18.56s 45m32s > > > > > > > >__________________________________________________ > >Do You Yahoo!? > >Tired of spam? Yahoo! Mail has the best spam protection around > >http://mail.yahoo.com > >_______________________________________________ > >Pw_forum mailing list > >Pw_forum at pwscf.org > >http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > > -- > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > Zilberman Silviu > > 213 Frick Laboratory, Department of Chemistry > Princeton University > Princeton, NJ 08544 > > phone: 609-258-1834 > fax: 609-258-6746 > silviu at Princeton.EDU > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > > > begin:vcard > fn:Silviu Zilberman > n:Zilberman;Silviu > org:Princeton University;Department of Chemistry > adr:;;213 Frick Laboratory;Princeton;New-Jersy;08544;USA > title:Research Associate > tel;work:1-609-258-1834 > tel;fax:1-609-258-6746 > x-mozilla-html:TRUE > version:2.1 > end:vcard > > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From yay451 at mail.usask.ca Thu Sep 22 09:09:31 2005 From: yay451 at mail.usask.ca (Yansun Yao) Date: Thu, 22 Sep 2005 01:09:31 -0600 Subject: [Pw_forum] error in scf calculation Message-ID: <1127372971.433258abc73a7@webmail.usask.ca> Dear Pwscf users, I am doing a scf calculation of Al(OH)3. At the very beginning of the output file the warning come as: Warning: card &IONS ignored Warning: card ION_DYNAMICS = 'DAMP', UPSCALE=20.0, ignored Warning: card / ignored Warning: card &CELL ignored Warning: card CELL_DYNAMICS = 'DAMP-PR', PRESS=0.0, ignored Warning: card / ignored warning: symmetry operation # 6 not allowed. fractional translation: -0.5000000 0.0000000 -0.4717471 in crystal coordinates This is the input file: &control calculation='scf', restart_mode='from_scratch', prefix='aloh3' pseudo_dir = '/home/espresso-2.1.4/pseudo/', outdir='/home/tmp/' / &system ibrav = 8, celldm(1) = 16.59746, celldm(2) = 0.58124 , celldm(3) = 1.05989, nat= 56, ntyp= 3, ecutwfc = 300.0, occupations = 'fixed' / &electrons mixing_beta = 0.7 conv_thr = 1.0d-10 / &ions ion_dynamics = 'damp', upscale=20.0, / &cell cell_dynamics = 'damp-pr', press=0.0, / ATOMIC_SPECIES Al 26.98 Al.pbe-rrkj.UPF O 15.9994 O.pbe-rrkjus.UPF H 1.008 H.pbe-rrkjus.UPF ATOMIC_POSITIONS Al 0.3348 0.0345 0.9857 Al 0.1652 -0.0345 1.4857 Al -0.3348 0.5345 -0.4857 Al 0.8348 0.4655 -0.9857 Al -0.3348 -0.0345 -0.9857 Al 0.8348 0.0345 -0.4857 Al 0.3348 0.4655 1.4857 Al 0.1652 0.5345 0.9857 O 0.9816 0.6360 0.8869 O -0.4816 -0.6360 1.3869 O -0.9816 1.1360 -0.3869 O 1.4816 -0.1360 -0.8869 O -0.9816 -0.6360 -0.8869 O 1.4816 0.6360 -0.3869 O 0.9816 -0.1360 1.3869 O -0.4816 1.1360 0.8869 O 0.3018 0.7334 0.8665 O 0.1982 -0.7334 1.3665 O -0.3018 1.2334 -0.3665 O 0.8018 -0.2334 -0.8665 O -0.3018 -0.7334 -0.8665 O 0.8018 0.7334 -0.3665 O 0.3018 -0.2334 1.3665 O 0.1982 1.2334 0.8665 O 0.8300 0.1387 0.9088 O -0.3300 -0.1387 1.4088 O -0.8300 0.6387 -0.4088 O 1.3300 0.3613 -0.9088 O -0.8300 -0.1387 -0.9088 O 1.3300 0.1387 -0.4088 O 0.8300 0.3613 1.4088 O -0.3300 0.6387 0.9088 H 0.9573 0.8260 0.8928 H -0.4573 -0.8260 1.3928 H -0.9573 1.3260 -0.3928 H 1.4573 -0.3260 -0.8928 H -0.9573 -0.8260 -0.8928 H 1.4573 0.8260 -0.3928 H 0.9573 -0.3260 1.3928 H -0.4573 1.3260 0.8928 H 0.3877 0.6600 0.8110 H 0.1123 -0.6600 1.3110 H -0.3877 1.1600 -0.3110 H 0.8877 -0.1600 -0.8110 H -0.3877 -0.6600 -0.8110 H 0.8877 0.6600 -0.3110 H 0.3877 -0.1600 1.3110 H 0.1123 1.1600 0.8110 H 0.8048 0.1696 0.8062 H -0.3048 -0.1696 1.3062 H -0.8048 0.6696 -0.3062 H 1.3048 0.3304 -0.8062 H -0.8048 -0.1696 -0.8062 H 1.3048 0.1696 -0.3062 H 0.8048 0.3304 1.3062 H -0.3048 0.6696 0.8062 K_POINTS {automatic} 3,6,3,0,0,0 From the previous posts, it looks like the symmetry warning is halmless? I have no idea about thoese card warnings. Could you please give me some help? Thank you in advance ! Yansun University of Saskatchewan From degironc at sissa.it Thu Sep 22 09:46:19 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Thu, 22 Sep 2005 09:46:19 +0200 Subject: [Pw_forum] error in scf calculation References: <1127372971.433258abc73a7@webmail.usask.ca> Message-ID: <4332614B.20703@sissa.it> IONS inamelist is read and used only if atoms are allowed to move (relax, md or vc-relax,vc-md type of calculation); CELL namelist is read and used only if you do a vc-md of vc-relax calculation not when doing just a scf calculation, hence the warning message saying that they are present but have been ignored. stefano Yansun Yao wrote: >Dear Pwscf users, > I am doing a scf calculation of Al(OH)3. At the very beginning of the output >file the warning come as: > Warning: card &IONS ignored > Warning: card ION_DYNAMICS = 'DAMP', UPSCALE=20.0, ignored > Warning: card / ignored > Warning: card &CELL ignored > Warning: card CELL_DYNAMICS = 'DAMP-PR', PRESS=0.0, ignored > Warning: card / ignored > warning: symmetry operation # 6 not allowed. fractional translation: > -0.5000000 0.0000000 -0.4717471 in crystal coordinates > This is the input file: > > &control > calculation='scf', > restart_mode='from_scratch', > prefix='aloh3' > pseudo_dir = '/home/espresso-2.1.4/pseudo/', > outdir='/home/tmp/' > / > &system > ibrav = 8, celldm(1) = 16.59746, celldm(2) = 0.58124 , celldm(3) = 1.05989, >nat= 56, ntyp= 3, > ecutwfc = 300.0, occupations = 'fixed' > / > &electrons > mixing_beta = 0.7 > conv_thr = 1.0d-10 > / > &ions > ion_dynamics = 'damp', upscale=20.0, > / > &cell > cell_dynamics = 'damp-pr', press=0.0, > / >ATOMIC_SPECIES > Al 26.98 Al.pbe-rrkj.UPF > O 15.9994 O.pbe-rrkjus.UPF > H 1.008 H.pbe-rrkjus.UPF >ATOMIC_POSITIONS > Al 0.3348 0.0345 0.9857 > Al 0.1652 -0.0345 1.4857 > Al -0.3348 0.5345 -0.4857 > Al 0.8348 0.4655 -0.9857 > Al -0.3348 -0.0345 -0.9857 > Al 0.8348 0.0345 -0.4857 > Al 0.3348 0.4655 1.4857 > Al 0.1652 0.5345 0.9857 > O 0.9816 0.6360 0.8869 > O -0.4816 -0.6360 1.3869 > O -0.9816 1.1360 -0.3869 > O 1.4816 -0.1360 -0.8869 > O -0.9816 -0.6360 -0.8869 > O 1.4816 0.6360 -0.3869 > O 0.9816 -0.1360 1.3869 > O -0.4816 1.1360 0.8869 > O 0.3018 0.7334 0.8665 > O 0.1982 -0.7334 1.3665 > O -0.3018 1.2334 -0.3665 > O 0.8018 -0.2334 -0.8665 > O -0.3018 -0.7334 -0.8665 > O 0.8018 0.7334 -0.3665 > O 0.3018 -0.2334 1.3665 > O 0.1982 1.2334 0.8665 > O 0.8300 0.1387 0.9088 > O -0.3300 -0.1387 1.4088 > O -0.8300 0.6387 -0.4088 > O 1.3300 0.3613 -0.9088 > O -0.8300 -0.1387 -0.9088 > O 1.3300 0.1387 -0.4088 > O 0.8300 0.3613 1.4088 > O -0.3300 0.6387 0.9088 > H 0.9573 0.8260 0.8928 > H -0.4573 -0.8260 1.3928 > H -0.9573 1.3260 -0.3928 > H 1.4573 -0.3260 -0.8928 > H -0.9573 -0.8260 -0.8928 > H 1.4573 0.8260 -0.3928 > H 0.9573 -0.3260 1.3928 > H -0.4573 1.3260 0.8928 > H 0.3877 0.6600 0.8110 > H 0.1123 -0.6600 1.3110 > H -0.3877 1.1600 -0.3110 > H 0.8877 -0.1600 -0.8110 > H -0.3877 -0.6600 -0.8110 > H 0.8877 0.6600 -0.3110 > H 0.3877 -0.1600 1.3110 > H 0.1123 1.1600 0.8110 > H 0.8048 0.1696 0.8062 > H -0.3048 -0.1696 1.3062 > H -0.8048 0.6696 -0.3062 > H 1.3048 0.3304 -0.8062 > H -0.8048 -0.1696 -0.8062 > H 1.3048 0.1696 -0.3062 > H 0.8048 0.3304 1.3062 > H -0.3048 0.6696 0.8062 >K_POINTS {automatic} > 3,6,3,0,0,0 > From the previous posts, it looks like the symmetry warning is halmless? I >have no idea about thoese card warnings. Could you please give me some help? >Thank you in advance ! > Yansun > University of Saskatchewan > > > > >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Thu Sep 22 12:07:20 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Thu, 22 Sep 2005 12:07:20 +0200 (CEST) Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <20050921174832.8605.qmail@web52006.mail.yahoo.com> Message-ID: On Wed, 21 Sep 2005, Konstantin Kudin wrote: KK> Hi, kostya, KK> I've done some parallel benchmarks for the CP code so I thought I'd KK> share them with the rest of the group. The system we have is a cluster KK> of dual Opterons 2.0 Ghz with 1Gbit ethernet. please keep in mind, that for reasonable scaling with car-parrinello MD you usually need a better interconnect than gigabit ethernet. some time ago i summarized some tests results for the CPMD code at: http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/cpmd-bench.html#parallel the general issues apply to the espresso CP codes as well as to CPMD. with gigabit (or TCP/IP at that) you suffer most from the very high latencies. this is especially bad for the all-to-all communications that are needed for the FFTs, and more visibly for dual-cpu. for quad-cpu nodes you also should be in the bandwidth limit. now, how much this becomes visible depends much on the size of the job. we have a cluster of dual-opteron 246 (2.0GHz) with two gigabit networks (data and MPI separately) and it usually does not pay to run jobs across more than 3-4 nodes. and even then you already 'waste' about 20% of your cpu power. the only saving grace is the fact, that a better interconnect will coste much more than one wasted node. KK> I looked at 2 different measures of time, CPU time, and wall time KK> computed as the difference between "This run was started" and "This run KK> was terminated". By the way, such wall time could probably be printed KK> by the code directly to be readily available. probably, but you can also get the number as easily by using the 'time' command to start the jobs. KK> The system is a reasonably sized simulation cell with 20 CP KK> (electronic+ionic) steps total. KK> KK> The compiler is IFC 9.0, GOTO library is for BLAS, and mpich 1.2.6 KK> used for the MPI. The CP version is the CVS from Aug. 20, 2005. KK> KK> What is crazy is that even for 2 cpus sitting in the same box there is KK> lots of cpu time just lost somewhere. The strange thing is that the KK> quad we have at 2.2 Ghz seems to lose just as much wall time as 2 duals KK> talking across the network. And note how 4 cpus are barely better than KK> 2x compared to single cpu performance if the wall clock time is KK> considered. please check, whether your MPI library does use shared memory communication properly and that your kernel supports setting the proper CPU and Memory affinity (and you set it). i have seen some numbers where this makes over 20% difference on a dual machine and i would expect it matters even more on quad machines. KK> I know Nicola Marzari has done some parallel benchmarks, but I do not KK> think that wall times were being paid attention to ... KK> KK> Kostya KK> KK> P.S. Any suggestions what might be going on here? you also have to take into account, that when you are running a gamma point only calculation, you are missing the most efficient parallelization (across k-points) that helps running, e.g., pw.x rather efficiently on 'not so high'-performance networks. best regards, axel. KK> KK> KK> Ncpu CPU time Wall time KK> 1 1h22m 1h24m KK> 2 45m33.41s 57m13s KK> 4 27m30.80s 44m21s KK> 6 18m22.71s 43m18s KK> 8 14m53.91s 45m56s KK> KK> 4(quad) 37m18.56s 45m32s KK> KK> KK> KK> __________________________________________________ KK> Do You Yahoo!? KK> Tired of spam? Yahoo! Mail has the best spam protection around KK> http://mail.yahoo.com KK> _______________________________________________ KK> Pw_forum mailing list KK> Pw_forum at pwscf.org KK> http://www.democritos.it/mailman/listinfo/pw_forum KK> KK> -- ======================================================================= Dr. Axel Kohlmeyer e-mail: axel.kohlmeyer at theochem.ruhr-uni-bochum.de Lehrstuhl fuer Theoretische Chemie Phone: ++49 (0)234/32-26673 Ruhr-Universitaet Bochum - NC 03/53 Fax: ++49 (0)234/32-14045 D-44780 Bochum http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/ ======================================================================= If you make something idiot-proof, the universe creates a better idiot. From shaposh at isp.nsc.ru Thu Sep 22 12:41:55 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Thu, 22 Sep 2005 17:41:55 +0700 Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <20050921174832.8605.qmail@web52006.mail.yahoo.com> References: <20050921174832.8605.qmail@web52006.mail.yahoo.com> Message-ID: <1127385715.3000.8.camel@m00> I hope you did not forget to set GOTO_NUM_THREADS=1 env. variable? I usually have 2.5 - 3 speedup for CP on Dual DualCore Opteron Machine. Gigabit network is really not fast enough for this kind of computations. It has latencies 10 times worse than myrinet and infiniband networks. The result for the quad machine is strange. Should be faster. If possible, provide your input file, i can test it on our DualCore boxes. On Wed, 2005-09-21 at 10:48 -0700, Konstantin Kudin wrote: > Hi, > > I've done some parallel benchmarks for the CP code so I thought I'd > share them with the rest of the group. The system we have is a cluster > of dual Opterons 2.0 Ghz with 1Gbit ethernet. > > I looked at 2 different measures of time, CPU time, and wall time > computed as the difference between "This run was started" and "This run > was terminated". By the way, such wall time could probably be printed > by the code directly to be readily available. > > The system is a reasonably sized simulation cell with 20 CP > (electronic+ionic) steps total. > > The compiler is IFC 9.0, GOTO library is for BLAS, and mpich 1.2.6 > used for the MPI. The CP version is the CVS from Aug. 20, 2005. > > What is crazy is that even for 2 cpus sitting in the same box there is > lots of cpu time just lost somewhere. The strange thing is that the > quad we have at 2.2 Ghz seems to lose just as much wall time as 2 duals > talking across the network. And note how 4 cpus are barely better than > 2x compared to single cpu performance if the wall clock time is > considered. > > I know Nicola Marzari has done some parallel benchmarks, but I do not > think that wall times were being paid attention to ... > > Kostya > > P.S. Any suggestions what might be going on here? > > > Ncpu CPU time Wall time > 1 1h22m 1h24m > 2 45m33.41s 57m13s > 4 27m30.80s 44m21s > 6 18m22.71s 43m18s > 8 14m53.91s 45m56s > > 4(quad) 37m18.56s 45m32s > > > > __________________________________________________ > Do You Yahoo!? > Tired of spam? Yahoo! Mail has the best spam protection around > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum From wmbmacam at lg.ehu.es Thu Sep 22 16:34:29 2005 From: wmbmacam at lg.ehu.es (=?UTF-8?B?TWlndWVsIE1hcnTDrW5leiBDYW5hbGVz?=) Date: Thu, 22 Sep 2005 16:34:29 +0200 Subject: [Pw_forum] Electron-Phonon: Two questions Message-ID: <4332C0F5.2090106@lg.ehu.es> Dear people, I am doing a "warm up" electron-phonon calculation with Al. I am basically looking for convergence and similar issues (i.e. gaussian broadening). I have the following issues (v2.1.3 & v.2.1.5): - One of the calculations, although it yields results, gives me the following error message: task # 0 from set_kplusq : error # 0 too many k points should I be worried? The loop at set_kplusq.f90 doesn't stop with this error message. Is this correct? - On the second non-selfconsistent run, with a dense grid (32 32 32 and higher) sometimes kpoint counting surpasses k(9999) and k(****) follows. Is this the source of the previous error? Is this normal? - I am also seeing some davcio errors, maybe these are related maybe not. If anyone wants the input file, I'd be glad to put it. Although I wouldn't bet my life, it seems that the program is well compiled, as examples seem to run just fine. Thanks in advance, Miguel -- ---------------------------------------- Miguel Mart?nez Canales Dto. F?sica de la Materia Condensada UPV/EHU Facultad de Ciencia y Tecnolog?a Apdo. 644 48080 Bilbao (Spain) Fax: +34 94 601 3500 Tlf: +34 94 601 5437 ---------------------------------------- 2005 International Year of Physics From massimiliano.bonomi at mi.infn.it Thu Sep 22 18:09:46 2005 From: massimiliano.bonomi at mi.infn.it (Massimiliano Bonomi) Date: Thu, 22 Sep 2005 18:09:46 +0200 Subject: [Pw_forum] electron-phonon coupling Message-ID: <1127405386.21641.14.camel@belfagor.mi.infn.it> Dear users, I'm calculating electron-phonon coupling for a solid assembled by cluster. In order to get lambda with espresso package i have to calculate lambda(q) for a regular q-grid and then make a summation (using lambda.x). In case of a very very flat phonon dispersion curve, do you think i can approximate lambda as a constant as a function of q, in order to get an extimate of lambda only evaluating it at Gamma point?? Thanks in advance. Massimiliano Bonomi. -- Massimiliano Bonomi Dipartimento di Fisica, Universit? degli Studi di Milano via Celoria, 16 - 20133 Milano, Italy ph. +39 02 50317 717 e-mail: massimiliano.bonomi at mi.infn.it, massimiliano.bonomi at unimi.it From giannozz at nest.sns.it Fri Sep 23 10:18:00 2005 From: giannozz at nest.sns.it (Paolo Giannozzi) Date: Fri, 23 Sep 2005 10:18:00 +0200 Subject: [Pw_forum] electron-phonon coupling In-Reply-To: <1127405386.21641.14.camel@belfagor.mi.infn.it> References: <1127405386.21641.14.camel@belfagor.mi.infn.it> Message-ID: <200509231018.00248.giannozz@nest.sns.it> On Thursday 22 September 2005 18:09, Massimiliano Bonomi wrote: > In case of a very very flat phonon dispersion curve, do you think > i can approximate lambda as a constant as a function of q, in order > to get an extimate of lambda only evaluating it at Gamma point?? it might work; in order to be sure you should compare the results of approximate formulations of the electron phonon coefficients, that are valid for molecular solids, with the results of a full calculation. Some work on electron-phonon coefficients in cluster-assembled materials has been done in Milan at Material Science: see for instance the PhD thesis of Ivan Spagnolatti. Paolo -- Paolo Giannozzi e-mail: giannozz at nest.sns.it Scuola Normale Superiore Phone: +39/050-509876, Fax:-563513 Piazza dei Cavalieri 7 I-56126 Pisa, Italy From silviu at Princeton.EDU Sun Sep 25 15:01:01 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Sun, 25 Sep 2005 15:01:01 +0200 Subject: [Pw_forum] pp freezes Message-ID: <43369F8D.9030302@Princeton.EDU> Hi, I've tried to play with the latest (cvs) version of pp.x, that incorporates chdens.x. The first step (corresponding to the original pp.x) seems to work properly, but the second phase (the chdens phase) halts without exiting. Below are the input and output files. I use two processors on a single AMD opteron node, intel V8 compiler. Is there anything wrong in the way I use the code? Thanks, Silviu. --------------------------------------------------------------------------------------------- INPUT FILE: &INPUTPP prefix = 'DTNH_CO_hbond' , outdir = '/pvfs/silviu/' , filplot = '/pvfs/silviu/DTNH23u.den' , plot_num = 7, kpoint = 1, kband = 23, lsign = .true. , / &PLOT filepp(1) = '/pvfs/silviu/DTNH23u.den' , weight(1) = 1.0 , iflag = 3 , output_format = 5 , fileout = '/pvfs/silviu/DTNH23u.xsf' , / --------------------------------------------------------------------------------------------- OUTPUT FILE: Program POST-PROC v.3.0 starts ... Today is 25Sep2005 at 6:57:28 Parallel version (MPI) Number of processors in use: 2 R & G space division: proc/pool = 2 Reading file DTNH_CO_hbond.save ... only dimensions read complete Reading file DTNH_CO_hbond.save ... all except wavefuctions read complete Planes per process (thick) : nr3 =144 npp = 72 ncplane =20736 Planes per process (smooth): nr3s= 90 npps= 45 ncplanes= 8100 Proc/ planes cols G planes cols G columns G Pool (dense grid) (smooth grid) (wavefct grid) 1 72 7457 685745 45 2987 173525 739 21689 2 72 7456 685736 45 2986 173446 742 21698 0 144 149131371481 90 5973 346971 1481 43387 nbndx = 68 nbnd = 68 natomwfc = 111 npwx = 10845 nelec = 135.00 nkb = 227 ngl = 3959 Calling punch_plot, plot_num = 7 Writing data to file /pvfs/silviu/DTNH23u.den Reading header from file /pvfs/silviu/DTNH23u.den -------------- next part -------------- A non-text attachment was scrubbed... Name: silviu.vcf Type: text/x-vcard Size: 272 bytes Desc: not available Url : /pipermail/attachments/20050925/fab4bb48/attachment.vcf From physicalreviewb at yahoo.com Mon Sep 26 09:12:15 2005 From: physicalreviewb at yahoo.com (Jigang Zhang) Date: Mon, 26 Sep 2005 00:12:15 -0700 (PDT) Subject: [Pw_forum] symmetry error Message-ID: <20050926071215.97712.qmail@web33605.mail.mud.yahoo.com> Hello Pwscf users, I am running a vc-relax calculation. The output gives this error: from setup : error # -1 Dynamics, you should have no symmetries But, according the output after this error, I think the program has got all the symmetries expected. Could anybody give me some explanation about it? Thank you so much in advance! Jigang Tsinghua University ______________________________________________________ Yahoo! for Good Donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From silviu at Princeton.EDU Mon Sep 26 10:57:56 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Mon, 26 Sep 2005 10:57:56 +0200 Subject: [Pw_forum] pp freezes In-Reply-To: <43369F8D.9030302@Princeton.EDU> References: <43369F8D.9030302@Princeton.EDU> Message-ID: <4337B814.5050007@Princeton.EDU> It works however for the parallel code when using just one processor. I think it fails somewhere in subroutine 'data_structure', which is called from allocate_fft. Why do we need a second call to allocate_fft, it is already done once in the first part of pp. Silviu. Silviu Zilberman wrote: > Hi, > > I've tried to play with the latest (cvs) version of pp.x, that > incorporates chdens.x. > > The first step (corresponding to the original pp.x) seems to work > properly, but the second phase (the chdens phase) halts without > exiting. Below are the input and output files. I use two processors on > a single AMD opteron node, intel V8 compiler. Is there anything wrong > in the way I use the code? > > Thanks, Silviu. > > --------------------------------------------------------------------------------------------- > > INPUT FILE: > > &INPUTPP > prefix = 'DTNH_CO_hbond' , > outdir = '/pvfs/silviu/' , > filplot = '/pvfs/silviu/DTNH23u.den' , > plot_num = 7, > kpoint = 1, > kband = 23, > lsign = .true. , > / > &PLOT > filepp(1) = '/pvfs/silviu/DTNH23u.den' , > weight(1) = 1.0 , > iflag = 3 , > output_format = 5 , > fileout = '/pvfs/silviu/DTNH23u.xsf' , > / > > --------------------------------------------------------------------------------------------- > > OUTPUT FILE: > > Program POST-PROC v.3.0 starts ... > Today is 25Sep2005 at 6:57:28 > > Parallel version (MPI) > > Number of processors in use: 2 > R & G space division: proc/pool = 2 > > Reading file DTNH_CO_hbond.save ... only dimensions > read complete > > Reading file DTNH_CO_hbond.save ... all except wavefuctions > read complete > > Planes per process (thick) : nr3 =144 npp = 72 ncplane =20736 > > Planes per process (smooth): nr3s= 90 npps= 45 ncplanes= 8100 > > Proc/ planes cols G planes cols G columns G > Pool (dense grid) (smooth grid) (wavefct grid) > 1 72 7457 685745 45 2987 173525 739 21689 > 2 72 7456 685736 45 2986 173446 742 21698 > 0 144 149131371481 90 5973 346971 1481 43387 > > > nbndx = 68 nbnd = 68 natomwfc = 111 npwx = 10845 > nelec = 135.00 nkb = 227 ngl = 3959 > > Calling punch_plot, plot_num = 7 > Writing data to file /pvfs/silviu/DTNH23u.den > Reading header from file /pvfs/silviu/DTNH23u.den > -- %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% Zilberman Silviu 213 Frick Laboratory, Department of Chemistry Princeton University Princeton, NJ 08544 phone: 609-258-1834 fax: 609-258-6746 silviu at Princeton.EDU %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% -------------- next part -------------- A non-text attachment was scrubbed... Name: silviu.vcf Type: text/x-vcard Size: 272 bytes Desc: not available Url : /pipermail/attachments/20050926/9c37ea64/attachment.vcf From massimiliano.bonomi at mi.infn.it Mon Sep 26 11:43:14 2005 From: massimiliano.bonomi at mi.infn.it (Massimiliano Bonomi) Date: Mon, 26 Sep 2005 11:43:14 +0200 Subject: [Pw_forum] lambda convergence Message-ID: <4337C2B2.50908@mi.infn.it> Dear pwscf users, I'm trying to calculate electron-phonon coupling constant at Gamma point for a solid assembled from clusters. I have used different mesh of k-points (800, 1600, 2200) and different value for the gaussian smearing. Plotting the DOS and the value for lambda I noticed a strange thing: the density of states is convergent for smearing in the range 0.02-0.03 while suddenly goes up for smearings larger than 0.05. Lambda is nearly constant for smearing > 0.06 and increases before. Should they converge in the same range of gaussian smearing? What is happening?? Thanks in advance. Massimiliano Bonomi -- Massimiliano Bonomi Dipartimento di Fisica, Universit? degli Studi di Milano via Celoria, 16 - 20133 Milano, Italy ph. +39 02 50317 717 e-mail: massimiliano.bonomi at mi.infn.it, massimiliano.bonomi at unimi.it From alberto.milani at chem.polimi.it Mon Sep 26 15:33:22 2005 From: alberto.milani at chem.polimi.it (Alberto Milani) Date: Mon, 26 Sep 2005 15:33:22 +0200 Subject: [Pw_forum] WARNING: something is going wrong Message-ID: <200509261533.22676.alberto.milani@chem.polimi.it> Dear all, I'm a new PWscf user and I have a question about a relax calculation. In the output I have found the following message: WARNING : something is going wrong resetting bfgs history new conv_thr = 0.0000000100 Then the calculation goes on but the number of bfgs steps is always the same till the end of the calculation and this looks strange to me (a larger part of the output follows at the end of this mail). This warning message occurs a different number of times in the output but the calculation seems to finish correctly. What does it exactly mean ? I have noticed that also another user had the same problem, but I found no answer in the forum. Thanks in advance The input for the calculation was: &control calculation='relax' title='Opt geometria cella fissa per $ verbosity = 'high', tstress = .TRUE., tprnfor = .TRUE., nstep = 1024, etot_conv_thr = 1.0D-6, forc_conv_thr = 1.0D-6, restart_mode='from_scratch', prefix='$nome', pseudo_dir = './pseudo/', outdir='./out/' / &system ibrav= 6, a = 10.0, b = 10.0, c = 2.5522, cosab = 0.0, cosac = 0.0, cosbc = 0.0, nat = 2, ntyp = 1, ecutwfc = 35, ecutrho = 350, nosym = .FALSE. , occupations ='tetrahedra', / &electrons conv_thr = 1.0d-6, mixing_mode = 'plain', mixing_beta = 0.7, electron_maxstep = 512, diagonalization = 'david', / &ions upscale = 100.0, ion_dynamics = 'bfgs', lbfgs_ndim = 1, / ATOMIC_SPECIES C 12.00 C.pbe-van_ak.UPF ATOMIC_POSITIONS {angstrom} C 0.0 0.0 -0.60022 C 0.0 0.0 0.60022 K_POINTS {automatic} 1 1 70 0 0 0 Output: number of scf cycles = 163 number of bfgs steps = 9 energy old = -22.6453264058 ryd energy new = -22.6453264054 ryd CASE: energy_new > energy_old new trust radius = 0.0000008222 bohr WARNING : something is going wrong resetting bfgs history new conv_thr = 0.0000000100 ATOMIC_POSITIONS (angstrom) C 0.000000000 0.000000000 -0.632152788 C 0.000000000 0.000000000 0.632152788 NEW-OLD atomic charge density approx. for the potential total cpu time spent up to now is 31761.70 secs Self-consistent Calculation iteration # 1 ecut= 35.00 ryd beta=0.70 Davidson diagonalization (with overlap) ethr = 1.00E-05, avg # of iterations = 1.0 Threshold (ethr) on eigenvalues was too large: Diagonalizing with lowered threshold Davidson diagonalization (with overlap) ethr = 2.65E-10, avg # of iterations = 1.0 total cpu time spent up to now is 31875.66 secs End of self-consistent calculation k = 0.0000 0.0000 0.0000 ( 6061 PWs) bands (ev): -20.6810 -13.4651 -9.7324 -9.7324 -0.2654 1.2893 1.2893 1.5008 k = 0.0000 0.0000 0.0557 ( 6077 PWs) bands (ev): -20.6784 -13.4708 -9.7280 -9.7280 -0.2608 1.2939 1.2939 1.5053 k = 0.0000 0.0000 0.1113 ( 6077 PWs) bands (ev): -20.6705 -13.4876 -9.7149 -9.7149 -0.2473 1.3078 1.3078 1.5187 k = 0.0000 0.0000 0.1670 ( 6077 PWs) bands (ev): -20.6573 -13.5155 -9.6929 -9.6929 -0.2247 1.3310 1.3310 1.5410 k = 0.0000 0.0000 0.2226 ( 6093 PWs) bands (ev): -20.6389 -13.5543 -9.6621 -9.6621 -0.1931 1.3635 1.3635 1.5723 k = 0.0000 0.0000 0.2783 ( 6105 PWs) bands (ev): -20.6152 -13.6038 -9.6226 -9.6226 -0.1524 1.4052 1.4052 1.6125 k = 0.0000 0.0000 0.3339 ( 6093 PWs) bands (ev): -20.5862 -13.6635 -9.5742 -9.5742 -0.1028 1.4562 1.4562 1.6616 k = 0.0000 0.0000 0.3896 ( 6101 PWs) bands (ev): -20.5521 -13.7332 -9.5171 -9.5171 -0.0441 1.5165 1.5165 1.7197 k = 0.0000 0.0000 0.4453 ( 6069 PWs) bands (ev): -20.5128 -13.8124 -9.4512 -9.4512 0.0236 1.5803 1.5861 1.7866 k = 0.0000 0.0000 0.5009 ( 6073 PWs) bands (ev): -20.4684 -13.9006 -9.3766 -9.3766 0.1002 1.6582 1.6649 1.8624 k = 0.0000 0.0000 0.5566 ( 6081 PWs) bands (ev): -20.4188 -13.9975 -9.2932 -9.2932 0.1859 1.7492 1.7530 1.9471 k = 0.0000 0.0000 0.6122 ( 6105 PWs) bands (ev): -20.3642 -14.1024 -9.2011 -9.2011 0.2805 1.8503 1.8503 2.0407 k = 0.0000 0.0000 0.6679 ( 6097 PWs) bands (ev): -20.3045 -14.2147 -9.1002 -9.1002 0.3840 1.9568 1.9569 2.1430 k = 0.0000 0.0000 0.7236 ( 6085 PWs) bands (ev): -20.2398 -14.3340 -8.9906 -8.9906 0.4965 2.0727 2.0727 2.2542 k = 0.0000 0.0000 0.7792 ( 6073 PWs) bands (ev): -20.1702 -14.4597 -8.8724 -8.8724 0.6178 2.1975 2.1975 2.3742 k = 0.0000 0.0000 0.8349 ( 6069 PWs) bands (ev): -20.0957 -14.5912 -8.7454 -8.7454 0.7481 2.3308 2.3308 2.5029 k = 0.0000 0.0000 0.8905 ( 6089 PWs) bands (ev): -20.0164 -14.7280 -8.6098 -8.6098 0.8871 2.3451 2.3451 2.4831 k = 0.0000 0.0000 0.9462 ( 6081 PWs) bands (ev): -19.9324 -14.8695 -8.4656 -8.4656 1.0349 1.9589 1.9589 2.6309 k = 0.0000 0.0000 1.0018 ( 6081 PWs) bands (ev): -19.8437 -15.0153 -8.3128 -8.3128 1.1915 1.5683 1.5683 2.7926 k = 0.0000 0.0000 1.0575 ( 6073 PWs) bands (ev): -19.7503 -15.1647 -8.1513 -8.1513 1.1786 1.1786 1.3568 2.9639 k = 0.0000 0.0000 1.1132 ( 6065 PWs) bands (ev): -19.6525 -15.3172 -7.9813 -7.9813 0.7911 0.7911 1.5307 3.1445 k = 0.0000 0.0000 1.1688 ( 6077 PWs) bands (ev): -19.5502 -15.4726 -7.8028 -7.8028 0.4063 0.4063 1.7131 3.3344 k = 0.0000 0.0000 1.2245 ( 6085 PWs) bands (ev): -19.4436 -15.6301 -7.6159 -7.6159 0.0251 0.0251 1.9041 3.6450 k = 0.0000 0.0000 1.2801 ( 6069 PWs) bands (ev): -19.3328 -15.7894 -7.4204 -7.4204 -0.3517 -0.3517 2.1033 3.8418 k = 0.0000 0.0000 1.3358 ( 6057 PWs) bands (ev): -19.2178 -15.9501 -7.2166 -7.2166 -0.7239 -0.7239 2.3108 4.0467 k = 0.0000 0.0000 1.3914 ( 6021 PWs) bands (ev): -19.0989 -16.1118 -7.0044 -7.0044 -1.0908 -1.0908 2.5264 4.2595 k = 0.0000 0.0000 1.4471 ( 6005 PWs) bands (ev): -18.9761 -16.2740 -6.7840 -6.7840 -1.4522 -1.4522 2.7498 4.4801 k = 0.0000 0.0000 1.5028 ( 6001 PWs) bands (ev): -18.8495 -16.4365 -6.5555 -6.5555 -1.8077 -1.8077 2.9810 4.7083 k = 0.0000 0.0000 1.5584 ( 6001 PWs) bands (ev): -18.7195 -16.5987 -6.3187 -6.3187 -2.1569 -2.1569 3.2194 4.9438 k = 0.0000 0.0000 1.6141 ( 5989 PWs) bands (ev): -18.5860 -16.7604 -6.0741 -6.0741 -2.4994 -2.4994 3.4649 5.1861 k = 0.0000 0.0000 1.6697 ( 5989 PWs) bands (ev): -18.4494 -16.9212 -5.8218 -5.8218 -2.8348 -2.8348 3.7168 5.4348 k = 0.0000 0.0000 1.7254 ( 5974 PWs) bands (ev): -18.3100 -17.0806 -5.5622 -5.5622 -3.1626 -3.1626 3.9747 5.6877 k = 0.0000 0.0000 1.7811 ( 5970 PWs) bands (ev): -18.1683 -17.2379 -5.2963 -5.2963 -3.4817 -3.4817 4.2377 5.8284 k = 0.0000 0.0000 1.8367 ( 5990 PWs) bands (ev): -18.0257 -17.3919 -5.0261 -5.0261 -3.7899 -3.7899 4.5046 5.5847 k = 0.0000 0.0000 1.8924 ( 5986 PWs) bands (ev): -17.8870 -17.5372 -4.7617 -4.7617 -4.0772 -4.0772 4.7731 5.3218 k = 0.0000 0.0000-1.9480 ( 6010 PWs) bands (ev): -17.7985 -17.6281 -4.5929 -4.5929 -4.2536 -4.2536 5.0037 5.0927 the Fermi energy is -4.5268 ev ! total energy = -22.64532640 ryd estimated scf accuracy < 2.2E-09 ryd band energy sum = -7.45594375 ryd one-electron contribution = -66.16342565 ryd hartree contribution = 33.77201522 ryd xc contribution = -6.79710281 ryd ewald contribution = 16.54318684 ryd convergence has been achieved Forces acting on atoms (Ry/au): atom 1 type 1 force = 0.00000000 0.00000000 -0.00000007 atom 2 type 1 force = 0.00000000 0.00000000 0.00000007 Total force = 0.000000 Total SCF correction = 0.000073 entering subroutine stress ... total stress (ryd/bohr**3) (kbar) P= -0.13 -0.00000076 -0.00000000 0.00000000 -0.11 -0.00 0.00 -0.00000000 -0.00000076 0.00000000 -0.00 -0.11 0.00 0.00000000 0.00000000 -0.00000108 0.00 0.00 -0.16 kinetic stress (kbar) 845.48 0.00 0.00 0.00 845.48 0.00 0.00 0.00 574.07 local stress (kbar) 206.44 0.00 -0.00 0.00 206.44 -0.00 -0.00 -0.00 -7202.63 nonloc. stress (kbar) 251.31 0.00 0.00 0.00 251.31 0.00 0.00 0.00 288.40 hartree stress (kbar) 0.57 -0.00 0.00 -0.00 0.57 -0.00 0.00 -0.00 2867.07 exc-cor stress (kbar) -183.30 -0.00 0.00 -0.00 -183.30 -0.00 0.00 -0.00 -173.32 corecor stress (kbar) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 ewald stress (kbar) -1120.63 -0.00 0.00 -0.00 -1120.63 0.00 0.00 0.00 3646.25 hubbard stress (kbar) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 bfgs converged in 164 scf cycles and 9 bfgs steps End of BFGS geometry calculation Final energy = -22.6453264050 ryd Saving the approximate inverse hessian CELL_PARAMETERS (alat) 1.000000000 0.000000000 0.000000000 0.000000000 1.000000000 0.000000000 0.000000000 0.000000000 0.256670000 ATOMIC_POSITIONS (angstrom) C 0.000000000 0.000000000 -0.632152788 C 0.000000000 0.000000000 0.632152788 Writing file cc.save for program phonon From asen at iitk.ac.in Mon Sep 26 18:32:08 2005 From: asen at iitk.ac.in (A Sen) Date: Mon, 26 Sep 2005 22:02:08 +0530 (IST) Subject: [Pw_forum] CPV compilation error Message-ID: <57978.172.26.113.66.1127752328.squirrel@webmail.iitk.ac.in> Dear All, Although I succeeded in compiling PW, PP and FPMD with Intel 9.0 compilers and mkl8.0 libraries in an Itanium workstation, I couldn't do so for CPV as I encountered the following errors: --------------------------------- cplib.o(.text+0x42f2): In function `dforce_': : undefined reference to `cfft3s_' cplib.o(.text+0x4c42): In function `dforce_': : undefined reference to `cfft3s_' cplib.o(.text+0x6322): In function `fwfftw_': : undefined reference to `cfft3s_' cplib.o(.text+0x11012): In function `newd_': : undefined reference to `cfft3b_' cplib.o(.text+0x123c2): In function `newd_': : undefined reference to `cfft3b_' cplib.o(.text+0x141e2): In function `newd_': : undefined reference to `cfft3b_' cplib.o(.text+0x4d1e2): In function `ivfftw_': : undefined reference to `cfft3s_' cplib.o(.text+0x4d2a2): In function `fwffts_': : undefined reference to `cfft3s_' cplib.o(.text+0x59922): In function `ivfftb_': : undefined reference to `cfft3b_' cplib.o(.text+0x61f22): In function `fwfft_': : undefined reference to `cfft3_' cplib.o(.text+0x62a52): In function `set_cc_': : undefined reference to `cfft3b_' cplib.o(.text+0x70662): In function `ivffts_': : undefined reference to `cfft3s_' cplib.o(.text+0x750a2): In function `add_cc_': : undefined reference to `cfft3_' cplib.o(.text+0x76322): In function `force_cc_': : undefined reference to `cfft3b_' cplib.o(.text+0xa1032): In function `fillgrad_': : undefined reference to `cfft3_' cplib.o(.text+0xa1912): In function `fillgrad_': : undefined reference to `cfft3_' cplib.o(.text+0xa25d2): In function `grad2_': : undefined reference to `cfft3_' cplib.o(.text+0xa2d02): In function `grad2_': : undefined reference to `cfft3_' cplib.o(.text+0xa3732): In function `grad2_': : undefined reference to `cfft3_' cplib.o(.text+0xb4062): more undefined references to `cfft3_' follow wf.o(.text+0x1cc62): In function `ivfftbold_': : undefined reference to `cfft3b_' make[1]: *** [cp.x] Error 1 make[1]: Leaving directory `/extra/soft/espresso-2.1.5/CPV' make: *** [cp] Error 2 ------------------------------------ Can you please guide me in resolving this? I am not able to find out cfft3_,cfft3s_ and cfft3b_ subroutines, which are to be called by the program cplib.f90 within the CPV directory. Regards, A. Sen From silviu at Princeton.EDU Tue Sep 27 15:15:05 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Tue, 27 Sep 2005 15:15:05 +0200 Subject: [Pw_forum] g95 compilation problem In-Reply-To: <20050927110733.39CC911280D@democritos.sissa.it> References: <20050927110733.39CC911280D@democritos.sissa.it> Message-ID: <433945D9.7080102@Princeton.EDU> Hi, Did any of you experienced the following (or similar) compilation problem with g95: g95 -O -D__FFTW -I../include -I. -I../Modules -I../PW -I../PH -I../iotk/src -c fft_scalar.F90 -o fft_scalar.o In file fft_scalar.F90:78 C_POINTER :: fw_plan( 3, ndims ) = 0 1 Error: Unclassifiable statement at (1) In file fft_scalar.F90:79 C_POINTER :: bw_plan( 3, ndims ) = 0 1 Error: Unclassifiable statement at (1) .... I am trying to get the cvs version running on a Cygwin + g95 platform. Thanks, Silviu. -------------- next part -------------- A non-text attachment was scrubbed... Name: silviu.vcf Type: text/x-vcard Size: 272 bytes Desc: not available Url : /pipermail/attachments/20050927/9dfb6304/attachment.vcf From ferretti.andrea at unimore.it Tue Sep 27 14:36:08 2005 From: ferretti.andrea at unimore.it (Andrea Ferretti) Date: Tue, 27 Sep 2005 14:36:08 +0200 (CEST) Subject: [Pw_forum] g95 compilation problem In-Reply-To: <433945D9.7080102@Princeton.EDU> References: <20050927110733.39CC911280D@democritos.sissa.it> <433945D9.7080102@Princeton.EDU> Message-ID: Hi Silviu, as far as I understand C_POINTER is a preprocessor macro which is defined in f_defs.h, as stems from (fft_scalar.f90, lines 72-75) 72 ! C_POINTER is defined in include/f_defs.h 73 ! for 32bit executables, C_POINTER is integer(4) 74 ! for 64bit executables, C_POINTER is integer(8) therefore it seems there's something wrong with the gcc precompiler producing fft_scalar.F90 ... while C_POINTER is explicitly written in the .f90 file, it should be expanded to the macro value ( integer(kind=4 or 8) ) in the .F90 file, which seems not to be the case here... Andrea On Tue, 27 Sep 2005, Silviu Zilberman wrote: > Hi, > > Did any of you experienced the following (or similar) compilation > problem with g95: > > g95 -O -D__FFTW -I../include -I. -I../Modules -I../PW -I../PH > -I../iotk/src -c fft_scalar.F90 -o fft_scalar.o > In file fft_scalar.F90:78 > > C_POINTER :: fw_plan( 3, ndims ) = 0 > 1 > Error: Unclassifiable statement at (1) > In file fft_scalar.F90:79 > > C_POINTER :: bw_plan( 3, ndims ) = 0 > 1 > Error: Unclassifiable statement at (1) > .... > > I am trying to get the cvs version running on a Cygwin + g95 platform. > > Thanks, Silviu. > > > From eyvaz_isaev at yahoo.com Tue Sep 27 21:57:43 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Tue, 27 Sep 2005 12:57:43 -0700 (PDT) Subject: [Pw_forum] CPV compilation error In-Reply-To: <57978.172.26.113.66.1127752328.squirrel@webmail.iitk.ac.in> Message-ID: <20050927195743.21621.qmail@web60320.mail.yahoo.com> Hi, You need to install FFTW3 library. Presumably, there is some effort to change from fftw2.1 to fftw3. As far as I know last releases of SuSe Linux contain precomiled fftw3 libraries. For sources see www.fftw.org, download sources, compile, and enjoy! Bests, Eyvaz. --- A Sen wrote: > Dear All, > Although I succeeded in compiling PW, PP > and FPMD with Intel 9.0 > compilers and mkl8.0 libraries in an Itanium > workstation, I > couldn't do so for CPV as I encountered the > following errors: > --------------------------------- > cplib.o(.text+0x42f2): In function `dforce_': > : undefined reference to `cfft3s_' > cplib.o(.text+0x4c42): In function `dforce_': > : undefined reference to `cfft3s_' > cplib.o(.text+0x6322): In function `fwfftw_': > : undefined reference to `cfft3s_' > cplib.o(.text+0x11012): In function `newd_': > : undefined reference to `cfft3b_' > cplib.o(.text+0x123c2): In function `newd_': > : undefined reference to `cfft3b_' > cplib.o(.text+0x141e2): In function `newd_': > : undefined reference to `cfft3b_' > cplib.o(.text+0x4d1e2): In function `ivfftw_': > : undefined reference to `cfft3s_' > cplib.o(.text+0x4d2a2): In function `fwffts_': > : undefined reference to `cfft3s_' > cplib.o(.text+0x59922): In function `ivfftb_': > : undefined reference to `cfft3b_' > cplib.o(.text+0x61f22): In function `fwfft_': > : undefined reference to `cfft3_' > cplib.o(.text+0x62a52): In function `set_cc_': > : undefined reference to `cfft3b_' > cplib.o(.text+0x70662): In function `ivffts_': > : undefined reference to `cfft3s_' > cplib.o(.text+0x750a2): In function `add_cc_': > : undefined reference to `cfft3_' > cplib.o(.text+0x76322): In function `force_cc_': > : undefined reference to `cfft3b_' > cplib.o(.text+0xa1032): In function `fillgrad_': > : undefined reference to `cfft3_' > cplib.o(.text+0xa1912): In function `fillgrad_': > : undefined reference to `cfft3_' > cplib.o(.text+0xa25d2): In function `grad2_': > : undefined reference to `cfft3_' > cplib.o(.text+0xa2d02): In function `grad2_': > : undefined reference to `cfft3_' > cplib.o(.text+0xa3732): In function `grad2_': > : undefined reference to `cfft3_' > cplib.o(.text+0xb4062): more undefined references to > `cfft3_' follow > wf.o(.text+0x1cc62): In function `ivfftbold_': > : undefined reference to `cfft3b_' > make[1]: *** [cp.x] Error 1 > make[1]: Leaving directory > `/extra/soft/espresso-2.1.5/CPV' > make: *** [cp] Error 2 > ------------------------------------ > Can you please guide me in resolving this? I am not > able to find out > cfft3_,cfft3s_ and cfft3b_ subroutines, which are to > be called by the > program cplib.f90 within the CPV directory. > > Regards, > A. Sen > > > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From naromero at gmail.com Tue Sep 27 23:19:11 2005 From: naromero at gmail.com (Nichols A. Romero) Date: Tue, 27 Sep 2005 17:19:11 -0400 Subject: [Pw_forum] definition of parameter conv_thr Message-ID: <6ac064b605092714195bf3c7eb@mail.gmail.com> Hi, Where can I find the definition of the conv_thr parameter? Is this a tolerance on the the wavefunction squared or potential residual ( i.e. analagous to tolwfr or tolvrs in ABINIT)? I searched the pw_forum archive but could not find a useful reference. Thanks, -- Nichols A. Romero, PhD. 1613 Denise Dr. Apt. D Forest Hill, MD 21050 (217) 417-5210 -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050927/59231774/attachment.htm From eyvaz_isaev at yahoo.com Wed Sep 28 00:01:51 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Tue, 27 Sep 2005 15:01:51 -0700 (PDT) Subject: [Pw_forum] definition of parameter conv_thr In-Reply-To: <6ac064b605092714195bf3c7eb@mail.gmail.com> Message-ID: <20050927220152.19882.qmail@web60318.mail.yahoo.com> Hi, Obviously, it is total energy treshold. Please have a look at /Doc/INPUT_PW in your Espresso directory. As a sequence, it is not the same as tolwfr and tolvrs in ABINIT. Bests, Eyvaz. --- "Nichols A. Romero" wrote: > Hi, > Where can I find the definition of the conv_thr > parameter? > Is this a tolerance on the the wavefunction squared > or potential residual ( > i.e. analagous to tolwfr or tolvrs in ABINIT)? > I searched the pw_forum archive but could not find > a useful reference. > Thanks, > -- > Nichols A. Romero, PhD. > 1613 Denise Dr. Apt. D > Forest Hill, MD 21050 > (217) 417-5210 > __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com From leoant21 at hotmail.com Wed Sep 28 02:56:47 2005 From: leoant21 at hotmail.com (=?gb2312?B?1cUguumx8g==?=) Date: Wed, 28 Sep 2005 00:56:47 +0000 Subject: [Pw_forum] Question on supercell calculation of MgB2 In-Reply-To: <20050924053636.1536.67609.Mailman@democritos.sissa.it> Message-ID: Dear all: Now I am doing a 2*2 supercell calculation of MgB2, however, there is a warning like this: Warning: ef = 1.818386 is above the highest band at k-point 490 e = 1.780277 and there are also warning on other k-points, but not all. Please tell me why. Thank you. Regards. _________________________________________________________________ ???? MSN Explorer: http://explorer.msn.com/lccn From eyvaz_isaev at yahoo.com Wed Sep 28 08:51:47 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Tue, 27 Sep 2005 23:51:47 -0700 (PDT) Subject: [Pw_forum] Question on supercell calculation of MgB2 In-Reply-To: Message-ID: <20050928065147.27365.qmail@web60323.mail.yahoo.com> Hi, This message was discussed before in the forum. It means unsufficient bands number which is defined as Number_of_valence_electrons*(1/2 + 20%). Just increase bands number (nbnd). Bests, Eyvaz. --- ?? ???? wrote: > Dear all: > Now I am doing a 2*2 supercell calculation of > MgB2, however, there is a > warning like this: > > Warning: ef = 1.818386 is above the highest band at > k-point 490 > e = 1.780277 > > and there are also warning on other k-points, but > not all. Please tell me > why. Thank you. > > Regards. > > _________________________________________________________________ > ???????? MSN Explorer: > http://explorer.msn.com/lccn > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From asen at iitk.ac.in Wed Sep 28 12:44:20 2005 From: asen at iitk.ac.in (A Sen) Date: Wed, 28 Sep 2005 16:14:20 +0530 (IST) Subject: [Pw_forum] CPV compilation error (Contd) Message-ID: <46174.172.28.37.46.1127904260.squirrel@webmail.iitk.ac.in> Dear Eyvaz, Thank you very much for your cooperation. I have installed both fftw-2.1.5 and fftw-3.0.1, but still I am coming across the same errors. Regarding fftw-3.0.1, I did not find any lib directory as such, but only libfftw3.la. Can you please suggest me what to do next if I have done something wrong in the installation of fftw-3.0.1. What efforts (and how) should I make in changing from fftw-2.1.5 to fftw-3.0.1? Please suggest me as to how I can resolve finally the PWscf compilation issue. Here is the make.sys.in file: --------------------- OSHOME = /extra/soft/espresso-2.1.5 DFLAGS = -D__LINUX64 -D__INTEL IFLAGS = -I$(OSHOME)/include -I./ CPPFLAGS = $(DFLAGS) $(IFLAGS) -no-ipo F90 = ifort F77 = ifort FFLAGS = -w -O2 -no-ipo F90FLAGS = $(FFLAGS) -fpp $(CPPFLAGS) -FR -no-ipo F77FLAGS = $(FFLAGS) -no-ipo F77FLAGS_NOOPT = -O0 CC = icc -c99 CCFLAGS = $(CPPFLAGS) -no-ipo MODULEFLAG= -I$(OSHOME)/Modules -I$(OSHOME)/PW -I$(OSHOME)/PH -I$(OSHOME)/PWCOND -I$(OSHOME)/PWNC -I$(OSHOME)/D3 -I$(OSHOME)/Gamma FFTW_LIB = -L/usr/local/lib -lfftw3 -lfftw -openmp LIBS = $(FFTW_LIB) -L/opt/intel/mkl80/lib/64 -lmkl_blas95 -lmkl_lapack95 -lmkl_lapack -lmkl_ipf -lguide -lpthread LD=$(F90) LDFLAGS = $(OSHOME)/flib/ptools.a $(OSHOME)/flib/flib.a $(OSHOME)/clib/clib.a -Vaxlib -no-ipo $(LIBS) AR = ar ARFLAGS = ruv RANLIB = echo -------------------------------- Thank you once again. Regards, A. Sen From silviu at Princeton.EDU Wed Sep 28 15:33:51 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Wed, 28 Sep 2005 15:33:51 +0200 Subject: [Pw_forum] g95 compilation problem In-Reply-To: References: <20050927110733.39CC911280D@democritos.sissa.it> <433945D9.7080102@Princeton.EDU> Message-ID: <433A9BBF.6070003@Princeton.EDU> Dear Andrea, Thank you very much for your reply, indeed I am experiencing strange behavior at the pre-processing stage and this problem is probably related. Is there someone out there that have a working code on cygwin/g95? I believe I got over the post-processing problem, but new ones keep coming. Thanks, Silviu. Andrea Ferretti wrote: >Hi Silviu, >as far as I understand C_POINTER is a preprocessor macro which is defined >in f_defs.h, as stems from (fft_scalar.f90, lines 72-75) > > 72 ! C_POINTER is defined in include/f_defs.h > 73 ! for 32bit executables, C_POINTER is integer(4) > 74 ! for 64bit executables, C_POINTER is integer(8) > >therefore it seems there's something wrong with the gcc precompiler >producing fft_scalar.F90 ... >while C_POINTER is explicitly written in the .f90 file, it should be >expanded to the macro value ( integer(kind=4 or 8) ) in the .F90 file, >which seems not to be the case here... > >Andrea > > >On Tue, 27 Sep 2005, Silviu Zilberman wrote: > > > >>Hi, >> >>Did any of you experienced the following (or similar) compilation >>problem with g95: >> >>g95 -O -D__FFTW -I../include -I. -I../Modules -I../PW -I../PH >>-I../iotk/src -c fft_scalar.F90 -o fft_scalar.o >>In file fft_scalar.F90:78 >> >> C_POINTER :: fw_plan( 3, ndims ) = 0 >> 1 >>Error: Unclassifiable statement at (1) >>In file fft_scalar.F90:79 >> >> C_POINTER :: bw_plan( 3, ndims ) = 0 >> 1 >>Error: Unclassifiable statement at (1) >>.... >> >>I am trying to get the cvs version running on a Cygwin + g95 platform. >> >>Thanks, Silviu. >> >> >> >> >> >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > -------------- next part -------------- A non-text attachment was scrubbed... Name: silviu.vcf Type: text/x-vcard Size: 247 bytes Desc: not available Url : /pipermail/attachments/20050928/bc7456f9/attachment.vcf From g.ballabio at cineca.it Wed Sep 28 15:01:54 2005 From: g.ballabio at cineca.it (Gerardo Ballabio) Date: Wed, 28 Sep 2005 15:01:54 +0200 (MEST) Subject: [Pw_forum] CPV compilation error (Contd) In-Reply-To: <46174.172.28.37.46.1127904260.squirrel@webmail.iitk.ac.in> (from asen@iitk.ac.in on Wed Sep 28 12:44:20 2005) References: <46174.172.28.37.46.1127904260.squirrel@webmail.iitk.ac.in> Message-ID: <1127912512l.21152l.0l@nb-ballabio.cineca.it> On 09/28/2005 12:44:20 PM, A Sen wrote: > Dear Eyvaz, > Thank you very much for your cooperation. I have > installed > both fftw-2.1.5 and fftw-3.0.1, but still I am coming across > the same errors. Regarding fftw-3.0.1, I did not find any lib > directory as such, but only libfftw3.la. If you installed fftw from source *and* did a "make install" as root after compilation, the generated libraries must have been placed in /usr/local/lib (if I remember correctly) unless you had specified a different place with the "--prefix" option to "configure". By the way Espresso didn't support fftw version 3 last time I checked, you have to use version 2. Anyway the above is valid for both versions. Gerardo From degironc at sissa.it Wed Sep 28 17:44:54 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Wed, 28 Sep 2005 17:44:54 +0200 Subject: [Pw_forum] CPV compilation error (Contd) In-Reply-To: <1127912512l.21152l.0l@nb-ballabio.cineca.it> References: <46174.172.28.37.46.1127904260.squirrel@webmail.iitk.ac.in> <1127912512l.21152l.0l@nb-ballabio.cineca.it> Message-ID: <1127922294.433aba7684598@webmail.sissa.it> Paolo Giannozzi asked to me to forward this message to the list stefano ----- On Mon, 26 Sep 2005 22:02:08 +0530 (IST) "A Sen" wrote: >[...] for CPV [...] I encountered the following errors: >[...] >: undefined reference to `cfft3s_' >[...] >: undefined reference to `cfft3b_' >[...] >: undefined reference to `cfft3_' all of these are contained in CPV/fft.f90 . Please check that CPV/fft.f90 is correctly compiled, i.e. that CPV/fft.o exists, and that it contains the above symbols ("nm CPV/fft.o | grep T | grep cfft3" for instance). Paolo ---------------------------------------------------------------- SISSA Webmail https://webmail.sissa.it/ Powered by Horde http://www.horde.org/ From degironc at sissa.it Wed Sep 28 17:48:07 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Wed, 28 Sep 2005 17:48:07 +0200 Subject: [Pw_forum] g95 compilation problem In-Reply-To: <433A9BBF.6070003@Princeton.EDU> References: <20050927110733.39CC911280D@democritos.sissa.it> <433945D9.7080102@Princeton.EDU> <433A9BBF.6070003@Princeton.EDU> Message-ID: <1127922487.433abb37a1aaa@webmail.sissa.it> Again a message from Paolo On Tue, 27 Sep 2005 15:15:05 +0200 Silviu Zilberman wrote: >g95 -O -D__FFTW -I../include -I. -I../Modules -I../PW >-I../PH -I../iotk/src -c fft_scalar.F90 -o fft_scalar.o >In file fft_scalar.F90:78 > > C_POINTER :: fw_plan( 3, ndims ) = 0 > 1 >Error: Unclassifiable statement at (1) preprocessing flags are not sufficient: I think you should have at least -D__LINUX -D__G95 -D__FFTW . C_POINTER has to be preprocessed to something sensible. ---------------------------------------------------------------- SISSA Webmail https://webmail.sissa.it/ Powered by Horde http://www.horde.org/ From degironc at sissa.it Wed Sep 28 17:51:03 2005 From: degironc at sissa.it (Stefano de Gironcoli) Date: Wed, 28 Sep 2005 17:51:03 +0200 Subject: [Pw_forum] definition of parameter conv_thr In-Reply-To: <6ac064b605092714195bf3c7eb@mail.gmail.com> References: <6ac064b605092714195bf3c7eb@mail.gmail.com> Message-ID: <1127922663.433abbe725222@webmail.sissa.it> On Tue, 27 Sep 2005 17:19:11 -0400 "Nichols A. Romero" wrote: > Where can I find the definition of the conv_thr >parameter? as specified in Doc/INPUT_PW: conv_thr is an estimate of the error on the energy due to the self-consistency error ---------------------------------------------------------------- SISSA Webmail https://webmail.sissa.it/ Powered by Horde http://www.horde.org/ From konstantin_kudin at yahoo.com Wed Sep 28 18:34:48 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Wed, 28 Sep 2005 09:34:48 -0700 (PDT) Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <1127385715.3000.8.camel@m00> Message-ID: <20050928163448.9694.qmail@web52012.mail.yahoo.com> Thanks to Axel and Alexander for suggestions on this issue! Some of them were quite helpful. I have investigated things further, in a more careful way. What happens is that the GOTO library uses its own threads, and that causes the CP90 time to be underestimated while the wall time stays roughly the same. This also makes 1cpu times look small, since the 2nd GOTO thread is quietly computing whenever the 2nd cpu is available. I used my own test case, with water and some organic stuff in it. The run does 20 CP steps starting from a restart file on a local disk. I eliminated the nfs issues by writing the stuff directly to the local disk on the head node. Also, for runs with 1 thread the 2nd cpu on the dual node was kept idle. For mpich1 it seems that there is little difference in wall times whether shared memory is used or not. However, CP90 times appear much smaller for sockets. For mpich2 shared memory just hangs, while sockets work. Also, mpich2 is worse in bad situations (more than 4 cpus) than mpich1. The launcher was compiled accordingly mpiexec from Ohio SC in all cases. Below are some actual numbers. 1 or 2 at the end of the job name means that the GOTO_NUM_THREADS was either 1 or 2. It looks that 2 dual nodes with 1Gbit is the fastest, and the speedup is about 3x. Beyond that things go downhill. Kostya ######################### mpich1 with shared memory job name G N_thread Wall CP90 sh1.out.0.1 1 106.5m 1h46m sh1.out.0.2 1 85.8m 1h24m sh1.out.1.1 2 62.6m 1h sh1.out.1.2 2 59.4m 48m sh1.out.2.1 4 34.8m 33m51.78s sh1.out.2.2 4 34.6m 27m46.72s sh1.out.3.1 6 42.7m 37m29.63s so1.out.3.2 (hangs) sh1.out.4.1 8 45.3m 37m43.38s sh1.out.4.2 8 45.6m 35m31.07s mpich1 with sockets job name G N_thread Wall CP90 so1.out.0.1 1 108.1m 1h47m so1.out.0.2 1 84.6m 1h23m so1.out.1.1 2 61.7m 1h so1.out.1.2 2 62.9m 50m32.96s so1.out.2.1 4 35.6m 33m10.21s so1.out.2.2 4 35.6m 27m20.13s so1.out.3.1 6 42.6m 22m59.73s so1.out.3.2 6 42.5m 18m50.57s so1.out.4.1 8 45.1m 18m31.35s so1.out.4.2 8 43.4m 14m47.97s mpich2 with sockets so2.out.0.1 1 105.3m 1h45m so2.out.0.2 1 84.5m 1h23m so2.out.1.1 2 60.7m 1h so2.out.1.2 2 59.4m 47m31.77s so2.out.2.1 4 34.6m 32m26.08s so2.out.2.2 4 34.3m 26m10.19s so2.out.3.1 6 53.5m 23m so2.out.3.2 6 53.0m 18m so2.out.4.1 8 54.0m 17m56.94s so2.out.4.2 8 53.2m 14m44.32s __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From marzari at MIT.EDU Wed Sep 28 19:15:20 2005 From: marzari at MIT.EDU (Nicola Marzari) Date: Wed, 28 Sep 2005 13:15:20 -0400 Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <20050928163448.9694.qmail@web52012.mail.yahoo.com> References: <20050928163448.9694.qmail@web52012.mail.yahoo.com> Message-ID: <433ACFA8.1000703@mit.edu> Dear Kostya, a long time ago I had collated some of our tests on parallel performance - parallelism over gigabit ethernet used to go very well up to 4 nodes, and then flatten out. All the details are in here: http://nnn.mit.edu/ESPRESSO/CP90_tests/ The "large" system has ~500 bands, so it's a reasonably challenging task. All the timings are always wall clock time. We have more recent tests in the group (maybe someone could post them) on Dual Xeon 3.4 Ghz, and 800 Mhz FSB. Our detailed experience there, with a lot of help from Intel, was that the second CPU, living on the same FSB, provides at best a 25%-30% performance boost. Opterons achieve on the other hand excellent scaling on the two CPUs on the same motherboard, but have worse performance at the single CPU level (I believe thanks to the extensive use of MKL by the CP code). My own feeling as the best platform for CP is nodes of 4 or 6 PIVs, with the fastest memory and FSB available. Gigabit will do fine for this small number of nodes. No idea about the recent dual core. Anyone would be more than welcome to provide updated numbers for the tests above (Axel did ! And I haven't collated them yet...) - do keep in mind that in order to be accurate and faithful, the machines needs to be completely empty (i.e. no other jobs running). Also, it makes a big difference running on 4 cpus on 4 dual xeon, where on each platform the second CPU is idle, than on 4 cpus on two dual xeons, all running. The OMP_NUM_THREADS is always set to 1. Let us know, nicola Konstantin Kudin wrote: > Thanks to Axel and Alexander for suggestions on this issue! Some of > them were quite helpful. > > I have investigated things further, in a more careful way. What > happens is that the GOTO library uses its own threads, and that causes > the CP90 time to be underestimated while the wall time stays roughly > the same. This also makes 1cpu times look small, since the 2nd GOTO > thread is quietly computing whenever the 2nd cpu is available. > > I used my own test case, with water and some organic stuff in it. The > run does 20 CP steps starting from a restart file on a local disk. I > eliminated the nfs issues by writing the stuff directly to the local > disk on the head node. Also, for runs with 1 thread the 2nd cpu on the > dual node was kept idle. > > For mpich1 it seems that there is little difference in wall times > whether shared memory is used or not. However, CP90 times appear much > smaller for sockets. For mpich2 shared memory just hangs, while sockets > work. Also, mpich2 is worse in bad situations (more than 4 cpus) than > mpich1. The launcher was compiled accordingly mpiexec from Ohio SC in > all cases. > > Below are some actual numbers. 1 or 2 at the end of the job name means > that the GOTO_NUM_THREADS was either 1 or 2. > > It looks that 2 dual nodes with 1Gbit is the fastest, and the speedup > is about 3x. Beyond that things go downhill. > > Kostya > > ######################### > mpich1 with shared memory > job name G N_thread Wall CP90 > sh1.out.0.1 1 106.5m 1h46m > sh1.out.0.2 1 85.8m 1h24m > sh1.out.1.1 2 62.6m 1h > sh1.out.1.2 2 59.4m 48m > sh1.out.2.1 4 34.8m 33m51.78s > sh1.out.2.2 4 34.6m 27m46.72s > sh1.out.3.1 6 42.7m 37m29.63s > so1.out.3.2 (hangs) > sh1.out.4.1 8 45.3m 37m43.38s > sh1.out.4.2 8 45.6m 35m31.07s > > mpich1 with sockets > job name G N_thread Wall CP90 > so1.out.0.1 1 108.1m 1h47m > so1.out.0.2 1 84.6m 1h23m > so1.out.1.1 2 61.7m 1h > so1.out.1.2 2 62.9m 50m32.96s > so1.out.2.1 4 35.6m 33m10.21s > so1.out.2.2 4 35.6m 27m20.13s > so1.out.3.1 6 42.6m 22m59.73s > so1.out.3.2 6 42.5m 18m50.57s > so1.out.4.1 8 45.1m 18m31.35s > so1.out.4.2 8 43.4m 14m47.97s > > mpich2 with sockets > so2.out.0.1 1 105.3m 1h45m > so2.out.0.2 1 84.5m 1h23m > so2.out.1.1 2 60.7m 1h > so2.out.1.2 2 59.4m 47m31.77s > so2.out.2.1 4 34.6m 32m26.08s > so2.out.2.2 4 34.3m 26m10.19s > so2.out.3.1 6 53.5m 23m > so2.out.3.2 6 53.0m 18m > so2.out.4.1 8 54.0m 17m56.94s > so2.out.4.2 8 53.2m 14m44.32s > > > > > __________________________________ > Yahoo! Mail - PC Magazine Editors' Choice 2005 > http://mail.yahoo.com > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum -- --------------------------------------------------------------------- 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 konstantin_kudin at yahoo.com Wed Sep 28 19:52:29 2005 From: konstantin_kudin at yahoo.com (Konstantin Kudin) Date: Wed, 28 Sep 2005 10:52:29 -0700 (PDT) Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <433ACFA8.1000703@mit.edu> Message-ID: <20050928175230.20376.qmail@web52001.mail.yahoo.com> Hi Nicola, Looking at how the code scales, it appears that in your case even 8 cpus are absolutely faster than 4, which is not the case for our cluster. Which MPI library was that? Also, I could not test on an empty cluster. However, the timings seem to be reproducible enough such as the other network traffic seems to not have too much of an effect in terms of fluctuations from run to run. I have run your AgI test as well, it looks as if something has changed since the latest code appears to be ~25% slower than it used to be in Jan '05. Kostya __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From shaposh at isp.nsc.ru Thu Sep 29 10:17:56 2005 From: shaposh at isp.nsc.ru (Alexander Shaposhnikov) Date: Thu, 29 Sep 2005 15:17:56 +0700 Subject: [Pw_forum] abysmal parallel performance of the CP code In-Reply-To: <433ACFA8.1000703@mit.edu> References: <20050928163448.9694.qmail@web52012.mail.yahoo.com> <433ACFA8.1000703@mit.edu> Message-ID: <1127981876.19427.4.camel@m00> (shorter version) Dear Nicola, I was kinda worried about low perfomance of Opterons you have found, cos we have switched to Opterons from Xeons recently. So i went to my home computer and made some testings with AgI.small.j and CP 2.1.5. Here are the details : System : Dual DualCore Opteron265 1.8GHz (Overclocked a bit to 1.96 GHz. By the way, it works perfectly at 2.35GHz, tested many times with Prime95, linpack, etc. ), 4Gb DDR400(430). Compilers : Intel icc 9.0 and ifort 9.0, EM64T. Flags : icc -O3 -xP , ifort -O2 -xP for all .f90 except cprstart.o ifort -O2 -xW for cprstart.o (to avoid cpu-check error at runtime) BLAS: acml2.7.0 MPI : MPICH2, ch3:shm device I tested with 1 and 4 threads. Timings: 10it 5it 1 11.50 7.35 4 4.0 2.21 With cp.x in Espresso-2.1, close results (a little faster, though) See also attached .out files. So, much faster even on much slower CPUs! Opterons are not that bad. In fact, in most cases they are significantly faster than Xeons with comparable model number. On Wed, 2005-09-28 at 13:15 -0400, Nicola Marzari wrote: > > Dear Kostya, > > > a long time ago I had collated some of our tests on parallel > performance - parallelism over gigabit ethernet > used to go very well up to 4 nodes, and then flatten out. > All the details are in here: > > http://nnn.mit.edu/ESPRESSO/CP90_tests/ > > The "large" system has ~500 bands, so it's a reasonably challenging > task. All the timings are always wall clock time. > > We have more recent tests in the group (maybe someone could post them) > on Dual Xeon 3.4 Ghz, and 800 Mhz FSB. Our detailed experience there, > with a lot of help from Intel, was that the second CPU, living on the > same FSB, provides at best a 25%-30% performance boost. Opterons > achieve on the other hand excellent scaling on the two CPUs on the same > motherboard, but have worse performance at the single CPU level (I > believe thanks to the extensive use of MKL by the CP code). > > My own feeling as the best platform for CP is nodes of 4 or 6 PIVs, > with the fastest memory and FSB available. Gigabit will do fine > for this small number of nodes. No idea about the recent dual core. > > Anyone would be more than welcome to provide updated numbers for the > tests above (Axel did ! And I haven't collated them yet...) - do > keep in mind that in order to be accurate and faithful, > the machines needs to be completely empty (i.e. no other jobs running). > Also, it makes a big difference running on 4 cpus on 4 dual xeon, > where on each platform the second CPU is idle, than on 4 cpus on > two dual xeons, all running. The OMP_NUM_THREADS is always set to 1. > > > > Let us know, > > nicola > > > > Konstantin Kudin wrote: > > > Thanks to Axel and Alexander for suggestions on this issue! Some of > > them were quite helpful. > > > > I have investigated things further, in a more careful way. What > > happens is that the GOTO library uses its own threads, and that causes > > the CP90 time to be underestimated while the wall time stays roughly > > the same. This also makes 1cpu times look small, since the 2nd GOTO > > thread is quietly computing whenever the 2nd cpu is available. > > > > I used my own test case, with water and some organic stuff in it. The > > run does 20 CP steps starting from a restart file on a local disk. I > > eliminated the nfs issues by writing the stuff directly to the local > > disk on the head node. Also, for runs with 1 thread the 2nd cpu on the > > dual node was kept idle. > > > > For mpich1 it seems that there is little difference in wall times > > whether shared memory is used or not. However, CP90 times appear much > > smaller for sockets. For mpich2 shared memory just hangs, while sockets > > work. Also, mpich2 is worse in bad situations (more than 4 cpus) than > > mpich1. The launcher was compiled accordingly mpiexec from Ohio SC in > > all cases. > > > > Below are some actual numbers. 1 or 2 at the end of the job name means > > that the GOTO_NUM_THREADS was either 1 or 2. > > > > It looks that 2 dual nodes with 1Gbit is the fastest, and the speedup > > is about 3x. Beyond that things go downhill. > > > > Kostya > > > > ######################### > > mpich1 with shared memory > > job name G N_thread Wall CP90 > > sh1.out.0.1 1 106.5m 1h46m > > sh1.out.0.2 1 85.8m 1h24m > > sh1.out.1.1 2 62.6m 1h > > sh1.out.1.2 2 59.4m 48m > > sh1.out.2.1 4 34.8m 33m51.78s > > sh1.out.2.2 4 34.6m 27m46.72s > > sh1.out.3.1 6 42.7m 37m29.63s > > so1.out.3.2 (hangs) > > sh1.out.4.1 8 45.3m 37m43.38s > > sh1.out.4.2 8 45.6m 35m31.07s > > > > mpich1 with sockets > > job name G N_thread Wall CP90 > > so1.out.0.1 1 108.1m 1h47m > > so1.out.0.2 1 84.6m 1h23m > > so1.out.1.1 2 61.7m 1h > > so1.out.1.2 2 62.9m 50m32.96s > > so1.out.2.1 4 35.6m 33m10.21s > > so1.out.2.2 4 35.6m 27m20.13s > > so1.out.3.1 6 42.6m 22m59.73s > > so1.out.3.2 6 42.5m 18m50.57s > > so1.out.4.1 8 45.1m 18m31.35s > > so1.out.4.2 8 43.4m 14m47.97s > > > > mpich2 with sockets > > so2.out.0.1 1 105.3m 1h45m > > so2.out.0.2 1 84.5m 1h23m > > so2.out.1.1 2 60.7m 1h > > so2.out.1.2 2 59.4m 47m31.77s > > so2.out.2.1 4 34.6m 32m26.08s > > so2.out.2.2 4 34.3m 26m10.19s > > so2.out.3.1 6 53.5m 23m > > so2.out.3.2 6 53.0m 18m > > so2.out.4.1 8 54.0m 17m56.94s > > so2.out.4.2 8 53.2m 14m44.32s > > > > > > > > > > __________________________________ > > Yahoo! Mail - PC Magazine Editors' Choice 2005 > > http://mail.yahoo.com > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > -------------- next part -------------- A non-text attachment was scrubbed... Name: AgI.out.1.10.gz Type: application/x-gzip Size: 5474 bytes Desc: not available Url : /pipermail/attachments/20050929/b05a38be/attachment.bin -------------- next part -------------- A non-text attachment was scrubbed... Name: AgI.out.1.5.gz Type: application/x-gzip Size: 4928 bytes Desc: not available Url : /pipermail/attachments/20050929/b05a38be/attachment-0001.bin -------------- next part -------------- A non-text attachment was scrubbed... Name: AgI.out.4.10.gz Type: application/x-gzip Size: 5567 bytes Desc: not available Url : /pipermail/attachments/20050929/b05a38be/attachment-0002.bin -------------- next part -------------- A non-text attachment was scrubbed... Name: AgI.out.4.5.gz Type: application/x-gzip Size: 5055 bytes Desc: not available Url : /pipermail/attachments/20050929/b05a38be/attachment-0003.bin From asen at iitk.ac.in Thu Sep 29 13:35:59 2005 From: asen at iitk.ac.in (A Sen) Date: Thu, 29 Sep 2005 17:05:59 +0530 (IST) Subject: [Pw_forum] CPV compilation successful Message-ID: <46668.172.28.37.46.1127993759.squirrel@webmail.iitk.ac.in> Dear Gerardo and Stefano, Thanks a million for the help. It compiles perfectly now. Regards, A Sen From wmbmacam at lg.ehu.es Thu Sep 29 18:20:36 2005 From: wmbmacam at lg.ehu.es (=?UTF-8?B?TWlndWVsIE1hcnTDrW5leiBDYW5hbGVz?=) Date: Thu, 29 Sep 2005 18:20:36 +0200 Subject: [Pw_forum] El-Ph funny error Message-ID: <433C1454.5060800@lg.ehu.es> Dear people, I'm still trying to get my hands on electron phonon calculations with Al. It finally seemed that I worked it out... but I was wrong. I have calculated the lambdas at a 8 8 8 regular grid (given by kpoints.x). I am using a PBE USPP, with 40 Ry and summing in a 24 24 24 kpoint grid. While some points work OK, some do not. I have done the phonon and the lambda calculation with the same nscf run (elph=trans=.true.). The error message I get is (al.elphon.out) the following: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% from phq_init : error # 1 wrong order of k points %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% stopping... Could anybody shed some light? Thanks in advance, Miguel PS: 1 processor, 3gb of memory are being used -- ---------------------------------------- Miguel Mart?nez Canales Dto. F?sica de la Materia Condensada UPV/EHU Facultad de Ciencia y Tecnolog?a Apdo. 644 48080 Bilbao (Spain) Fax: +34 94 601 3500 Tlf: +34 94 601 5437 ---------------------------------------- 2005 International Year of Physics From leoant21 at hotmail.com Fri Sep 30 08:35:05 2005 From: leoant21 at hotmail.com (=?gb2312?B?1cUguumx8g==?=) Date: Fri, 30 Sep 2005 06:35:05 +0000 Subject: [Pw_forum] Re: Question on supercell calculation of MgB2 In-Reply-To: <20050929053629.22002.11760.Mailman@democritos.sissa.it> Message-ID: Dear Eyvaz: Thank you very much for your reply, now I am calculating and I want to see whether the results with this warning is the same with a correct one, so I haven't stopped the program. However, I want to mention that I am just doing a scf calculation now, there's no nband parameter setted.please give me some comments. Regards > >Message: 1 >Date: Tue, 27 Sep 2005 23:51:47 -0700 (PDT) >From: Eyvaz Isaev >Subject: Re: [Pw_forum] Question on supercell calculation of MgB2 >To: pw_forum at pwscf.org >Reply-To: pw_forum at pwscf.org > >Hi, > >This message was discussed before in the forum. >It means unsufficient bands number which is defined as >Number_of_valence_electrons*(1/2 + 20%). Just increase >bands number (nbnd). > >Bests, >Eyvaz. > >--- ? ?? wrote: > > > Dear all: > > Now I am doing a 2*2 supercell calculation of > > MgB2, however, there is a > > warning like this: > > > > Warning: ef = 1.818386 is above the highest band at > > k-point 490 > > e = 1.780277 > > > > and there are also warning on other k-points, but > > not all. Please tell me > > why. Thank you. > > > > Regards. > > > > >_________________________________________________________________ > > ???? MSN Explorer: > > http://explorer.msn.com/lccn > > > > _______________________________________________ > > Pw_forum mailing list > > Pw_forum at pwscf.org > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > >__________________________________ >Yahoo! Mail - PC Magazine Editors' Choice 2005 >http://mail.yahoo.com > >--__--__-- _________________________________________________________________ ??????????????? MSN Hotmail? http://www.hotmail.com From hqzhou at nju.edu.cn Fri Sep 30 11:39:34 2005 From: hqzhou at nju.edu.cn (Huiqun Zhou) Date: Fri, 30 Sep 2005 17:39:34 +0800 Subject: [Pw_forum] About compiling cd.x and fpmd.x Message-ID: <008901c5c5a2$ddf19ca0$1d00a8c0@solarflare> dear list users, I have successfully compiled all programs of espresso 2.1.5 on a P4 cluster with very aggressive optimization options (-O3 -tpp7 -xN -unroll -r8 -pc64). pw.x works very well both in serial and parallel runs but cp.x and fpmd.x don't. Although cp.x works in serial run, it failed in parallel execution. fpmd.x is the worst, it fails in both serial and parallel runs. fpmd.x (example21) both serial and parallel runs (serial is run on local computer) got same error message as below (memory addresses are different): ... some results of calculation (mostly initialization, I think) ... running the calculation with 32 H2O molecules... *** glibc detected *** free(): invalid next size (fast) : 0x1bebfad8 *** forrtl: error (76): IOT trap signal cp.x (example18) serial run is OK. parallel run got error below: ... some results of calculation ... forrtl: No such file or directory forrtl: severe (29): file not found, unit 93, file /my/home/espresso-2.1.5/examples/example18/results/ .//fort.93 ...... Note: the serial and parallel runs use same parallel version of cp.x and fpmd.x. I also tried to compile a less optimized version with option -O2 -tpp7 -r8 -pc64, but got same problem. In both aggressive and less optimized versions, Intel MKL libraries were linked statically (-i-static option). Other libraries may be linked dynamically. But my compute nodes are homogenous architecture with precisely equivalent OS and other stuff, including Intel icc, ifort, and MKL. Thanks in advance. Huiqun Zhou Can anybody give me a pointer? -------------- next part -------------- An HTML attachment was scrubbed... URL: /pipermail/attachments/20050930/19d831d8/attachment.htm From eyvaz_isaev at yahoo.com Fri Sep 30 14:45:34 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 30 Sep 2005 05:45:34 -0700 (PDT) Subject: [Pw_forum] Re: Question on supercell calculation of MgB2 In-Reply-To: Message-ID: <20050930124534.12629.qmail@web60321.mail.yahoo.com> Hi, By default the number of bands, nbnd, used for integration over the BZ is Number_of_valence_electrons*(1/2 + 1/5) (i.e. additional 20% of bands number). For "nelec" see your output file. During integration over the BZ ef for some reason became higher then the largest energy for a given k-point (e_k(nbnd)), because you have not enough number of bands, nbnd, so that you have a leak of electrons. Bests, Eyvaz. --- ?? ???? wrote: > Dear Eyvaz: > Thank you very much for your reply, now I am > calculating and I want to > see whether the results with this warning is the > same with a correct one, > so I haven't stopped the program. > However, I want to mention that I am just doing a > scf calculation now, > there's no nband parameter setted.please give me > some comments. > Regards > > > >Message: 1 > >Date: Tue, 27 Sep 2005 23:51:47 -0700 (PDT) > >From: Eyvaz Isaev > >Subject: Re: [Pw_forum] Question on supercell > calculation of MgB2 > >To: pw_forum at pwscf.org > >Reply-To: pw_forum at pwscf.org > > > >Hi, > > > >This message was discussed before in the forum. > >It means unsufficient bands number which is defined > as > >Number_of_valence_electrons*(1/2 + 20%). Just > increase > >bands number (nbnd). > > > >Bests, > >Eyvaz. > > > >--- ?? ???? wrote: > > > > > Dear all: > > > Now I am doing a 2*2 supercell calculation of > > > MgB2, however, there is a > > > warning like this: > > > > > > Warning: ef = 1.818386 is above the highest band > at > > > k-point 490 > > > e = 1.780277 > > > > > > and there are also warning on other k-points, > but > > > not all. Please tell me > > > why. Thank you. > > > > > > Regards. > > > > > > > >_________________________________________________________________ > > > ???????? MSN Explorer: > > > http://explorer.msn.com/lccn > > > > > > _______________________________________________ > > > Pw_forum mailing list > > > Pw_forum at pwscf.org > > > > http://www.democritos.it/mailman/listinfo/pw_forum > > > > > > > > > > > > >__________________________________ > >Yahoo! Mail - PC Magazine Editors' Choice 2005 > >http://mail.yahoo.com > > > >--__--__-- > > _________________________________________________________________ > ?????????????????????????????? MSN Hotmail?? > http://www.hotmail.com > > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > ______________________________________________________ Yahoo! for Good Donate to the Hurricane Katrina relief effort. http://store.yahoo.com/redcross-donate3/ From eyvaz_isaev at yahoo.com Fri Sep 30 14:56:54 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 30 Sep 2005 05:56:54 -0700 (PDT) Subject: [Pw_forum] El-Ph funny error In-Reply-To: <433C1454.5060800@lg.ehu.es> Message-ID: <20050930125655.21950.qmail@web60322.mail.yahoo.com> Hi, In my experience this kind of error appeares due to incomplete calculations in the previous step. So, check your output files. Bests, Eyvaz. --- Miguel Mart?nez Canales wrote: > Dear people, > > I'm still trying to get my hands on electron phonon > calculations with Al. > It finally seemed that I worked it out... but I was > wrong. I have > calculated the lambdas at a 8 8 8 regular grid > (given by kpoints.x). I am > using a PBE USPP, with 40 Ry and summing in a 24 24 > 24 kpoint grid. > > While some points work OK, some do not. I have done > the phonon and the > lambda calculation with the same nscf run > (elph=trans=.true.). The error > message I get is (al.elphon.out) the following: > > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > from phq_init : error # 1 > wrong order of k points > %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% > stopping... > > Could anybody shed some light? Thanks in advance, > > Miguel > > PS: 1 processor, 3gb of memory are being used > > -- > ---------------------------------------- > Miguel Mart?nez Canales > Dto. F?sica de la Materia Condensada > UPV/EHU > Facultad de Ciencia y Tecnolog?a > Apdo. 644 > 48080 Bilbao (Spain) > Fax: +34 94 601 3500 > Tlf: +34 94 601 5437 > ---------------------------------------- > > 2005 International Year of Physics > _______________________________________________ > Pw_forum mailing list > Pw_forum at pwscf.org > http://www.democritos.it/mailman/listinfo/pw_forum > __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com From vranjan at ncsu.edu Fri Sep 30 19:54:34 2005 From: vranjan at ncsu.edu (Vivek Ranjan) Date: Fri, 30 Sep 2005 13:54:34 -0400 (EDT) Subject: [Pw_forum] problem with neb calculation Message-ID: <32886.152.14.74.99.1128102874.squirrel@webmail.ncsu.edu> Dear Users, I ran the attached input file on three different machines with : (i) Xeon processors, (ii) IBM p690, (iii) another machine with IBM p690. I could run the programme successfully on the first two, but the run on the third machine stopped with the following statement : ------------------------------ iteration 1 ------------------------ tcpu = 0.1; self-consistency for image 1 ** ON ENTRY TO DSYGV PARAMETER NUMBER 2 HAD AN ILLEGAL VALUE -------------------------------------------------------------------- Usually everything is similar on the second and the third machines mentioned above. As a result the same makefile works well on both of them. I would also like to mention that I have successfully performed many calculations using PWSCF on all three machines. So, my PWSCF compilation seems to be alright. Any hints ? Thank you, Vivek __________________________________________________________________________ web : http://www.geocities.com/vivekranjanweb __________________________________________________________________________ Vivek Ranjan Email : vranjan at chips.ncsu.edu Center for High Performance Simulation : vranjan at ncsu.edu Department of Physics NC State University Phone : +1-919-513-0613 (Office) Campus Box 7518 Fax : +1-919-513-4804 Raleigh, NC 27695 -------------------------------------------------------------------------- -------------- next part -------------- A non-text attachment was scrubbed... Name: run-neb.sc Type: application/octet-stream Size: 1661 bytes Desc: not available Url : /pipermail/attachments/20050930/7bddbe3e/attachment.obj From silviu at Princeton.EDU Fri Sep 30 22:11:11 2005 From: silviu at Princeton.EDU (Silviu Zilberman) Date: Fri, 30 Sep 2005 22:11:11 +0200 Subject: [Pw_forum] g95 compilation problem In-Reply-To: <1127922487.433abb37a1aaa@webmail.sissa.it> References: <20050927110733.39CC911280D@democritos.sissa.it> <433945D9.7080102@Princeton.EDU> <433A9BBF.6070003@Princeton.EDU> <1127922487.433abb37a1aaa@webmail.sissa.it> Message-ID: <433D9BDF.1050502@Princeton.EDU> Dear Paolo and Stefano, Thank you for your comments. I had to introduce a couple of other minor changes but now the code compiles. Running however seems still to be a problem, but I hope to make it work. Thanks, Silviu. Stefano de Gironcoli wrote: >Again a message from Paolo > >On Tue, 27 Sep 2005 15:15:05 +0200 > Silviu Zilberman wrote: > > > >>g95 -O -D__FFTW -I../include -I. -I../Modules -I../PW >>-I../PH -I../iotk/src -c fft_scalar.F90 -o fft_scalar.o >>In file fft_scalar.F90:78 >> >> C_POINTER :: fw_plan( 3, ndims ) = 0 >> 1 >>Error: Unclassifiable statement at (1) >> >> > >preprocessing flags are not sufficient: I think you >should have at least -D__LINUX -D__G95 -D__FFTW . >C_POINTER has to be preprocessed to something sensible. > > >---------------------------------------------------------------- > SISSA Webmail https://webmail.sissa.it/ > Powered by Horde http://www.horde.org/ > > >_______________________________________________ >Pw_forum mailing list >Pw_forum at pwscf.org >http://www.democritos.it/mailman/listinfo/pw_forum > > -------------- next part -------------- A non-text attachment was scrubbed... Name: silviu.vcf Type: text/x-vcard Size: 247 bytes Desc: not available Url : /pipermail/attachments/20050930/38faf1c7/attachment.vcf From eyvaz_isaev at yahoo.com Fri Sep 30 22:32:22 2005 From: eyvaz_isaev at yahoo.com (Eyvaz Isaev) Date: Fri, 30 Sep 2005 13:32:22 -0700 (PDT) Subject: [Pw_forum] problem with neb calculation In-Reply-To: <32886.152.14.74.99.1128102874.squirrel@webmail.ncsu.edu> Message-ID: <20050930203222.62914.qmail@web60319.mail.yahoo.com> Hi, It looks like the next message: ** ON ENTRY TO ZHPEV PARAMETER NUMBER 1 HAD AN ILLEGAL VALUE There is Paolo's comments on this: >ESSL contains a version of ZHPEV that is incompatible > with ZHPEV provided by LAPACK. If __AIX is defined, > the former is used. > Load ESSL before LAPACK. Try to follow this advice, i.e. recompile PWSCF with options -lessl -llapack. Bests, Eyvaz. --- Vivek Ranjan wrote: > Dear Users, > > I ran the attached input file on three different > machines with : (i) Xeon > processors, (ii) IBM p690, (iii) another machine > with IBM p690. I could > run the programme successfully on the first two, but > the run on the third > machine stopped with the following statement : > > > ------------------------------ iteration 1 > ------------------------ > tcpu = 0.1; self-consistency for image > 1 > ** ON ENTRY TO DSYGV PARAMETER NUMBER 2 HAD AN > ILLEGAL VALUE > > > -------------------------------------------------------------------- > > Usually everything is similar on the second and the > third machines > mentioned above. As a result the same makefile > works well on both of > them. I would also like to mention that I have > successfully performed many > calculations using PWSCF on all three machines. So, > my PWSCF compilation > seems to be alright. > > Any hints ? > > Thank you, > > Vivek > > __________________________________________________________________________ > web : http://www.geocities.com/vivekranjanweb > __________________________________________________________________________ > Vivek Ranjan Email : > vranjan at chips.ncsu.edu > Center for High Performance Simulation : > vranjan at ncsu.edu > Department of Physics > NC State University Phone : > +1-919-513-0613 (Office) > Campus Box 7518 Fax : > +1-919-513-4804 > Raleigh, NC 27695 > -------------------------------------------------------------------------- __________________________________ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com