From radbalu at comcast.net Tue Feb 1 13:47:52 2005 From: radbalu at comcast.net (Rad Balu) Date: Tue, 1 Feb 2005 07:47:52 -0500 Subject: [CPMD-list] Very Large Negative Frequencies Message-ID: <001901c5085c$3e515c20$1262f345@D3B9FD51> Dear All, I am trying to do a vibrational analysis on a system (around 50 atoms) using CPMD version 3.9. I am getting large negative frequencies as part of the output as shown here: HARMONIC FREQUENCIES [cm**-1]: -23281.9544 -23272.5876 -22483.4345 -22463.8015 -9610.4496 -9576.3246 -6998.5338 -6654.0035 -6359.3503 -6311.7344 -4500.7134 -4161.7275 -3625.4781 -3623.1194 -2297.5825 -2046.2618 -1957.2234 -1920.9401 -1794.1146 -1721.5709 I did optimize the geometry before the vibrational analysis and a portion of the input is shown here: &CPMD OPTIMIZE GEOMETRY TSDC LBFGS LANCZOS DIAGONALIZATION LANCZOS PARAMETERS 1 6 10 1.D-8 SPLINE RANGE 6.0 SPLINE POINTS 3000 RESTLINE 2 STRUCTURE BONDS ANGLES DIHEDRALS CONVERGENCE ADAPT 0.000001 RESTART LATEST I tried to calculate the frequencies using PBC (w/o PBC the frequencies are normal) and here is my input for the run &CPMD VIBRATIONAL ANALYSIS LBFGS HESSIAN DISCO SPLINE RANGE 6.0 SPLINE POINTS 3000 RESTLINE 2 STRUCTURE BONDS ANGLES DIHEDRALS CONVERGENCE ADAPT 0.000001 BIG MEMORY RESTART LATEST I would appreciate if somebody can provide some suggestions to improve the results. Please let me know if you need more information on how I set up the input. Thanks Rad Balu -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20050201/6d3869ff/attachment.html From Ari.P.Seitsonen at iki.fi Tue Feb 1 14:54:20 2005 From: Ari.P.Seitsonen at iki.fi (Ari P Seitsonen) Date: Tue, 1 Feb 2005 14:54:20 +0100 (CET) Subject: [CPMD-list] Very Large Negative Frequencies In-Reply-To: <001901c5085c$3e515c20$1262f345@D3B9FD51> References: <001901c5085c$3e515c20$1262f345@D3B9FD51> Message-ID: Dear Rad Balu, Did you specify RESTART GEOMETRY as well? Have you checked that your optimised geometry was indeed okay, and that it was indeed used (by checking the output of the corresponding calculation) when calculating the frequencies? Did you have the corresponding, optimised cell? I'd be somewhat careful taking the geometry from a calculation with 'TSDC', unless you checked later that the cell is indeed properly optimised (for example with respect to Pulay stresses). Greetings from Paris, apsi -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=- Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/ Tel +41 1 635 44 97 / Fax +41 1 635 68 38 / GSM +41 79 719 09 35 Anschrift: Physikalisch Chemisches Institut (PCI), Universit?t Z?rich (UniZh) Indirizzo: Winterthurerstrasse 190, CH-8057 Z?rich Address: Schweiz / Svizzera / Suisse / Svizra / Switzerland On Tue, 1 Feb 2005, Rad Balu wrote: > Dear All, > > I am trying to do a vibrational analysis on a system (around 50 atoms) using CPMD version 3.9. I am getting large negative > frequencies as part of the output as shown here: > HARMONIC FREQUENCIES [cm**-1]: > > -23281.9544 -23272.5876 -22483.4345 -22463.8015 > -9610.4496 -9576.3246 -6998.5338 -6654.0035 > -6359.3503 -6311.7344 -4500.7134 -4161.7275 > -3625.4781 -3623.1194 -2297.5825 -2046.2618 > -1957.2234 -1920.9401 -1794.1146 -1721.5709 > > > I did optimize the geometry before the vibrational analysis and a portion of the input is shown here: > &CPMD > OPTIMIZE GEOMETRY > TSDC > LBFGS > LANCZOS DIAGONALIZATION > LANCZOS PARAMETERS > 1 6 10 1.D-8 > SPLINE RANGE > 6.0 > SPLINE POINTS > 3000 > RESTLINE > 2 > STRUCTURE BONDS ANGLES DIHEDRALS > CONVERGENCE ADAPT > 0.000001 > RESTART LATEST > > I tried to calculate the frequencies using PBC (w/o PBC the frequencies are normal) and here is my input for the run > > &CPMD > VIBRATIONAL ANALYSIS > LBFGS > HESSIAN DISCO > SPLINE RANGE > 6.0 > SPLINE POINTS > 3000 > RESTLINE > 2 > STRUCTURE BONDS ANGLES DIHEDRALS > CONVERGENCE ADAPT > 0.000001 > BIG MEMORY > RESTART LATEST > > I would appreciate if somebody can provide some suggestions to improve the results. Please let me know if you need more information > on how I set up the input. > > Thanks > Rad Balu From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Tue Feb 1 21:27:26 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Tue, 01 Feb 2005 21:27:26 +0100 Subject: [CPMD-list] cpmd tutorial webpages available Message-ID: <200502012027.j11KRQx32072@yello.theochem.ruhr-uni-bochum.de> hello everybody on cpmd-list, i am very pleased to announce the availability of an initial version of a CPMD beginner's tutorial at http://www.theochem.ruhr-uni-bochum.de/go/cpmd-tutor.html the material was initially put together by Carme Rovira, Roger Rousseau, and myself and first presented in its initial form during a CPMD tutorial at the Centre for Research in Theoretical Chemistry of the Parc Cientific de Barcelona in november 2004. i've since then rearranged it and cleaned it up a little bit for publishing on the web. it is still work in progress and there is a lot of stuff that can be improved or added, but it should be useful enough to announce it publically. nevertheless, i'd like to invite all of you to contribute by sending or suggesting further exercises, improvements or corrections. we also have come up with a way to organize the tutorial pages so they can be combined it with a knoppix-like installation. (this is how we used it in barcelona). due to the various license restrictions on Knoppix and CPMD, we cannot distribute it or publish a script that will create it, so you have to contact me personally to learn how this can be done, e.g. if you want to do such a course on your on. my future plans are, time permitting, to have it moved to cpmd.org as soon as it has reached some finished form and add a section with 'advanced exercises', that will demonstrate many of the cool features in CPMD that go beyond standard CP-MD are not so easy to use (at least for less experience users). for the first of these goals i'd need especially the feedback from less experienced CPMD users, so i know what is missing or could be explained better. for the latter goal it would be nice to get some contributions from people who have implemented or successfully used those advanced features. finally i'd like to send a special thank you to Carme and Roger, since these pages would have never happend without their 'interference'. happy CPMD-ing, axel kohlmeyer. -- ======================================================================= 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 axel.kohlmeyer at theochem.ruhr-uni-bochum.de Wed Feb 2 11:26:21 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Wed, 2 Feb 2005 11:26:21 +0100 (CET) Subject: [CPMD-list] from younes ansari In-Reply-To: <20050127075315.17521.qmail@web50908.mail.yahoo.com> Message-ID: On Wed, 26 Jan 2005, Younes Ansari wrote: YA> Dear sir : dear younes ansari, YA> I have checked out deferent methods and just got the same results. YA> firstly I think the definitions of atoms which is in PPLIBNEW directory YA> are deferent from other programs like Gaussian so I want to get the same YA> result in both Gaussian and CPMD. this is simply not possible, mainly for three reasons: - you are using pseudopotentials to describe atoms with CPMD (or any other pseudopotential/plane wave/DFT code), so you will only get the energy of the (valence) electrons not included in the pseudocore. - you are always calculating the energy for the whole supercell on a grid. so the size of the supercell and the size of the grid (usually implicitely defined via the plane wave cutoff) have an influence on the total energy. - you have different types of basis functions (gaussians vs. plane waves) with different properties that represent parts of the wavefunction differently well. this all, however, should have little impact on _relative_ energy differences, e.g. between different conformations or configurations with the same number of atoms electrons and the same cell/pseudopotential/cutoff settings and the equivalent gaussian calculation. provided you are reasonably converged with respect to the basis set. since the dominant application of the CPMD code are Car-Parrinello MD calculations people care mainly about a proper representation of the interatomic forces in the most efficient way (so you don't need insane amounts of computer power for your simulations). this works (usually) very well with the plane-wave-pseudopotential approach. please check out some of the introductory material on electron structure calculations with plane-waves and pseudopotentials. there are nice collections of them or links to it on juerg hutter's homepage: http://www.pci.unizh.ch/gruppe.hutter/e/information.html and on cpmd.org: http://www.cpmd.org/cpmd_thecode.html the bottom line is, that the approach used by gaussian and by CPMD are quite complementary, i.e. Gaussian is very efficient for small isolated molecules, while CPMD is very at describing bulk liquids and solids/surfaces. YA> YA> As a result I have checked the ionization energies for Na...NA+ , YA> K...K+ , Rb...Rb+ and figured out that the results for K and Rb are YA> comparable with literatures but Na has a lot differences. alkali metals are especially tricky with pseudopotentials. for accurate results you usually need a semi-core pseudopotential (i.e. a pseudopotential, that not only includes the outer s electron but also electrons from the filled core shells in its valence) and/or non-linear core correction. YA> Please tell me how I can get a Gaussian type calculation (I mean the YA> Etot results) in the CPMD program. you cannot. see above. YA> secondly, I have two kinds of Au,Rb,K,....in GEODECKER for example YA> (K-q1,Kq9) which are the definitions are better to use for a simple YA> GEOMETRY calculation. the difference between K-q1 and K-q9 is exactly the difference between a regular and a semi-core pseudopotential. please have a look at the CPMD output and compare the number of electrons used in the calculation. the q1 potential only includes the 4s electron, while the q9 potential also includes the 3s and 3p electrons. which one to prefer depends on your system and the properties you want to calculate. with the q9 pseudopotential you will need a much larger basis set, i.e. a higher cutoff for converged results, but then again, the pseudopotential may be much more transferable. for example in the case of potassium, the K-q1 pseudopotential should good for metallic potassium but for K+ ions, you will probably need the K-q9 since in the case of the K-q1 there would be next to no electron density left at the K ion which should result in a poor description of the total valence structure. regards, axel kohlmeyer. YA> Thanks... YA> > YA> YA> YA> YA> --------------------------------- YA> Do you Yahoo!? YA> Yahoo! Search presents - Jib Jab's 'Second Term' -- ======================================================================= 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 dogbe at unr.nevada.edu Wed Feb 2 22:45:26 2005 From: dogbe at unr.nevada.edu (John Kofi Dogbe) Date: Wed, 2 Feb 2005 13:45:26 -0800 (PST) Subject: [CPMD-list] Concerning CPMD Tutorials In-Reply-To: <200502021100.MAA16932@cpmd.org> References: <200502021100.MAA16932@cpmd.org> Message-ID: Dear CPMD User, I just visited the CPMD tutorial page put together by Dr. Axel Kohlmeyer. I do not have any questions at this point but I would like express my gratitude to him as a CPMD user for putting up the page. Even though I have been using CPMD for sometime now and already familiar with a large part of the stuff on the page, I still find many other hints and links here and there very useful. It also gives me the confidence that I've been doing things right. So, Dr. Axel thank you for putting up the tutorial page and I look forward to visiting the page more often to take advantage of the more advance features of CPMD as you post them. By the way are there any plans to give a list of subroutines that users with guts to delve into the code and see what CPMD is really doing? Together with the manual, one can get a clearer picture on usage ... It's time consuming to do so right now since one has to go guessing the appropriate routine. Ok. Thanks again Dr. Axel and all other who contribute to the tutorial page. Regards from the sunny but freezing desert town of Nevada(?). John -- **The first is not necessarily the Leader** "If I have spoken evil, bear witness of the evil: but if well, why smitest thou me?" -- Jesus Christ (John, 18:23) From symsun69 at yahoo.com.cn Thu Feb 3 01:59:03 2005 From: symsun69 at yahoo.com.cn (Yuming Sun) Date: Thu, 3 Feb 2005 08:59:03 +0800 (CST) Subject: [CPMD-list] problem with keyword POLYMER Message-ID: <20050203005903.30283.qmail@web15705.mail.cnb.yahoo.com> Dear dall, I have tried calculating carbon nanotube using keyword POLYMER in cpmd 3.9.1, bu I was given error which did not occur for cpmd 3.7.2 : INITIALIZATION TIME: 1.57 SECONDS *** WFOPTS| THE NEW SIZE OF THE PROGRAM IS 3640 KBytes *** *** PHFAC| THE NEW SIZE OF THE PROGRAM IS 3672 KBytes *** *** ATOMWF| THE NEW SIZE OF THE PROGRAM IS 4008 KBytes *** ATRHO| CHARGE(R-SPACE): 112.000000 (G-SPACE): 112.000000 [ DSYGVX| INFO= 111 DSYGVX| FAILED TO CONVERGE WHEN DIAGONALIZING A PROGRAM STOPS IN SUBROUTINE DSYGVX| FAILED TO DIAGONALIZE [PROC= 0] 0] MPI Abort by user Aborting program ! [0] MPI Abort by user Aborting program ! [0] Kill all processes of the task! below is my input file: &CPMD OPTIMIZE WAVEFUNCTION ODIIS 4 &END &SYSTEM POLYMER SYMMETRY 0 CELL 4.3128 4.0 4.0 0.0 0.0 0.0 CUTOFF 35. POISSON SOLVER MORTENSEN &END &ATOMS Carbon *C_MT_GIA_BLYP KLEINMANN-BYLANDER LMAX=P LOC=P 28 7.188010E-01 2.499351E+00 1.203624E+00 3.594005E+00 2.499351E+00 1.203624E+00 2.875204E+00 1.729605E+00 2.168856E+00 1.437602E+00 1.729605E+00 2.168856E+00 7.188011E-01 6.172889E-01 2.704519E+00 3.594005E+00 6.172887E-01 2.704519E+00 2.875204E+00 -6.172887E-01 2.704519E+00 1.437600E+00 -6.172929E-01 2.704518E+00 7.187998E-01 -1.729607E+00 2.168854E+00 3.594003E+00 -1.729608E+00 2.168853E+00 2.875203E+00 -2.499352E+00 1.203622E+00 1.437601E+00 -2.499352E+00 1.203623E+00 7.188014E-01 -2.774071E+00 5.379503E-07 3.594005E+00 -2.774071E+00 -1.341569E-06 2.875205E+00 -2.499352E+00 -1.203624E+00 1.437603E+00 -2.499352E+00 -1.203623E+00 7.188030E-01 -1.729607E+00 -2.168854E+00 3.594006E+00 -1.729606E+00 -2.168855E+00 2.875206E+00 -6.172921E-01 -2.704518E+00 1.437605E+00 -6.172930E-01 -2.704518E+00 7.187996E-01 6.172912E-01 -2.704518E+00 3.594008E+00 6.172846E-01 -2.704520E+00 2.875203E+00 1.729607E+00 -2.168854E+00 1.437601E+00 1.729606E+00 -2.168855E+00 7.188013E-01 2.499351E+00 -1.203624E+00 3.594005E+00 2.499352E+00 -1.203623E+00 2.875205E+00 2.774071E+00 -3.385571E-07 1.437603E+00 2.774071E+00 -1.278317E-06 &END &DFT FUNCTIONAL BLYP &END How can I overcome this problem? Thanks a lot. Yuming Sun Dr. Yuming Sun Department of Physics Yantai University Yantai, Shandong 264005 P.R. China Tel. 86-535-6904965 Fax 86-535-6901947 --------------------------------- Do You Yahoo!? 150????MP3???????????????????????? ?????????????????????????????????????? 1G????1000?????????????????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20050203/f2f37cca/attachment.html From cur at zurich.ibm.com Thu Feb 3 09:38:27 2005 From: cur at zurich.ibm.com (Alessandro Curioni) Date: Thu, 3 Feb 2005 09:38:27 +0100 Subject: Fw: [CPMD-list] Concerning CPMD Tutorials Message-ID: John, just take the htmldoc.tar.gz file in the contrib section of the cpmd distribution ( thanks to T. Deutsch). It will create a full tree of the cpmd distribution where you can see dependencies, include files etc. It is really helpfull. Regards. Alessandro CURIONI, PhD Research Staff Member Computational Biochemistry and Material Science group IBM Research Division - Zurich Research Laboratory Saumerstrasse 4 8003 Rueschlikon - Switzerland e-mail: cur at zurich.ibm.com www: www.zurich.ibm.com Tel: +41-1-7248633 Fax: +41-1-7248958 From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Thu Feb 3 09:59:22 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Thu, 3 Feb 2005 09:59:22 +0100 (CET) Subject: [CPMD-list] problem with keyword POLYMER In-Reply-To: <20050203005903.30283.qmail@web15705.mail.cnb.yahoo.com> Message-ID: On Thu, 3 Feb 2005, Yuming Sun wrote: YS> Dear dall, dear yuming sun, there is one problem with your input. you seem to be using coordinates in angstrom, but you are not using the ANGSTROM keyword. unfortunately, even after fixing that using POLYMER produces the same error. YS> I have tried calculating carbon nanotube using keyword POLYMER in cpmd YS> 3.9.1, bu I was given error which did not occur for cpmd 3.7.2 : sorry to disagree, but i get the same error with 3.7.2 here. YS> INITIALIZATION TIME: 1.57 SECONDS YS> *** WFOPTS| THE NEW SIZE OF THE PROGRAM IS 3640 KBytes *** YS> *** PHFAC| THE NEW SIZE OF THE PROGRAM IS 3672 KBytes *** YS> *** ATOMWF| THE NEW SIZE OF THE PROGRAM IS 4008 KBytes *** YS> ATRHO| CHARGE(R-SPACE): 112.000000 (G-SPACE): 112.000000 YS> [ DSYGVX| INFO= 111 YS> DSYGVX| FAILED TO CONVERGE WHEN DIAGONALIZING A YS> YS> PROGRAM STOPS IN SUBROUTINE DSYGVX| FAILED TO DIAGONALIZE [PROC= 0] YS> 0] MPI Abort by user Aborting program ! YS> [0] MPI Abort by user Aborting program ! YS> [0] Kill all processes of the task! i currently don't have the time to look into this, but this error reminds me of examples where you have significant electron density in an area, where there should be none due to the boundary conditions for the poisson solver. for your system, using just symmetry 8 an no poisson solver should be a good alternative. perhaps somebody else is able to help you more. regards, axel kohlmeyer. [...] -- ======================================================================= 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 axel.kohlmeyer at theochem.ruhr-uni-bochum.de Thu Feb 3 10:29:07 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Thu, 3 Feb 2005 10:29:07 +0100 (CET) Subject: [CPMD-list] Concerning CPMD Tutorials In-Reply-To: Message-ID: On Wed, 2 Feb 2005, John Kofi Dogbe wrote: JD> Dear CPMD User, dear john, JD> I just visited the CPMD tutorial page put together by Dr. Axel Kohlmeyer. JD> I do not have any questions at this point but I would like express my JD> gratitude to him as a CPMD user for putting up the page. Even though I JD> have been using CPMD for sometime now and already familiar with a large JD> part of the stuff on the page, I still find many other hints and links JD> here and there very useful. It also gives me the confidence that I've been JD> doing things right. thank you very much. i hope that it will motivate more people to share their experiences using CPMD, so everybody can benefit from it. occasionally i hear people complaining (mostly off-list) about the lack of good documentation for CPMD or people doing stupid things with CPMD, but unless 'those who know'(TM) help to provide the information it will stay that way. personally, i find it a bit embarrassing for the CPMD 'community', that despite the many very smart people coming up with the algorithms, methods and codes, there is so little done to make them accessable to others. in that context it may be interesting to note, that i have been exposed to the CPMD code for less than five years and have been using it seriously and trying to understand the physics behind it for less than three years. JD> So, Dr. Axel thank you for putting up the tutorial page and I look forward JD> to visiting the page more often to take advantage of the more advance let me take this opportunity to thank those of you, who post questions on cpmd-list, since this is the place where most of the inspiration and motivation comes from. in retrospect i find it extremely amazing, how much one can learn simply by trying to answer questions... JD> features of CPMD as you post them. By the way are there any plans to give JD> a list of subroutines that users with guts to delve into the code and see JD> what CPMD is really doing? Together with the manual, one can get a clearer JD> picture on usage ... It's time consuming to do so right now since one has JD> to go guessing the appropriate routine. please have a look at the htmldoc.tar.gz (written by thierry deutsch) package in the contrib section of the cpmd.org download area. this contains a python script which will transform an existing CPMD source tree into a interconnected tree of html or xml pages, so you can browse through the sources and follow the flow of the code with a web browser. regards, axel. JD> JD> Ok. Thanks again Dr. Axel and all other who contribute to the tutorial JD> page. JD> JD> Regards from the sunny but freezing desert town of Nevada(?). JD> John JD> -- JD> **The first is not necessarily the Leader** JD> JD> "If I have spoken evil, bear witness of the evil: JD> but if well, why smitest thou me?" JD> -- Jesus Christ (John, 18:23) JD> _______________________________________________ JD> CPMD-list mailing list JD> CPMD-list at cpmd.org JD> http://cpmd.org/mailman/listinfo/cpmd-list JD> JD> -- ======================================================================= 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 frank at blue.cup.uni-muenchen.de Thu Feb 3 11:37:16 2005 From: frank at blue.cup.uni-muenchen.de (Irmgard Frank) Date: Thu, 3 Feb 2005 11:37:16 +0100 (CET) Subject: [CPMD-list] CPMD-list: 87. International Bunsen Discussion Meeting Message-ID: Dear Colleagues and Friends, we are happy to announce the 87. International Bunsen Discussion Meeting 'Mechanically Induced Chemistry: Theory and Experiment' which will take place in Tutzing near Munich, October 3 - 6, 2005. The meeting focusses on the nano-manipulation of molecular systems with AFM and STM techniques and on their theoretical investigation. Confirmed invited speakers are Frank Achenbach Vincent Croquette Harald Fuchs Helmut Grubmueller Michael Klein Dominik Marx Michele Parrinello Matthias Rief Klaus Schulten Gotthard Seifert Besides the invited talks, there will be room for contributed talks and posters. We are looking forward to receiving abstracts of your contributions (deadline: 1. 4. 2005). Tutzing is located south of Munich at the Starnberger See. The Evangelische Akademie resides in a beautiful chateau (Schloss Tutzing) that dates back to the 19th century (http://www.ev-akademie-tutzing.de, web pages in German). The number of participants is limited to 60, all of them can be accommodated in the comfortable and quiet rooms of the Schloss. For detailed information about the meeting please have a look at our webpage: http://www.chemie.uni-muenchen.de/pc/frank/bunsen Best regards, Irmgard Frank Andreas Zumbusch Department Chemie und Biochemie LMU Munich Butenandtstr. 11 81377 Munich Germany From iivanov at vitae.cmm.upenn.edu Thu Feb 3 20:51:39 2005 From: iivanov at vitae.cmm.upenn.edu (Ivaylo Ivanov) Date: Thu, 3 Feb 2005 14:51:39 -0500 (EST) Subject: [CPMD-list] linear response calculation Message-ID: <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> Hello, I'm trying to do vibrational analysis with the option LR (Linear Response). The manual says that "A rule of thumb is that good results are obtained with a 10 times stricter convergence on the ground state orbitals compared to that of the response orbitals." What is not clear to me is how to change the convergence criterion for the response orbitals from the default value (1.0D-5). Any suggestions would be appreciated. Thanks very much in advance. --Ivaylo Ivanov -- Ivaylo Ivanov Center for Molecular Modeling | University of Pennsylvania 231 S.34 Street, | Philadelphia,PA 19104-6323 tel: (215) 573 8697 | fax: (215) 573 6233 e-mail: iivanov at cmm.chem.upenn.edu From Ari.P.Seitsonen at iki.fi Fri Feb 4 11:29:29 2005 From: Ari.P.Seitsonen at iki.fi (Ari P Seitsonen) Date: Fri, 4 Feb 2005 11:29:29 +0100 (CET) Subject: [CPMD-list] linear response calculation In-Reply-To: <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> References: <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> Message-ID: Dear Ivaylo, Please have a look on the keyword 'CONVERGENCE' in the section '&RESP' (i.e. not in the section '&CPMD', where the criterion for convergence of the wave fucntions in the ground state is given; the syntax and meaning is similar, though). Greetings, apsi -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=- Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/ Tel +41 1 635 44 97 / Fax +41 1 635 68 38 / GSM +41 79 719 09 35 Anschrift: Physikalisch Chemisches Institut (PCI), Universit?t Z?rich (UniZh) Indirizzo: Winterthurerstrasse 190, CH-8057 Z?rich Address: Schweiz / Svizzera / Suisse / Svizra / Switzerland On Thu, 3 Feb 2005, Ivaylo Ivanov wrote: > Hello, > I'm trying to do vibrational analysis with the option LR (Linear > Response). The manual says that "A rule of thumb is that good results are > obtained with a 10 times stricter convergence on the ground state orbitals > compared to that of the response orbitals." What is not clear to me is how > to change the convergence criterion for the response orbitals from the > default value (1.0D-5). Any suggestions would be appreciated. Thanks very > much in advance. > > --Ivaylo Ivanov > > -- > Ivaylo Ivanov > Center for Molecular Modeling | University of Pennsylvania > 231 S.34 Street, | Philadelphia,PA 19104-6323 > tel: (215) 573 8697 | fax: (215) 573 6233 > e-mail: iivanov at cmm.chem.upenn.edu > _______________________________________________ > CPMD-list mailing list > CPMD-list at cpmd.org > http://cpmd.org/mailman/listinfo/cpmd-list > From wolfi at mittelerde.physik.uni-konstanz.de Fri Feb 4 12:10:07 2005 From: wolfi at mittelerde.physik.uni-konstanz.de (Wolfram Quester) Date: Fri, 4 Feb 2005 12:10:07 +0100 Subject: [CPMD-list] linear response calculation In-Reply-To: <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> References: <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> Message-ID: <20050204111006.GC21955@halley.zuhause> Hi Ivaylo, On Thu, Feb 03, 2005 at 02:51:39PM -0500, Ivaylo Ivanov wrote: > Hello, > I'm trying to do vibrational analysis with the option LR (Linear > Response). The manual says that "A rule of thumb is that good results are > obtained with a 10 times stricter convergence on the ground state orbitals > compared to that of the response orbitals." What is not clear to me is how > to change the convergence criterion for the response orbitals from the > default value (1.0D-5). Any suggestions would be appreciated. Thanks very > much in advance. I've never worked with linear response, but having a glance though the manual I guess this is what you need: CONVERGENCE Section: &LINRES Convergence criterion for linear response calculations. Default value is 10-5 This means that you have to put the CONVERGENCE keyword into the &LINRES...&END section of your input-file and give the convergence in the next line as in &LINRES CONVERGENCE 1.0.D-6 &END HTH, Wolfi -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 189 bytes Desc: Digital signature Url : http://cpmd.org/pipermail/cpmd-list/attachments/20050204/bd47a0e7/attachment.pgp From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Fri Feb 4 12:12:10 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Fri, 04 Feb 2005 12:12:10 +0100 Subject: [CPMD-list] linear response calculation In-Reply-To: Your message of "Thu, 03 Feb 2005 14:51:39 EST." <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> Message-ID: <200502041112.j14BCAf26435@yello.theochem.ruhr-uni-bochum.de> >>> "II" == Ivaylo Ivanov writes: II> Hello, II> I'm trying to do vibrational analysis with the option LR (Linear II> Response). The manual says that "A rule of thumb is that good results are hi, two remarks: 1.) VIBRATIONAL ANALYSIS LR is currently broken (and has been for quite some time). you can do a LINEAR RESPONSE calculation with &RESP PHONON &END instead. check out the vibanal subdirectory in the CPMD test suite. 2.) the convergence of the ground state is controlled with CONVERGENCE ORBITALS. the default value is good enough to start an MD, but you should use stricter convergence for phonon calculations. you should also check, whether you need to increase the plane wave cutoff. best regards, axel. II> obtained with a 10 times stricter convergence on the ground state orbitals II> compared to that of the response orbitals." What is not clear to me is how II> to change the convergence criterion for the response orbitals from the II> default value (1.0D-5). Any suggestions would be appreciated. Thanks very II> much in advance. II> --Ivaylo Ivanov II> -- II> Ivaylo Ivanov II> Center for Molecular Modeling | University of Pennsylvania II> 231 S.34 Street, | Philadelphia,PA 19104-6323 II> tel: (215) 573 8697 | fax: (215) 573 6233 II> e-mail: iivanov at cmm.chem.upenn.edu II> _______________________________________________ II> CPMD-list mailing list II> CPMD-list at cpmd.org II> http://cpmd.org/mailman/listinfo/cpmd-list -- ======================================================================= 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 hutter at pci.unizh.ch Fri Feb 4 13:07:24 2005 From: hutter at pci.unizh.ch (Juerg Hutter) Date: Fri, 4 Feb 2005 13:07:24 +0100 (MET) Subject: [CPMD-list] linear response calculation In-Reply-To: <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> References: <55166.132.239.16.199.1107460299.squirrel@cmm.upenn.edu> Message-ID: Hi this needs probably some explanations. There are three ways you can calculate second derivatives in CPMD 1) &CPMD VIBRATIONAL ANALYSIS FD &END --> Finite difference calculation This should work for all cases, needs the same specifications as a normal wavefunction optimization. Additional keywords are "FINITE DIFFERENCES" for the step length 2) &CPMD VIBRATIONAL ANALYSIS LR &END --> Linear response calculation This method has several known bugs, if you want to use it test it for your special type of calculations first. Additional input in the &LINRES section 3) &CPMD LINEAR RESPONSE &END &RESP PHONON &END The same method as 2) but different implementation. All additional input (like Convergence) through the &RESP section. General: Geometries should be optimized as good as possible, with lower convergence criteria for both, geometry and wfn. regards Juerg ---------------------------------------------------------- Juerg Hutter Phone : ++41 1 635 4491 Physical Chemistry Institute FAX : ++41 1 635 6838 University of Zurich E-mail: hutter at pci.unizh.ch Winterthurerstrasse 190 CH-8057 Zurich, Switzerland ---------------------------------------------------------- On Thu, 3 Feb 2005, Ivaylo Ivanov wrote: > Hello, > I'm trying to do vibrational analysis with the option LR (Linear > Response). The manual says that "A rule of thumb is that good results are > obtained with a 10 times stricter convergence on the ground state orbitals > compared to that of the response orbitals." What is not clear to me is how > to change the convergence criterion for the response orbitals from the > default value (1.0D-5). Any suggestions would be appreciated. Thanks very > much in advance. > > --Ivaylo Ivanov > > -- > Ivaylo Ivanov > Center for Molecular Modeling | University of Pennsylvania > 231 S.34 Street, | Philadelphia,PA 19104-6323 > tel: (215) 573 8697 | fax: (215) 573 6233 > e-mail: iivanov at cmm.chem.upenn.edu > _______________________________________________ > CPMD-list mailing list > CPMD-list at cpmd.org > http://cpmd.org/mailman/listinfo/cpmd-list > From ZRLSRB at ch.ibm.com Fri Feb 4 15:22:12 2005 From: ZRLSRB at ch.ibm.com (Salomon Billeter) Date: Fri, 4 Feb 2005 15:22:12 +0100 Subject: [CPMD-list] linear response calculation Message-ID: Hi, > 1.) VIBRATIONAL ANALYSIS LR is currently broken (and has been for > quite some time). you can do a LINEAR RESPONSE calculation with > &RESP PHONON &END instead. check out the vibanal subdirectory > in the CPMD test suite. The default AUTOMATIC optimizer (switching between optimizers depending on the gradient) will run into troubles. However, if you are selecting the DIIS or the PCG optimizers to solve the LR equations, the code will run just fine and produce reasonable results in many cases. It's a good idea anyway to check and to add a section &LINRES ... &END instead of going with the defaults (like in the test examples methan*). I'll look into the problems of AUTOMATIC when I find the time. Best regards, Salomon From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Fri Feb 4 16:07:35 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Fri, 04 Feb 2005 16:07:35 +0100 Subject: [CPMD-list] linear response calculation In-Reply-To: Your message of "Fri, 04 Feb 2005 15:22:12 +0100." Message-ID: <200502041507.j14F7Zp11728@yello.theochem.ruhr-uni-bochum.de> >>> "SB" == Salomon Billeter writes: SB> Hi, >> 1.) VIBRATIONAL ANALYSIS LR is currently broken (and has been for >> quite some time). you can do a LINEAR RESPONSE calculation with >> &RESP PHONON &END instead. check out the vibanal subdirectory >> in the CPMD test suite. SB> The default AUTOMATIC optimizer (switching between optimizers depending on SB> the gradient) will run into troubles. However, if you are selecting the SB> DIIS or the PCG optimizers to solve the LR equations, the code will run SB> just fine and produce reasonable results in many cases. It's a good idea SB> anyway to check and to add a section &LINRES ... &END instead of going with SB> the defaults (like in the test examples methan*). I'll look into the dear salomon, i beg to differ here. defaults should be set to something that works, even if it is not very efficient. if one cannot produce reasonable defaults, there should be a test in the code demanding additional input (e.g. a &LINRES section in the current example) and specific instructions or at least some hints in the manual. how else should somebody find this out? there are far too many documented (and undocumented) flags and options to try them out systematically and in this case there have been no examples available. best regards, axel. SB> problems of AUTOMATIC when I find the time. SB> Best regards, SB> Salomon SB> _______________________________________________ SB> CPMD-list mailing list SB> CPMD-list at cpmd.org SB> http://cpmd.org/mailman/listinfo/cpmd-list -- ======================================================================= 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 ZRLSRB at ch.ibm.com Fri Feb 4 16:44:22 2005 From: ZRLSRB at ch.ibm.com (Salomon Billeter) Date: Fri, 4 Feb 2005 16:44:22 +0100 Subject: [CPMD-list] linear response calculation Message-ID: Dear Axel, > i beg to differ here. defaults should be set to something that > works, even if it is not very efficient. if one cannot produce > reasonable defaults, there should be a test in the code demanding > additional input (e.g. a &LINRES section in the current example) > and specific instructions or at least some hints in the manual. you are right, we don't differ here. Defaults should be sensible. That's why I wrote I'll look into the problem when I have the time. Best regards, Salomon From ZRLSRB at ch.ibm.com Mon Feb 7 12:33:55 2005 From: ZRLSRB at ch.ibm.com (Salomon Billeter) Date: Mon, 7 Feb 2005 12:33:55 +0100 Subject: [CPMD-list] linear response calculation Message-ID: Hi, appended is the promised bug fix to make the default automatic LR wavefunction optimizer work as well. It works for some test systems, but there might be more bugs. Best regards, Salomon (See attached file: opt_lr.F.patch)(See attached file: lr_upd.F.patch) -------------- next part -------------- A non-text attachment was scrubbed... Name: opt_lr.F.patch Type: application/octet-stream Size: 681 bytes Desc: not available Url : http://cpmd.org/pipermail/cpmd-list/attachments/20050207/65716d4e/attachment.obj -------------- next part -------------- A non-text attachment was scrubbed... Name: lr_upd.F.patch Type: application/octet-stream Size: 2239 bytes Desc: not available Url : http://cpmd.org/pipermail/cpmd-list/attachments/20050207/65716d4e/attachment-0001.obj From J.Schuele at TU-BS.de Mon Feb 7 14:04:28 2005 From: J.Schuele at TU-BS.de (=?ISO-8859-1?Q?=22J=2E_Sch=FCle=22?=) Date: Mon, 07 Feb 2005 14:04:28 +0100 Subject: [CPMD-list] cpmd2cube error Message-ID: <4207675C.60209@TU-BS.de> Hi, got cpmd and cpmd2cube and tried to run the B2H6 example given in the properties directory. Unfortunately cpmd2cube shows a fortran read error on the ELF-File. I scanned through the sources and found cpmd2cube wants to read a FORMATTED file, while CMPD writes a UNFORMATTED file. Any comments and solutions for this? Josef Sch?le From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Feb 7 14:45:20 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 7 Feb 2005 14:45:20 +0100 (CET) Subject: [CPMD-list] cpmd2cube error In-Reply-To: <4207675C.60209@TU-BS.de> Message-ID: On Mon, 7 Feb 2005, "J. Sch?le" wrote: JS> Hi, JS> got cpmd and cpmd2cube and tried to run the B2H6 example given in the JS> properties directory. Unfortunately cpmd2cube shows a fortran read error JS> on the ELF-File. JS> I scanned through the sources and found cpmd2cube wants to read a JS> FORMATTED file, while CMPD writes a UNFORMATTED file. JS> JS> Any comments and solutions for this? could you please state what OS/compiler combination you are using and please also post the error message. are you using the same combination for writing the ELF file? AFAICT, all of the FORMATTED opens are for writing. best regards, axel. JS> JS> Josef Sch?le JS> JS> _______________________________________________ JS> CPMD-list mailing list JS> CPMD-list at cpmd.org JS> http://cpmd.org/mailman/listinfo/cpmd-list JS> JS> -- ======================================================================= 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 Ari.P.Seitsonen at iki.fi Mon Feb 7 14:54:22 2005 From: Ari.P.Seitsonen at iki.fi (Ari P Seitsonen) Date: Mon, 7 Feb 2005 14:54:22 +0100 (CET) Subject: [CPMD-list] cpmd2cube error In-Reply-To: <4207675C.60209@TU-BS.de> References: <4207675C.60209@TU-BS.de> Message-ID: Dear Josef, cpmd2cube should read UNFORMATTED files. That's why you might have problems if you transfer files between different architectures, that is, you run CPMD and cpmd2cube in a different architectures. At least in the options for the PGI compiler you can control the format of floating point numbers in the I/O ('byteswapio' or similar); so if you're for example running both CPMD and cpmd2cube in a linux machine you should make sure that you use those options consistently, either having them in both or neither Makefile. Another solution for your problem might be the option '-double'. Hopefully this helps. Greetings from Paris, apsi -=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=- Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/ Tel +41 1 635 44 97 / Fax +41 1 635 68 38 / GSM +41 79 719 09 35 Anschrift: Physikalisch Chemisches Institut (PCI), Universit?t Z?rich (UniZh) Indirizzo: Winterthurerstrasse 190, CH-8057 Z?rich Address: Schweiz / Svizzera / Suisse / Svizra / Switzerland On Mon, 7 Feb 2005, "J. Sch?le" wrote: > Hi, > got cpmd and cpmd2cube and tried to run the B2H6 example given in the > properties directory. Unfortunately cpmd2cube shows a fortran read error > on the ELF-File. > I scanned through the sources and found cpmd2cube wants to read a > FORMATTED file, while CMPD writes a UNFORMATTED file. > > Any comments and solutions for this? > > Josef Sch?le > > _______________________________________________ > CPMD-list mailing list > CPMD-list at cpmd.org > http://cpmd.org/mailman/listinfo/cpmd-list > From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Feb 7 15:18:04 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 7 Feb 2005 15:18:04 +0100 (CET) Subject: [CPMD-list] cpmd2cube error In-Reply-To: <4207675C.60209@TU-BS.de> Message-ID: On Mon, 7 Feb 2005, "J. Sch?le" wrote: JS> Hi, JS> got cpmd and cpmd2cube and tried to run the B2H6 example given in the JS> properties directory. Unfortunately cpmd2cube shows a fortran read error JS> on the ELF-File. JS> I scanned through the sources and found cpmd2cube wants to read a JS> FORMATTED file, while CMPD writes a UNFORMATTED file. JS> JS> Any comments and solutions for this? please compare your cpmd and cpmd2cube makefiles. i just saw, that some kind soul - who probably does his postprocessing on a big endian machine - had put a '-convert big_endian' into the configuration for DEC-ALPHA for cpmd but not for cpmd2cube. hhmmmm. please make sure that you compile _both_ programs either with or without that flag. regards, axel. JS> JS> Josef Sch?le JS> JS> _______________________________________________ JS> CPMD-list mailing list JS> CPMD-list at cpmd.org JS> http://cpmd.org/mailman/listinfo/cpmd-list JS> JS> -- ======================================================================= 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 ara_1357_2416 at yahoo.com Mon Feb 7 16:56:48 2005 From: ara_1357_2416 at yahoo.com (Younes Ansari) Date: Mon, 7 Feb 2005 07:56:48 -0800 (PST) Subject: [CPMD-list] from younes ansari Message-ID: <20050207155648.54181.qmail@web50904.mail.yahoo.com> Dear CPMD users I have a very simple question: once a CPMD run is through, the trajectories in VMD environment look to have extra line as if some atoms are bonded by more than their bonds by more than their valance bond. For example, an H atom is bonded to other three C atoms. This is the case for other elements. The input file representing the molecule is originally made in HYPERCHEM, and then operated for appropriate coordinate in GAUSSIAN 98 where it allows the input file for CPMD be formatted suitably. It might be due by the representation and not really extra bonds are involved. Could any of you give comment on this! --------------------------------- Do you Yahoo!? Yahoo! Search presents - Jib Jab's 'Second Term' -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20050207/3a73ed88/attachment.html From ara_1357_2416 at yahoo.com Mon Feb 7 16:57:10 2005 From: ara_1357_2416 at yahoo.com (Younes Ansari) Date: Mon, 7 Feb 2005 07:57:10 -0800 (PST) Subject: [CPMD-list] from younes ansari Message-ID: <20050207155710.30111.qmail@web50902.mail.yahoo.com> Dear CPMD users I have a very simple question: once a CPMD run is through, the trajectories in VMD environment look to have extra line as if some atoms are bonded by more than their bonds by more than their valance bond. For example, an H atom is bonded to other three C atoms. This is the case for other elements. The input file representing the molecule is originally made in HYPERCHEM, and then operated for appropriate coordinate in GAUSSIAN 98 where it allows the input file for CPMD be formatted suitably. It might be due by the representation and not really extra bonds are involved. Could any of you give comment on this! --------------------------------- Do you Yahoo!? Yahoo! Search presents - Jib Jab's 'Second Term' -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20050207/8b2e75c0/attachment.html From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Feb 7 17:53:14 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 7 Feb 2005 17:53:14 +0100 (CET) Subject: [CPMD-list] from younes ansari In-Reply-To: <20050207155648.54181.qmail@web50904.mail.yahoo.com> Message-ID: On Mon, 7 Feb 2005, Younes Ansari wrote: YA> Dear CPMD users dear younes ansari, first off a few remarks. - please use a subject line that has some relevance to the question you ask. this makes it easier to follow a thread. we already know from the From: line, that you have sent the mail. - there is no need to send your mail twice, and also using HTML encoding is at best redundant and annoying to others. please see: http://expita.com/nomime.html - in a previous mail i have already pointed you to: http://www.catb.org/~esr/faqs/smart-questions.html please read it again and follow the advice, if you want to get useful answers. it is extremely difficult to guess the (vital) parts, that you are leaving out, so chances are, that people will ignore you, or - even worse - give bad advice. YA> I have a very simple question: once a CPMD run is through, the YA> trajectories in VMD environment look to have extra line as if some YA> atoms are bonded by more than their bonds by more than their valance bond. YA> For example, an H atom is bonded to other three C atoms. This is the case YA> for other elements. please give more details here and post a (short) example. if i assume that you are loading a GEOMETRY.xyz or TRAJEC.xyz file generated by CPMD into the VMD visualization program (note, that people here may know CPMD, but not necessarily VMD), then wrong bonding like you describe it, most likely originates from providing coordinates in angstrom while not using the ANGSTROM keyword. please note, that VMD does not recompute the bonding when visualizing a trajectory unless you use the Dynamic Bonds representation (or something more sophisticated). you can find an example at: http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/cpmd-vmd/part2.html#chap4_sect5 YA> The input file representing the molecule is originally made in YA> HYPERCHEM, and then operated for appropriate coordinate in GAUSSIAN 98 YA> where it allows the input file for CPMD be formatted suitably. i am sorry, but i cannot follow you here. how should a run with gaussian help you to get correctly formatted CPMD input? can you please explain the procedure in more detail. YA> YA> It might be due by the representation and not really extra bonds are YA> involved. Could any of you give comment on this! as far as VMD visualization is concerned, the VMD mailing list may be a better place to discuss that. but please first make sure, that your CPMD input is correct. a finished CPMD run without any error messages is a prerequisite for that but not a sufficient indicator. best regards, axel kohlmeyer. YA> YA> YA> YA> --------------------------------- YA> Do you Yahoo!? YA> Yahoo! Search presents - Jib Jab's 'Second Term' -- ======================================================================= 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 bob at bob.usuhs.mil Tue Feb 8 16:50:42 2005 From: bob at bob.usuhs.mil (Robert Williams) Date: Tue, 08 Feb 2005 10:50:42 -0500 Subject: [CPMD-list] atomic units: Hartrees or Rydbergs? Message-ID: <4208DFD2.50304@bob.usuhs.mil> Dear list, A.U. seems to mean different things to different folks. I assume in CPMD that energy units are Hartrees, but I see that the energy cutoff units are Rydbergs. Also, in a previous post I noted that the value for the van der Waals strength (alpha) given in the contrib vdw examples was a factor of ten greater than the values provided by the literature reference (after conversion to hartrees). So... Just to be sure, what are the units of energy used internally? Bob -- Dr. Robert Williams Department of Biomedical Informatics Uniformed Services University of the Health Sciences 301-295-3568 From hutter at pci.unizh.ch Tue Feb 8 19:29:43 2005 From: hutter at pci.unizh.ch (Juerg Hutter) Date: Tue, 8 Feb 2005 19:29:43 +0100 (MET) Subject: [CPMD-list] atomic units: Hartrees or Rydbergs? In-Reply-To: <4208DFD2.50304@bob.usuhs.mil> References: <4208DFD2.50304@bob.usuhs.mil> Message-ID: Hi CPMD uses Hartree as internal units. The input in Rydbergs for the PW cutoff is only for historical reasons. regards Juerg Hutter ---------------------------------------------------------- Juerg Hutter Phone : ++41 1 635 4491 Physical Chemistry Institute FAX : ++41 1 635 6838 University of Zurich E-mail: hutter at pci.unizh.ch Winterthurerstrasse 190 CH-8057 Zurich, Switzerland ---------------------------------------------------------- On Tue, 8 Feb 2005, Robert Williams wrote: > Dear list, > > A.U. seems to mean different things to different folks. > I assume in CPMD that energy units are Hartrees, > but I see that the energy cutoff units are Rydbergs. > Also, in a previous post I noted that the value > for the van der Waals strength (alpha) given in the > contrib vdw examples was a factor of ten greater than > the values provided by the literature reference > (after conversion to hartrees). So... > > Just to be sure, what are the units of energy used internally? > > Bob > -- > Dr. Robert Williams > Department of Biomedical Informatics > Uniformed Services University of the Health Sciences > 301-295-3568 > _______________________________________________ > CPMD-list mailing list > CPMD-list at cpmd.org > http://cpmd.org/mailman/listinfo/cpmd-list > From langel at uni-greifswald.de Tue Feb 8 20:19:57 2005 From: langel at uni-greifswald.de (Walter Langel) Date: Tue, 08 Feb 2005 20:19:57 +0100 Subject: [CPMD-list] cpmd2cube on SunFire Message-ID: <420910DD.8030409@uni-greifswald.de> Hi, has someone compiled the cpmd2cube program on a SunFire architecture and would be so kind to send me the options for Configure or the Makefile. I understand that the program has to be recompiled on every architecture used, since it works on binaries, and that the current Configure does not contain much Sun. Thanks Regards Walter Langel -- Prof. Dr. Walter Langel Institut fuer Chemie und Biochemie Universitaet Greifswald Soldmannstrasse 23 D-17487 Greifswald Germany Tel +49 3834 86 4423 Fax +49 3834 86 4413 http://www.chemie.uni-greifswald.de/~plasma From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Wed Feb 9 09:44:28 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Wed, 9 Feb 2005 09:44:28 +0100 (CET) Subject: [CPMD-list] cpmd2cube on SunFire In-Reply-To: <420910DD.8030409@uni-greifswald.de> Message-ID: On Tue, 8 Feb 2005, Walter Langel wrote: WL> Hi, hi! WL> has someone compiled the cpmd2cube program on a SunFire architecture and WL> would be so kind to send me the options for Configure or the Makefile. yes. please try the attached makefile. you may also want to try to update your cpmd2cube with the attached patch. i am not 100% certain, whether this applies cleanly to the latest release from cpmd.org, but there should be new release on cpmd.org any day now. to 'port' cpmd2cube to a new platform, you usually only need to use the same preprocessor directives compilation like for a serial cpmd executable and then use a FFLAGS definition, that selects compilation of free format fortran90 with preprocessing. WL> I understand that the program has to be recompiled on every architecture WL> used, since it works on binaries, and that the current Configure does WL> not contain much Sun. only if your binary format is not IEEE754 compliant. you can usually read binary data from a sun, e.g. on a linux pc when you either compile for big endian binary i/o or use on-the-fly conversion. for more details, please see: http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/cpmd-linux.html#endian best regards, axel. WL> WL> Thanks WL> WL> Regards WL> WL> Walter Langel WL> WL> -- ======================================================================= 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. -------------- next part -------------- #-----------------------------------------------------------------------------# # Makefile for cpmd2cube.x # Configuration: SUN-ULTRA3-32bit # Creation of Makefile: Feb 9 2005 # on Linux monteverdi.theochem.ruhr-uni-bochum.de 2.4.20-20.7 #1 Mon Aug 18 15:05:54 EDT 2003 i686 unknown # Author: monteverdi.theochem.ruhr-uni-bochum.de!akohlmey #---------------------------------------------------------------------------- # SHELL = /bin/sh # #--------------- Configuration FC = f90 FFLAGS = -D__SUN -DFFT_DEFAULT -fast -free -fns=no -xtarget=ultra3 -xarch=v8plusb LFLAGS = -xtarget=ultra3 -xarch=v8plusb -fast LIBS = -xlibmopt # uncomment and set only if needed #CC = cc #CFLAGS = #--------------- End of Configuration ############## EXE = cpmd2cube.x OBJECTS = \ kinds.o periodic_table.o util.o grids.o cpmd2cube.o \ main.o grid_types.o atom_types.o readwrites.o \ fftsg_lib.o fft_tools.o fftsg.o CFGDEST = . CFGMACH = SUN-ULTRA3-32bit SRC=. ############################# all: $(EXE) trimcube $(EXE): $(OBJECTS) $(FC) -o $(EXE) $(LFLAGS) $(OBJECTS) $(LIBS) trimcube: trimcube.o $(CC) $(CFLAGS) -o $@ trimcube.o ############################# clean: rm -f *.o *.mod *.L *~ # Create a new makefile with new dependencies but retain the configuration. reconfig: ( cd $(SRC) ; ./Configure -r -m -SRC=$(SRC) -DEST=$(CFGDEST) $(CFGMACH) ; cd $(CFGDEST) ) ############################# atom_types.o: $(SRC)/atom_types.F $(FC) -c $(FFLAGS) $(SRC)/$*.F cpmd2cube.o: $(SRC)/cpmd2cube.F $(FC) -c $(FFLAGS) $(SRC)/$*.F fft_tools.o: $(SRC)/fft_tools.F $(FC) -c $(FFLAGS) $(SRC)/$*.F grid_types.o: $(SRC)/grid_types.F $(FC) -c $(FFLAGS) $(SRC)/$*.F grids.o: $(SRC)/grids.F $(FC) -c $(FFLAGS) $(SRC)/$*.F main.o: $(SRC)/main.F $(FC) -c $(FFLAGS) $(SRC)/$*.F periodic_table.o: $(SRC)/periodic_table.F $(FC) -c $(FFLAGS) $(SRC)/$*.F readwrites.o: $(SRC)/readwrites.F $(FC) -c $(FFLAGS) $(SRC)/$*.F util.o: $(SRC)/util.F $(FC) -c $(FFLAGS) $(SRC)/$*.F kinds.o: $(SRC)/kinds.F $(FC) -c $(FFLAGS) $(SRC)/$*.F fftsg.o: $(SRC)/fftsg.F $(FC) -c $(FFLAGS) $(SRC)/$*.F fftsg_lib.o: $(SRC)/fftsg_lib.F $(FC) -c $(FFLAGS) $(SRC)/$*.F trimcube.o: $(SRC)/trimcube.c $(CC) -c $(CFLAGS) $(SRC)/$*.c # Dependencies atom_types.o: kinds.o \ periodic_table.o cpmd2cube.o: atom_types.o \ fft_tools.o \ grid_types.o \ grids.o \ kinds.o \ periodic_table.o \ readwrites.o \ util.o fft_tools.o: grid_types.o \ fftsg_lib.o \ kinds.o grid_types.o: kinds.o grids.o: grid_types.o \ kinds.o \ util.o main.o: cpmd2cube.o periodic_table.o: kinds.o readwrites.o: atom_types.o \ grid_types.o \ kinds.o util.o: kinds.o -------------- next part -------------- A non-text attachment was scrubbed... Name: cp2c-update.diff.gz Type: application/x-gzip Size: 14795 bytes Desc: Url : http://cpmd.org/pipermail/cpmd-list/attachments/20050209/cb744067/attachment.gz From t.laino at sns.it Thu Feb 10 18:55:16 2005 From: t.laino at sns.it (t.laino at sns.it) Date: Thu, 10 Feb 2005 11:55:16 -0600 Subject: [CPMD-list] Mail Delivery (failure cpmd-list@cpmd.org) Message-ID: <200502101806.TAA05044@internet-fence.zurich.ihost.com> -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20050210/a0023be2/attachment.html -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: audio/x-wav Size: 29568 bytes Desc: not available Url : http://cpmd.org/pipermail/cpmd-list/attachments/20050210/a0023be2/attachment.wav From g0306324 at nus.edu.sg Fri Feb 11 09:02:28 2005 From: g0306324 at nus.edu.sg (Dai Ling) Date: Fri, 11 Feb 2005 16:02:28 +0800 Subject: [CPMD-list] artificial atoms Message-ID: <396AEE87751E5A498D2F422E3A9FF68BCC1E3B@MBOX23.stu.nus.edu.sg> Dear Sirs: Is it possible to create artificial atoms in CPMD? For example, I want to cut a B-terminated surface from cubic BN bulk, the bottom layer is N-terminated which will be saturated with hydrogen atoms (H). As BN surface is polar surface, so I hope the H atoms for situration have 3/4 electrons and the mass is 20. Could you please kindly advise us how to implement it? Thanks! Scott From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Fri Feb 11 12:12:32 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Fri, 11 Feb 2005 12:12:32 +0100 (CET) Subject: [CPMD-list] artificial atoms In-Reply-To: <396AEE87751E5A498D2F422E3A9FF68BCC1E3B@MBOX23.stu.nus.edu.sg> Message-ID: On Fri, 11 Feb 2005, Dai Ling wrote: DL> Dear Sirs: dear scott, DL> Is it possible to create artificial atoms in CPMD? For example, I want DL> to cut a B-terminated surface from cubic BN bulk, the bottom layer is DL> N-terminated which will be saturated with hydrogen atoms (H). As BN DL> surface is polar surface, so I hope the H atoms for situration have 3/4 DL> electrons and the mass is 20. Could you please kindly advise us how to DL> implement it? changing the mass is easily done via the ISOTOPE keyword. to get the 3/4 electrons, you'd have to create a proper pseudopotential with the corresponding core charge (as this is where CPMD takes the number of electrons from) _and_ either wait for the release of CPMD v3.9.2 or get back to me for a patch to CPMD to add support for pseudopotentials with fractional core charges (with the currently available versions, core charges are rounded). AFAICT, the total number of electrons has to be integer. regards, axel. DL> DL> Thanks! DL> DL> Scott DL> _______________________________________________ DL> CPMD-list mailing list DL> CPMD-list at cpmd.org DL> http://cpmd.org/mailman/listinfo/cpmd-list DL> DL> -- ======================================================================= 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 chokosabe at yahoo.co.uk Sat Feb 12 12:02:39 2005 From: chokosabe at yahoo.co.uk (Abraham Alaka) Date: Sat, 12 Feb 2005 11:02:39 +0000 (GMT) Subject: [CPMD-list] Error message during PIMD with Vanderbilt pseudopotentials Message-ID: <20050212110240.77565.qmail@web25704.mail.ukl.yahoo.com> Hi all, I was running a pimd calculation and I ended up with the error message; *** PI_MDPT| THE NEW SIZE OF THE PROGRAM IS 28640/ 173648 kBYTES *** RV30| WARNING! NO WAVEFUNCTION VELOCITIES RESTART INFORMATION READ ON FILE ./RESTART_1.1 RESTART INFORMATION READ ON FILE ./RESTART_2.1 RESTART INFORMATION READ ON FILE ./RESTART_3.1 RESTART INFORMATION READ ON FILE ./RESTART_4.1 RESTART INFORMATION READ ON FILE ./RESTART_5.1 RESTART INFORMATION READ ON FILE ./RESTART_6.1 RESTART INFORMATION READ ON FILE ./RESTART_7.1 RESTART INFORMATION READ ON FILE ./RESTART_8.1 DEGREES OF FREEDOM FOR SYSTEM: 39 QMDOF| NUMBER OF QUANTUM DEGREES OF FREEDOM FOR IP=1 : 42 QMDOF| NUMBER OF QUANTUM DEGREES OF FREEDOM FOR IP>1 : 42 *** PHFAC| THE NEW SIZE OF THE PROGRAM IS 60080/ 193220 kBYTES *** EWALD| SUM IN REAL SPACE OVER 11*11*11 CELLS *************************************************************** * CHARACTERISTIC RADII * *************************************************************** SPECIES R OF GYRATION FREE PARTICLE (IN ANGSTROM) 1 Fe 0.02990 0.02690 2 O 0.05414 0.05027 R OF SUSCEPTIBILITY 1 Fe 0.05009 0.03805 2 O 0.09932 0.07109 R OF CORRELATION 1 Fe 0.05292 0.04660 2 O 0.09450 0.08706 *************************************************************** FILE ENERGIES EXISTS, NEW DATA WILL BE APPENDED *** PI_MD| THE NEW SIZE OF THE PROGRAM IS 148508/ 194092 kBYTES *** CPU TIME FOR INITIALIZATION 139.06 SECONDS RORTHO:DIFGAM= ??????????????????????? ITER = 31 > 30 MAXIMUM NUMBER OF ITERATIONS EXCEEDED STOPGM! STACK OF MAIN CALLS: STOPGM! CALL POSUPA STOPGM! CALL RORTOG PROGRAM STOPS IN SUBROUTINE RORTOG| The input file is below. (Sorry about the length of this post). ****************************************************************************** ** INPUT FILE ** ****************************************************************************** ** &CPMD ** ** MIRROR ** ** LSD ** ** RHOOUT ** ** PRINT ON FORCES WANNIER ** ** 100 ** ** PATH INTEGRALS ** ** MOLECULAR DYNAMICS CP ** ** PCG ** ** lANCZOS dIAGONALIZATION aLL ** ** RESTART LATEST ** ** RESTART WAVEFUNCTION COORDINATES vELOCITIES ** ** TRAJECTORY SAMPLE ** ** 90 ** ** PARRINELLO-RAHMAN NPT ** ** TEMPERATURE ** ** 270 ** ** TEMPCONTROL IONS ** ** 300 10 ** ** nOSE iONS mASSIVE ** ** 900 29311.d0 ** ** nOSE eLECTRONS mASSIVE ** ** 9.5 150000 ** ** SPLINE POINTS ** ** 3000 ** ** TIMESTEP ** ** 9.0 ** ** MAXSTEP ** ** 1000 ** ** DIPOLE DYNAMICS SAMPLE WANNIER ** ** 100 ** ** &END ** ** ** ** ** ** ** ** ** ** &SYSTEM ** ** MULTIPLICITY ** ** 2 ** ** ANGSTROM ** ** CHARGE ** ** 1 ** ** CELL VECTORS ** ** 0.0 4.1948636 4.1948636 4.1948636 0.0 4.1948636 4.1948636 4.1948636 0.0 ** ** CUTOFF ** ** 35 ** ** TESR ** ** 5 ** ** &END ** ** ** ** &ATOMS ** ** *026-Fe-gpbe-nlcc-bm.uspp BINARY NLCC ** ** LMAX=D ** ** 6 ** ** 1.0487159095 1.0487159095 1.0487159095 ** ** -1.0487159095 -1.0487159095 -1.0487159095 ** ** 4.1948636484 4.1948636484 4.1948636484 ** ** 4.1948636484 2.0974318242 2.0974318242 ** ** 2.0974318242 4.1948636484 2.0974318242 ** ** 2.0974318242 2.0974318242 4.1948636484 ** ** ** ** *008-O-gpbe--bm.uspp BINARY ** ** LMAX=P ** ** 8 ** ** 2.1346484286 2.1346484286 2.1346484286 ** ** 2.1346484286 -0.0372166056 -0.0372166056 ** ** -0.0372166056 2.1346484286 -0.0372166056 ** ** -0.0372166056 -0.0372166056 2.1346484286 ** ** -2.1346484286 -2.1346484286 -2.1346484286 ** ** -2.1346484286 0.0372166056 0.0372166056 ** ** 0.0372166056 -2.1346484286 0.0372166056 ** ** 0.0372166056 0.0372166056 -2.1346484286 ** ** &END ** ** ** ** &DFT ** ** FUNCTIONAL PBE ** ** GC-CUTOFF ** ** 5.D-6 &END ** ** ** ** ** ** &PIMD ** ** TROTTER DIMENSION ** ** 8 ** ** NORMAL MODES ** ** 1.d0 ** ** FACMASS ** ** 1.d0 ** ** DEBROGLIE CENTROID ** ** 300.d0 ** ** PRINT LEVEL ** ** 1 ** ** &END ** ** ** ****************************************************************************** ****************************************************************************** ___________________________________________________________ ALL-NEW Yahoo! Messenger - all new features - even more fun! http://uk.messenger.yahoo.com From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Sat Feb 12 21:34:59 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Sat, 12 Feb 2005 21:34:59 +0100 Subject: [CPMD-list] Error message during PIMD with Vanderbilt pseudopotentials In-Reply-To: Your message of "Sat, 12 Feb 2005 11:02:39 GMT." <20050212110240.77565.qmail@web25704.mail.ukl.yahoo.com> Message-ID: <200502122035.j1CKZ0h10059@yello.theochem.ruhr-uni-bochum.de> >>> "AA" == Abraham Alaka writes: AA> Hi all, [...] AA> *************************************************************** AA> FILE ENERGIES EXISTS, NEW DATA WILL BE APPENDED AA> *** PI_MD| THE NEW SIZE OF THE PROGRAM IS AA> 148508/ 194092 kBYTES *** AA> CPU TIME FOR INITIALIZATION 139.06 SECONDS AA> RORTHO:DIFGAM= ??????????????????????? ITER = AA> 31 > 30 AA> MAXIMUM NUMBER OF ITERATIONS EXCEEDED AA> STOPGM! STACK OF MAIN CALLS: AA> STOPGM! CALL POSUPA AA> STOPGM! CALL RORTOG AA> PROGRAM STOPS IN SUBROUTINE RORTOG| could you please check, whether you get the same error with a closed shell PI-calculation and/or a regular (non-PI) CPMD job. there were some problems with open shell calculations using the PBE functional, that were found (and hopefully fixed) only a few weeks ago. if the closed shell run works, but the non-PI fails, i can provide a patch that should fix the problem. regards, axel. AA> The input file is below. (Sorry about the length of AA> this post). [...] -- ======================================================================= 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 g0403127 at nus.edu.sg Mon Feb 14 08:23:54 2005 From: g0403127 at nus.edu.sg (Hong Won Keon) Date: Mon, 14 Feb 2005 15:23:54 +0800 Subject: [CPMD-list] how to run ld1 programme Message-ID: <9E4557B19235FE47B1C7B52DE3A536A33E046B@MBOX23.stu.nus.edu.sg> Hi, everybody, I'm beginner in CPMD. Now I'm trying to get psedopotential of Pt. So I downloaded a package from the site below: http://leopardi.cmp.sns.it/~giannozz/ As a first step, I put information in "ld1.in" file as follow: &input atom = 'Pt' config = '[Xe] 4f14.0 5d9.0 6s1.0' dft = 'pw91', file_wavefunctions = 'pts1d9.wfc' &end However, I cannot find out how to run this program - ld1 - in manual. I think the answer may be short and simple. Please help me with this trivial problem. Thanks in advance. W.K. Hong -------------- next part -------------- An HTML attachment was scrubbed... URL: http://cpmd.org/pipermail/cpmd-list/attachments/20050214/5cec5ad2/attachment.html From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Feb 14 09:38:18 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 14 Feb 2005 09:38:18 +0100 (CET) Subject: [CPMD-list] how to run ld1 programme In-Reply-To: <9E4557B19235FE47B1C7B52DE3A536A33E046B@MBOX23.stu.nus.edu.sg> Message-ID: On Mon, 14 Feb 2005, Hong Won Keon wrote: WKH> Hi, everybody, hello! WKH> WKH> I'm beginner in CPMD. WKH> WKH> Now I'm trying to get psedopotential of Pt. first off, as a beginner in using CPMD you probably don't want to start with some of the more trickier stuff like generating a pseudopotential for a transition metal. there are two ready-to-use LDA Pt pseudopotentials (one 'regular' and one 'semi-core' in the goedecker library from the contrib section in www.cpmd.org). WKH> So I downloaded a package from the site below: WKH> WKH> http://leopardi.cmp.sns.it/~giannozz/ please note, that this package is (almost) obsolete. an updated and extended version of this code is distributed with the PWscf (a.k.a quantum-ESPRESSO) package via http://www.pwscf.org/. please also check out http://www.fhi-berlin.mpg.de/th/fhi98md/fhi98PP/ which comes with a very good tutorial on pseudopotential generation. to convert the resulting pseudopotential files to format that is readable by CPMD you can use this program: http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/file/fhi2cpmd.c WKH> As a first step, I put information in "ld1.in" file as follow: WKH> WKH> &input WKH> WKH> atom = 'Pt' config = '[Xe] 4f14.0 5d9.0 6s1.0' WKH> WKH> dft = 'pw91', file_wavefunctions = 'pts1d9.wfc' WKH> WKH> &end WKH> WKH> However, I cannot find out how to run this program - ld1 - in manual. like with most (scientific) codes, if it is not documented, you have to read the source code. ld1 reads data from standard input (see file ld1.f, line 139) so you have to use i/o redirection, e.g, ld1 < ld1.in please see the attached shell script for an example of the whole PP generation procedure using paolo gianozzi's package (please don't use the resulting pseudopotential, it has not been tested, this is just a small example). regards, axel kohlmeyer. p.s.: please don't use HTML encoded mails (especially not those generated by Microsoft Word) on mailing lists. see http://expita.com/nomime.html WKH> WKH> I think the answer may be short and simple. WKH> WKH> Please help me with this trivial problem. Thanks in advance. WKH> WKH> WKH> WKH> W.K. Hong WKH> WKH> -- ======================================================================= 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. -------------- next part -------------- A non-text attachment was scrubbed... Name: gen-si-lda.sh Type: application/x-sh Size: 631 bytes Desc: Url : http://cpmd.org/pipermail/cpmd-list/attachments/20050214/ba3fd120/attachment.sh From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Mon Feb 14 12:48:57 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Mon, 14 Feb 2005 12:48:57 +0100 Subject: [CPMD-list] how to run ld1 programme In-Reply-To: Your message of "Mon, 14 Feb 2005 09:38:18 +0100." Message-ID: <200502141148.j1EBmvj12333@yello.theochem.ruhr-uni-bochum.de> >>> "AK" == Axel Kohlmeyer writes: [...] AK> please also check out http://www.fhi-berlin.mpg.de/th/fhi98md/fhi98PP/ AK> which comes with a very good tutorial on pseudopotential generation. AK> to convert the resulting pseudopotential files to format that is AK> readable by CPMD you can use this program: AK> http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/file/fhi2cpmd.c ooops. there is a typo. this correct URL is http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/files/fhi2cpmd.c sorry, axel. -- ======================================================================= 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 jernej at cmm.ki.si Tue Feb 15 09:57:13 2005 From: jernej at cmm.ki.si (jernej at cmm.ki.si) Date: Tue, 15 Feb 2005 09:57:13 +0100 Subject: [CPMD-list] problems with MD -- conservation of energy and translational drift Message-ID: <1108457833.4211b96985ff1@arg.cmm.ki.si> Dear CPMD community, I wonder if somebody can help me with the following two issues regarding CP molecular dynamics: 1) Conservation of EHAM: when starting CPMD simulation from a previously minimized structure, EHAM is by no means constant and features a significant drift. HOWEVER, if I kill the dynamics after a few (say 10) steps and restart it with the 'RESTART WAVEFUNCTION COORDINATES VELOCITIES NOSE ACCUMULATORS LATEST' and 'QUENCH BO' options, EHAM does not drift anymore. This weird feature does not make any sense to me. It appears with either of the three popular pseudopotentials I've used so far (i.e., Goedecker, Trouiller-Martins, and Vanderbilt). I've been using the BLYP functional throughout. Can somebody tell me if I did something wrong in the simulation setup, or otherwise explain this feature? Is it possible that this problem has anything to do with the way in which the code was compiled (I had quite a few problems with this)? 2) Translational drift: even after "solving" the problem with EHAM, as depicted above, it happens that after running the MD for quite a while, say 20.000 steps of 5 a.u., the DIS value that was reasonably small (up to 1.) and stable within the first 20.000 steps, starts increasing considerably -- it reaches a value of more than 25. after 60.000 steps! Then the whole MD chrases -- it produces meaningless characters in the output: RESTART INFORMATION WRITTEN ON FILE ./RESTART.1 60601 0.00699 335.1 -221.25864 -221.23337 -221.22638 26.429 1.91 60602 0.00697 332.8 -221.25838 -221.23335 -221.22638 26.429 2.05 60603 0.00696 330.9 -221.25816 -221.23334 -221.22638 26.429 1.95 60604 0.00694 329.3 -221.25800 -221.23332 -221.22638 26.430 1.95 60605 0.00692 328.1 -221.25786 -221.23330 -221.22638 26.430 1.99 60606 0.00688 327.1 -221.25774 -221.23326 -221.22638 26.431 1.98 60607 0.00684 326.3 -221.25764 -221.23322 -221.22638 26.432 1.95 60608 0.00679 325.7 -221.25757 -221.23317 -221.22638 26.433 1.94 60609 0.00676 325.2 -221.25752 -221.23314 -221.22638 26.434 1.91 60610 0.00675 324.7 -221.25750 -221.23313 -221.22638 26.436 1.97 60611+++++++++ 324.2 -221.25749 -221.23314++++++++++++++ 26.438 1.96 60612??????????????????????????????????????????????????????????? 26.440 1.86 60613??????????????????????????????????????????????????????????????????? 1.83 60614??????????????????????????????????????????????????????????????????? 1.98 60615??????????????????????????????????????????????????????????????????? 1.88 I've checked the center of mass of the system and indeed it features a drift -- therefore there is considerable translation in the simulation, especially after the DIS value had increased. I'm a bit puzzled about the origin of this drift, for I've made sure that I started my MD from a previously minimized structure! Can somebody tell me what is the source of this problem and how to avoid it? Could the drift have anything to do with the weird conservation of EHAM depicted in 1) ? Thanks in advance for your help, Dr. Jernej Stare Los Alamos National Laboratory, Los Alamos, NM, USA **** Below is the input file that I've been using for my simulations. &CPMD MOLECULAR DYNAMICS CP RESTART WAVEFUNCTION COORDINATES VELOCITIES NOSE ACCUMULATORS LATEST TIMESTEP 5. EMASS 600. NOSE IONS 300 3200 QUENCH BO MAXSTEP 500000 STORE 50 &END &DFT FUNCTIONAL BLYP &END &SYSTEM ANGSTROM SYMMETRY TRICLINIC CELL ABSOLUTE DEGREE 6.1026 3.4867 11.954 90.0 105.79 90.0 CUTOFF 70.0 POINT GROUP AUTO &END &ATOMS *O_MT_BLYP KLEINMAN-BYLANDER LMAX=D 12 0.031558 -0.193163 1.727506 -1.470649 0.849709 0.417095 -1.338586 0.463731 3.698642 -0.031558 0.193163 -1.727506 1.470649 -0.849709 -0.417095 1.338586 -0.463731 -3.698642 1.456458 1.936513 7.478956 -0.045749 0.893641 6.168545 0.086314 1.279619 9.450092 1.393342 1.550187 4.023944 2.895549 2.593059 5.334355 2.763486 2.207081 2.052808 *C_MT_BLYP KLEINMAN-BYLANDER LMAX=D 4 -0.442416 0.205018 0.597921 0.442416 -0.205018 -0.597921 0.982484 1.538332 6.349371 1.867316 1.948368 5.153529 *H_MT_BLYP KLEINMAN-BYLANDER LMAX=S 12 -0.581475 0.077056 2.563421 -0.770136 0.692807 4.458524 -2.003665 -0.159691 4.034067 0.581475 -0.077056 -2.563421 0.770136 -0.692807 -4.458524 2.003665 0.159691 -4.034067 0.843425 1.666294 8.314871 0.654764 1.050543 10.209974 -0.578765 1.903041 9.785517 2.006375 1.820406 3.188029 2.195036 2.436157 1.292926 3.428565 1.583659 1.717383 &END From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Tue Feb 15 11:49:06 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Tue, 15 Feb 2005 11:49:06 +0100 (CET) Subject: [CPMD-list] problems with MD -- conservation of energy and translational drift In-Reply-To: <1108457833.4211b96985ff1@arg.cmm.ki.si> Message-ID: On Tue, 15 Feb 2005 jernej at cmm.ki.si wrote: JS> Dear CPMD community, dear jernej, JS> JS> I wonder if somebody can help me with the following two issues regarding CP JS> molecular dynamics: yes, i may have one or two suggestions. ;-) JS> 1) Conservation of EHAM: when starting CPMD simulation from a previously JS> minimized structure, EHAM is by no means constant and features a significant JS> drift. HOWEVER, if I kill the dynamics after a few (say 10) steps and restart JS> it with the 'RESTART WAVEFUNCTION COORDINATES VELOCITIES NOSE ACCUMULATORS JS> LATEST' and 'QUENCH BO' options, EHAM does not drift anymore. when you start a CP-MD without velocities, then it'll will take a little time until the electronic degrees of freedom have picked up some kinetic energy and are in equilibrium (20-100 steps). if you have a very large timestep (like your 5 a.u.) chances are, that the ions move to fast to stay close enough to the BO-surface for a 'good' CP-dynamics. with the QUENCH BO you re-optimize the wavefunction. please note, that to restart the nose-hoover chains for the ions, you have to use RESTART NOSEP (RESTART NOSE will have no effect, see the code in control.F). JS> This weird feature does not make any sense to me. It appears with either of the JS> three popular pseudopotentials I've used so far (i.e., Goedecker, JS> Trouiller-Martins, and Vanderbilt). I've been using the BLYP functional JS> throughout. Can somebody tell me if I did something wrong in the simulation JS> setup, or otherwise explain this feature? Is it possible that this problem has JS> anything to do with the way in which the code was compiled (I had quite a few JS> problems with this)? no, this should not be a compilation problem. what platform/compiler/libraries do you use and what were your problems? JS> 2) Translational drift: even after "solving" the problem with EHAM, as depicted JS> above, it happens that after running the MD for quite a while, say 20.000 steps JS> of 5 a.u., the DIS value that was reasonably small (up to 1.) and stable within JS> the first 20.000 steps, starts increasing considerably -- it reaches a value of JS> more than 25. after 60.000 steps! Then the whole MD chrases -- it produces JS> meaningless characters in the output: JS> [...] JS> 1.91 JS> 60610 0.00675 324.7 -221.25750 -221.23313 -221.22638 26.436 JS> 1.97 JS> 60611+++++++++ 324.2 -221.25749 -221.23314++++++++++++++ 26.438 JS> 1.96 JS> 60612??????????????????????????????????????????????????????????? 26.440 this happens, when your wavefunction 'diverges', e.g. when an ion moves too fast for the electronic degrees of freedom to be able to follow. JS> I've checked the center of mass of the system and indeed it features a drift -- JS> therefore there is considerable translation in the simulation, especially after JS> the DIS value had increased. I'm a bit puzzled about the origin of this drift, the drift is a consequence of discretization and rounding errors (you are solving the equations of motions numerically for finite time steps and you keep your electron structure on a grid), like it happens for all (long-running) MD simulations. to counter this effect, you can add to your &CPMD section: SUBTRACT COMVEL 1000 which will cancel the center of mass velocity every 1000 steps (the energy is conserved by distributing the removed COM kinetic energy to the other DOFs). please note, that this code does not (yet) take nose-hoover thermostats into account, so you'll see a (small) discontinuity in the total energy, if you use it in your example. JS> for I've made sure that I started my MD from a previously minimized structure! JS> Can somebody tell me what is the source of this problem and how to avoid it? JS> Could the drift have anything to do with the weird conservation of EHAM JS> depicted in 1) ? i have a few more comments regarding your input file. JS> **** Below is the input file that I've been using for my simulations. JS> JS> &CPMD JS> MOLECULAR DYNAMICS CP JS> RESTART WAVEFUNCTION COORDINATES VELOCITIES NOSE ACCUMULATORS LATEST as stated above, you need NOSEP instead of NOSE here. JS> TIMESTEP JS> 5. a timestep of 5.0 a.u. is pretty large for a system containing (light and thus fast moving) hydrogen atoms JS> EMASS JS> 600. my guess is you need to increase EMASS to 600 to get the CP-MD working at all. right? JS> NOSE IONS JS> 300 3200 JS> QUENCH BO you should not do a QUENCH BO regularly, when restarting a trajectory. this will disrupt the dynamics. unless you also do a QUENCH ELECTRONS, your electronic DOFs will retain their old velocities, so you won't gain a lot from it here. if you are concerned about 'loosing' your electron structure you should better use a thermostat for the electrons as well. ideally, you'll thermostat to the average value of EKINC from a non-thermostatted run, e.g. something like 0.0065 in your example. of course you run into the risk of 'thermostatting away' the real problem, which to me seems mainly a too large timestep. JS> MAXSTEP JS> 500000 JS> STORE JS> 50 JS> &END JS> &DFT JS> FUNCTIONAL BLYP with gradient corrected functionals, especially when using a small cutoff (and even more if you run an MD of an isolated molecule) you should use GC-CUTOFF here (e.g. with a value of 1.0e-6 in your case) for a much improved energy conservation. the BLYP functional due to its functional form seems very sensitive to this. JS> &END JS> &SYSTEM JS> ANGSTROM JS> SYMMETRY JS> TRICLINIC JS> CELL ABSOLUTE DEGREE JS> 6.1026 3.4867 11.954 90.0 105.79 90.0 hmmm. this is quite a small cell. did you do any checks, whether your k-point sampling is sufficient with a gamma-point calculation? this should not affect your energy conservation issues, though. JS> CUTOFF JS> 70.0 a plane wave cutoff of 70.0 is very much pushing the limit for your system, if you use the standard MT pseudopotentials. a very nice discussion of what are reasonable parameters for CP-MD is in Kuo et al., J. Phys. Chem. B 2004, 108, 12990-12998. JS> POINT GROUP JS> AUTO JS> &END JS> &ATOMS JS> *O_MT_BLYP KLEINMAN-BYLANDER JS> LMAX=D unless you use custom pseudopotentials, you should most likely use LMAX=P here. [...] JS> 2.895549 2.593059 5.334355 JS> 2.763486 2.207081 2.052808 JS> *C_MT_BLYP KLEINMAN-BYLANDER JS> LMAX=D same as above. [...] JS> *H_MT_BLYP KLEINMAN-BYLANDER JS> LMAX=S here KLEINMANN-BYLANDER is not really needed, but should do nor harm either. best regards, axel. -- ======================================================================= 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 s101bayu at mail.chem.itb.ac.id Wed Feb 16 05:38:45 2005 From: s101bayu at mail.chem.itb.ac.id (s101bayu at mail.chem.itb.ac.id) Date: Wed, 16 Feb 2005 11:38:45 +0700 (WIT) Subject: [CPMD-list] "MEMORY| ALLOCATION FAILED" error message Message-ID: <45712.167.205.72.13.1108528725.squirrel@webmail.chem.itb.ac.id> Dear All, I am having some issues in running the test programs that is provided at the cpmd.org site. When I ran the code using the test inputs I get, "MEMORY| ALLOCATION FAILED" error message. My machine is an Intel Pentium 4 based running Fedora Core 3, Clustering with OSCAR and I compiled the code using intel fortran compiler. Here is my make file: #QMMM_FLAGS = -D__QMECHCOUPL #QMMM_LIBS = -L. -lmm FFLAGS = -c -r8 -w95 -O3 -pc64 -axiKMWN -tpp7 -unroll -cm -tune pn4 -arch pn4 LFLAGS = -L. -latlas -lsvml -Vaxlib -static $(QMMM_LIBS) CFLAGS = -c -O2 -Wall CPP = /lib/cpp -P -C -traditional CPPFLAGS = -DADD_ONE_UNDERSCORE -D__Linux -D__PGI -DLAPACK -DFFT_DEFAULT -DPARALLEL -DMP_LIBRARY=__MPI -DMYRINET -DLINUX_IFC NOOPT_FLAG = CC = cc FC = env LAMHF77=ifort mpif77 LD = env LAMHF77=ifort mpif77 Here is the error message: **************************************************************** PROCESSOR 0 ALLOCATION OF 64672 WORDS OF MEMORY FAILED **************************************************************** *** MEMORY| THE NEW SIZE OF THE PROGRAM IS 2212/ 21808 kBYTES *** ================================================================ BIG MEMORY ALLOCATIONS TAU0 402 VELP 402 NGHCOM 200 WSG 200 NGHTOL 200 RGH 200 WGH 200 LVELINI 136 BL 24 RCL 24 ---------------------------------------------------------------- [PEAK NUMBER 11] PEAK MEMORY 2012 = 0.0 MBytes ================================================================ PROGRAM STOPS IN SUBROUTINE MEMORY| ALLOCATION FAILED (GNL) [PROC= 0] Any suggestions to help me resolve this issue will be appreciated. Thanks Bayu From s101bayu at mail.chem.itb.ac.id Wed Feb 16 05:41:46 2005 From: s101bayu at mail.chem.itb.ac.id (s101bayu at mail.chem.itb.ac.id) Date: Wed, 16 Feb 2005 11:41:46 +0700 (WIT) Subject: [CPMD-list] "MEMORY| ALLOCATION FAILED" error message Message-ID: <45727.167.205.72.13.1108528906.squirrel@webmail.chem.itb.ac.id> Dear All, I am having some issues in running the test programs that is provided at the cpmd.org site. When I ran the code using the test inputs I get, "MEMORY| ALLOCATION FAILED" error message. My machine is an Intel Pentium 4 based running Fedora Core 3, Clustering with OSCAR and I compiled the code using intel fortran compiler. Here is my make file: #QMMM_FLAGS = -D__QMECHCOUPL #QMMM_LIBS = -L. -lmm FFLAGS = -c -r8 -w95 -O3 -pc64 -axiKMWN -tpp7 -unroll -cm -tune pn4 -arch pn4 LFLAGS = -L. -latlas -lsvml -Vaxlib -static $(QMMM_LIBS) CFLAGS = -c -O2 -Wall CPP = /lib/cpp -P -C -traditional CPPFLAGS = -DADD_ONE_UNDERSCORE -D__Linux -D__PGI -DLAPACK -DFFT_DEFAULT -DPARALLEL -DMP_LIBRARY=__MPI -DMYRINET -DLINUX_IFC NOOPT_FLAG = CC = cc FC = env LAMHF77=ifort mpif77 LD = env LAMHF77=ifort mpif77 Here is the error message: **************************************************************** PROCESSOR 0 ALLOCATION OF 64672 WORDS OF MEMORY FAILED **************************************************************** *** MEMORY| THE NEW SIZE OF THE PROGRAM IS 2212/ 21808 kBYTES *** ================================================================ BIG MEMORY ALLOCATIONS TAU0 402 VELP 402 NGHCOM 200 WSG 200 NGHTOL 200 RGH 200 WGH 200 LVELINI 136 BL 24 RCL 24 ---------------------------------------------------------------- [PEAK NUMBER 11] PEAK MEMORY 2012 = 0.0 MBytes ================================================================ PROGRAM STOPS IN SUBROUTINE MEMORY| ALLOCATION FAILED (GNL) [PROC= 0] Any suggestions to help me resolve this issue will be appreciated. Thanks Bayu From jernej at cmm.ki.si Wed Feb 16 09:29:04 2005 From: jernej at cmm.ki.si (jernej at cmm.ki.si) Date: Wed, 16 Feb 2005 09:29:04 +0100 Subject: [CPMD-list] problems with MD -- conservation of energy and translational drift Message-ID: <1108542544.4213045049588@arg.cmm.ki.si> Dear Axel, thanks a lot for your help! As soon as our computer cluster is put back to life -- our system admin is currently rebuilding it -- I'll implement your suggestions in my MD run and let you know about the results. Regarding compilation -- we've been using the Intel Fortran compiler package and it took quite a long time before successfully compiling the parallel code for mpich. Here are the main flags with which we finally made it: ************************************* FFLAGS = -c -r8 -w90 -w95 -O3 -pc64 -axM -ip -tpp7 LFLAGS = -llapack-ifc -lf77blas-ifc -latlas-ifc -Vaxlib $(QMMM_LIBS) LFLAGS = -L/net/pro/opt/intel/mkl/lib/32 -L. -lmkl_lapack -lmkl_def -Vaxlib -lpthread $(QMMM_LIBS) -static CFLAGS = -c -O2 -Wall -I/net/pro/opt/intel/mkl/include CPP = /net/pro/lib/cpp -P -C -traditional CPPFLAGS = -D__Linux -D__PGI -DLAPACK -DFFT_DEFAULT -DLINUX_IFC \ -DPARALLEL -DMP_LIBRARY=__MPI NOOPT_FLAG = CC = gcc-3.2 FC = $(HOME)/mpi/bin/mpif77 LD = $(HOME)/mpi/bin/mpif77 ************************************* Thanks again, and best regards, Jernej From chokosabe at yahoo.co.uk Wed Feb 16 09:39:07 2005 From: chokosabe at yahoo.co.uk (Abraham Alaka) Date: Wed, 16 Feb 2005 08:39:07 +0000 (GMT) Subject: [CPMD-list] "MEMORY| ALLOCATION FAILED" error message In-Reply-To: <45727.167.205.72.13.1108528906.squirrel@webmail.chem.itb.ac.id> Message-ID: <20050216083907.49027.qmail@web25702.mail.ukl.yahoo.com> Check the intel fortran compiler notes that you get with the download. try (in terminal) "ulimit -s unlimited" to set stacksize space to unlimited. Hope you find this helpful. --- s101bayu at mail.chem.itb.ac.id wrote: > Dear All, > > I am having some issues in running the test programs > that is provided at the > cpmd.org site. When I ran the code using the test > inputs I get, "MEMORY| > ALLOCATION FAILED" error message. > > My machine is an Intel Pentium 4 based running > Fedora Core 3, Clustering > with OSCAR and I > compiled the code using intel fortran compiler. Here > is my make file: > > #QMMM_FLAGS = -D__QMECHCOUPL > #QMMM_LIBS = -L. -lmm > FFLAGS = -c -r8 -w95 -O3 -pc64 -axiKMWN -tpp7 > -unroll -cm -tune pn4 -arch pn4 > LFLAGS = -L. -latlas -lsvml -Vaxlib -static > $(QMMM_LIBS) > CFLAGS = -c -O2 -Wall > CPP = /lib/cpp -P -C -traditional > CPPFLAGS = -DADD_ONE_UNDERSCORE -D__Linux -D__PGI > -DLAPACK -DFFT_DEFAULT > -DPARALLEL -DMP_LIBRARY=__MPI -DMYRINET -DLINUX_IFC > NOOPT_FLAG = > CC = cc > FC = env LAMHF77=ifort mpif77 > LD = env LAMHF77=ifort mpif77 > > > Here is the error message: > > **************************************************************** > PROCESSOR 0 ALLOCATION OF 64672 WORDS OF > MEMORY FAILED > > **************************************************************** > > *** MEMORY| THE NEW SIZE OF THE PROGRAM IS > 2212/ 21808 kBYTES *** > > > ================================================================ > BIG MEMORY ALLOCATIONS > TAU0 402 VELP > 402 > NGHCOM 200 WSG > 200 > NGHTOL 200 RGH > 200 > WGH 200 LVELINI > 136 > BL 24 RCL > 24 > > ---------------------------------------------------------------- > [PEAK NUMBER 11] PEAK MEMORY 2012 = > 0.0 MBytes > > ================================================================ > > > PROGRAM STOPS IN SUBROUTINE MEMORY| ALLOCATION > FAILED (GNL) [PROC= 0] > > > Any suggestions to help me resolve this issue will > be appreciated. > > Thanks > Bayu > _______________________________________________ > CPMD-list mailing list > CPMD-list at cpmd.org > http://cpmd.org/mailman/listinfo/cpmd-list > ___________________________________________________________ ALL-NEW Yahoo! Messenger - all new features - even more fun! http://uk.messenger.yahoo.com From fritsch at hmi.de Wed Feb 16 10:47:53 2005 From: fritsch at hmi.de (Wolfgang Fritsch) Date: Wed, 16 Feb 2005 10:47:53 +0100 Subject: [CPMD-list] density calculation: memory errors Message-ID: <421316C9.5090502@hmi.de> Dear in the knowing, for a band structure calculation I have used an input file for the density, that I have modified from file found somewhere in this list here and which seems to have run successfully. Now this does not run at all here. I am getting memory errors. No matter what I can think of in terms of cutting back on the size of the calculation, the memory errors stay on though the details of the complaints look different. Surely something I am doing is odd? This is on a cluster of parallel alpha workstation, with some 4 GB memory, 500 MB is available at each node. I am appending first the input file and then the output from the job, calculated with the serial version of the code. Thank you very much for any suggestions, Wolfgang &CPMD OPTIMIZE WAVEFUNCTION LANCZOS PARAMETER N=4 10 8 0 1.e-10 0.1 1.e-12 0.01 1.e-14 0.001 1.e-16 CONVERGENCE 1.e-4 1.e-3 TEMPERATURE ELECTRON 300.0 STORE WAVEFUNCTION OFF RHOOUT ENERGYBANDS &END &DFT FUNCTIONAL LDA &END &SYSTEM SYMMETRY 8 CUTOFF 120.0 CELL 10.6 1.0 1.961 0.0 0.0 0.0 KPOINTS MONKHORST-PACK BLOCK=32 ALL 4 4 4 &END &ATOMS *Cu-q11 LMAX=D 4 0.0 0.0 0.0 5.3101 0.0 5.1967 5.3101 5.3101 10.3935 0.0 5.3101 15.5902 *Ga-q13 LMAX=D 4 5.3101 5.3101 0.0 0.0 5.9101 5.1967 0.0 0.0 10.3935 5.3101 0.0 15.5902 *Se-q6 LMAX=D 8 2.6752 7.6524 2.6161 7.9058 2.9698 2.6293 2.7936 3.0020 7.6496 7.7490 7.5688 7.5935 2.7621 7.6604 13.1190 7.8472 2.9307 13.1248 2.7714 2.9129 18.0796 7.8473 7.7280 17.0830 &END PROGRAM CPMD STARTED AT: Tue Feb 15 16:42:52 2005 ****** ****** **** **** ****** ******* ******* ********** ******* *** ** *** ** **** ** ** *** ** ** *** ** ** ** ** ** ** ******* ** ** ** ** *** ****** ** ** ** *** ******* ** ** ** ******* ****** ** ** ** ****** VERSION 3.9.1 COPYRIGHT IBM RESEARCH DIVISION MPI FESTKOERPERFORSCHUNG STUTTGART The CPMD consortium WWW: http://www.cpmd.org Mailinglist: cpmd-list at cpmd.org E-mail: cpmd at cpmd.org *** Dec 1 2004 -- 10:47:46 *** THE INPUT FILE IS: densitymod THIS JOB RUNS ON: darling7 THE CURRENT DIRECTORY IS: /hmi/sf5/pfr/CPMD-3.9.1/cases/cugase/MD THE TEMPORARY DIRECTORY IS: /hmi/sf5/pfr/CPMD-3.9.1/cases/cugase/MD THE PROCESS ID IS: 4283681 SINGLE POINT DENSITY OPTIMIZATION PATH TO THE RESTART FILES: ./ GRAM-SCHMIDT ORTHOGONALIZATION MAXIMUM NUMBER OF STEPS: 10000 STEPS PRINT INTERMEDIATE RESULTS EVERY 10001 STEPS STORE INTERMEDIATE RESULTS EVERY 10001 STEPS DO NOT STORE WAVEFUNCTIONS IN RESTART FILE NUMBER OF DISTINCT RESTART FILES: 1 TEMPERATURE IS CALCULATED ASSUMING EXTENDED BULK BEHAVIOR STORE ELECTRON DENSITY AT THE END OF THE RUN FICTITIOUS ELECTRON MASS: 400.0000 TIME STEP FOR ELECTRONS: 5.0000 TIME STEP FOR IONS: 5.0000 CONVERGENCE CRITERIA FOR WAVEFUNCTION OPTIMIZATION: 1.0000E-04 WAVEFUNCTION OPTIMIZATION BY PRECONDITIONED DIIS THRESHOLD FOR THE WF-HESSIAN IS 0.5000 MAXIMUM NUMBER OF VECTORS RETAINED FOR DIIS: 10 STEPS UNTIL DIIS RESET ON POOR PROGRESS: 10 FULL ELECTRONIC GRADIENT IS USED SPLINE INTERPOLATION IN G-SPACE FOR PSEUDOPOTENTIAL FUNCTIONS NUMBER OF SPLINE POINTS: 5000 EXCHANGE CORRELATION FUNCTIONALS LDA EXCHANGE: NONE LDA XC THROUGH PADE APPROXIMATION S.GOEDECKER, J.HUTTER, M.TETER PRB 54 1703 (1996) *** DETSP| THE NEW SIZE OF THE PROGRAM IS 36216 kBYTES *** ***************************** ATOMS **************************** NR TYPE X(bohr) Y(bohr) Z(bohr) MBL 1 Cu 0.000000 0.000000 0.000000 3 2 Cu 5.310100 0.000000 5.196700 3 3 Cu 5.310100 5.310100 10.393500 3 4 Cu 0.000000 5.310100 15.590200 3 5 Ga 5.310100 5.310100 0.000000 3 6 Ga 0.000000 5.910100 5.196700 3 7 Ga 0.000000 0.000000 10.393500 3 8 Ga 5.310100 0.000000 15.590200 3 9 Se 2.675200 7.652400 2.616100 3 10 Se 7.905800 2.969800 2.629300 3 11 Se 2.793600 3.002000 7.649600 3 12 Se 7.749000 7.568800 7.593500 3 13 Se 2.762100 7.660400 13.119000 3 14 Se 7.847200 2.930700 13.124800 3 15 Se 2.771400 2.912900 18.079600 3 16 Se 7.847300 7.728000 17.083000 3 **************************************************************** ******************* SPECIAL K-POINTS GENERATION **************** DIMENSIONS ARE: NUMBER OF ATOMS 16 K POINTS MONKHORST-PACK MESH 4 4 4 MAXIMAL NUMBER OF K POINTS 64 CONSTANT VECTOR SHIFT (MACDONALD) 0.000 0.000 0.000 NON SYMMETRIC SPECIAL K POINTS --> (SEE THE FILE KPTS_GENERATION FOR MORE INFORMATION) NUMBER OF SPECIAL K POINTS (IN CARTESIAN COORDINATES): 32 NKP KX KY KZ WEIGHT 1 0.375000 0.375000 0.191229 0.031250 2 0.375000 0.375000 0.063743 0.031250 3 0.375000 0.375000 -0.063743 0.031250 4 0.375000 0.375000 -0.191229 0.031250 5 0.375000 0.125000 0.191229 0.031250 6 0.375000 0.125000 0.063743 0.031250 7 0.375000 0.125000 -0.063743 0.031250 8 0.375000 0.125000 -0.191229 0.031250 9 0.375000 -0.125000 0.191229 0.031250 10 0.375000 -0.125000 0.063743 0.031250 11 0.375000 -0.125000 -0.063743 0.031250 12 0.375000 -0.125000 -0.191229 0.031250 13 0.375000 -0.375000 0.191229 0.031250 14 0.375000 -0.375000 0.063743 0.031250 15 0.375000 -0.375000 -0.063743 0.031250 16 0.375000 -0.375000 -0.191229 0.031250 17 0.125000 0.375000 0.191229 0.031250 18 0.125000 0.375000 0.063743 0.031250 19 0.125000 0.375000 -0.063743 0.031250 20 0.125000 0.375000 -0.191229 0.031250 21 0.125000 0.125000 0.191229 0.031250 22 0.125000 0.125000 0.063743 0.031250 23 0.125000 0.125000 -0.063743 0.031250 24 0.125000 0.125000 -0.191229 0.031250 25 0.125000 -0.125000 0.191229 0.031250 26 0.125000 -0.125000 0.063743 0.031250 27 0.125000 -0.125000 -0.063743 0.031250 28 0.125000 -0.125000 -0.191229 0.031250 29 0.125000 -0.375000 0.191229 0.031250 30 0.125000 -0.375000 0.063743 0.031250 31 0.125000 -0.375000 -0.063743 0.031250 32 0.125000 -0.375000 -0.191229 0.031250 **************************************************************** NUMBER OF STATES: 72 NUMBER OF ELECTRONS: 144.00000 CHARGE: 0.00000 ELECTRON TEMPERATURE(KELVIN): 0.00000 OCCUPATION 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 2.0 Goedecker/Hartwigsen s ? PP Goedecker/Hartwigsen s ? PP Goedecker/Hartwigsen s ? PP **************************************************************** * ATOM MASS RAGGIO NLCC PSEUDOPOTENTIAL * * Cu 63.5400 1.2000 NO GOEDECKER S NONLOCAL * * P NONLOCAL * * D NONLOCAL * * F LOCAL * * Ga 69.7200 1.2000 NO GOEDECKER S NONLOCAL * * P NONLOCAL * * D NONLOCAL * * F LOCAL * * Se 78.9600 1.2000 NO GOEDECKER S NONLOCAL * * P NONLOCAL * * D NONLOCAL * * F LOCAL * **************************************************************** OPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPEN NUMBER OF CPUS PER TASK 1 OPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPENMPOPEN *** RGGEN| THE NEW SIZE OF THE PROGRAM IS 48062 kBYTES *** *** RKPNT| THE NEW SIZE OF THE PROGRAM IS 166461 kBYTES *** ************************** SUPERCELL *************************** SYMMETRY: ORTHORHOMBIC LATTICE CONSTANT(a.u.): 10.60000 CELL DIMENSION: 10.6000 1.0000 1.9610 0.0000 0.0000 0.0000 VOLUME(OMEGA IN BOHR^3): 2335.58238 LATTICE VECTOR A1(BOHR): 10.6000 0.0000 0.0000 LATTICE VECTOR A2(BOHR): 0.0000 10.6000 0.0000 LATTICE VECTOR A3(BOHR): 0.0000 0.0000 20.7866 RECIP. LAT. VEC. B1(2Pi/BOHR): 0.0943 0.0000 0.0000 RECIP. LAT. VEC. B2(2Pi/BOHR): 0.0000 0.0943 0.0000 RECIP. LAT. VEC. B3(2Pi/BOHR): 0.0000 0.0000 0.0481 REAL SPACE MESH: 80 80 150 WAVEFUNCTION CUTOFF(RYDBERG): 120.00000 DENSITY CUTOFF(RYDBERG): (DUAL= 4.00) 480.00000 NUMBER OF PLANE WAVES PER WAVEFUNCTION 58387 NUMBER OF PLANE WAVES AT GAMMA POINT 51917 NUMBER OF PLANE WAVES FOR DENSITY CUTOFF: 207425 KPOINTS (IN CARTESIAN COORDINATES AS INPUT): 32 NKP KX KY KZ WEIGHT NGW 1 0.37500 0.37500 0.19123 0.03125 51782 2 0.37500 0.37500 0.06374 0.03125 51812 3 0.37500 0.37500 -0.06374 0.03125 51812 4 0.37500 0.37500 -0.19123 0.03125 51782 5 0.37500 0.12500 0.19123 0.03125 51862 6 0.37500 0.12500 0.06374 0.03125 51856 7 0.37500 0.12500 -0.06374 0.03125 51856 8 0.37500 0.12500 -0.19123 0.03125 51862 9 0.37500 -0.12500 0.19123 0.03125 51862 10 0.37500 -0.12500 0.06374 0.03125 51856 11 0.37500 -0.12500 -0.06374 0.03125 51856 12 0.37500 -0.12500 -0.19123 0.03125 51862 13 0.37500 -0.37500 0.19123 0.03125 51782 14 0.37500 -0.37500 0.06374 0.03125 51812 15 0.37500 -0.37500 -0.06374 0.03125 51812 16 0.37500 -0.37500 -0.19123 0.03125 51782 17 0.12500 0.37500 0.19123 0.03125 51862 18 0.12500 0.37500 0.06374 0.03125 51856 19 0.12500 0.37500 -0.06374 0.03125 51856 20 0.12500 0.37500 -0.19123 0.03125 51862 21 0.12500 0.12500 0.19123 0.03125 51863 22 0.12500 0.12500 0.06374 0.03125 51885 23 0.12500 0.12500 -0.06374 0.03125 51885 24 0.12500 0.12500 -0.19123 0.03125 51863 25 0.12500 -0.12500 0.19123 0.03125 51863 26 0.12500 -0.12500 0.06374 0.03125 51885 27 0.12500 -0.12500 -0.06374 0.03125 51885 28 0.12500 -0.12500 -0.19123 0.03125 51863 29 0.12500 -0.37500 0.19123 0.03125 51862 30 0.12500 -0.37500 0.06374 0.03125 51856 31 0.12500 -0.37500 -0.06374 0.03125 51856 32 0.12500 -0.37500 -0.19123 0.03125 51862 **************************************************************** *** RINFORCE| THE NEW SIZE OF THE PROGRAM IS 885178 kBYTES *** *** FFTPRP| THE NEW SIZE OF THE PROGRAM IS 885178 kBYTES *** GENERATE ATOMIC BASIS SET Cu SLATER ORBITALS 3D ALPHA= 5.2833 OCCUPATION=10.00 4S ALPHA= 1.0000 OCCUPATION= 1.00 Ga SLATER ORBITALS 3D ALPHA= 5.9500 OCCUPATION=10.00 4S ALPHA= 1.3514 OCCUPATION= 2.00 4P ALPHA= 1.3514 OCCUPATION= 1.00 Se SLATER ORBITALS 4S ALPHA= 1.8784 OCCUPATION= 2.00 4P ALPHA= 1.8784 OCCUPATION= 4.00 INITIALIZATION TIME: 28.33 SECONDS MEMORY| MEMORY REQUIRED: 269051914 WORDS PROGRAM STOPS IN SUBROUTINE MEMORY| TOO BIG VALUE From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Wed Feb 16 11:31:06 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Wed, 16 Feb 2005 11:31:06 +0100 Subject: [CPMD-list] problems with MD -- conservation of energy and translational drift In-Reply-To: Your message of "Wed, 16 Feb 2005 09:29:04 +0100." <1108542544.4213045049588@arg.cmm.ki.si> Message-ID: <200502161031.j1GAV6K06551@yello.theochem.ruhr-uni-bochum.de> >>> "JS" == jernej writes: JS> Dear Axel, dear jernej, [...] JS> Regarding compilation -- we've been using the Intel Fortran compiler package and JS> it took quite a long time before successfully compiling the parallel code for JS> mpich. Here are the main flags with which we finally made it: JS> ************************************* JS> FFLAGS = -c -r8 -w90 -w95 -O3 -pc64 -axM -ip -tpp7 your main problem is the -ip flag. for a large and in its time critical subroutines already optimized code like CPMD there is no benefit from inter-procedural optimization. i recently made some tests and found a _decrease_ in cpu speed of about 5%. i found similar effects when using the SIMD vectorizer (-axM). finally replacing -O3 with -O2 -unroll also increased the CPMD performance (while at the same time cutting down compilation time). so my 'fastest' FFLAGS setting for intel on P4 are: FFLAGS = -c -r8 -w95 -O2 -pc64 -tpp7 -unroll -cm -tune pn4 -arch pn4 compared to your original flags i found a total increase in execution speed of over 15% in some cases with this 'reduced' optmization. JS> LFLAGS = -llapack-ifc -lf77blas-ifc -latlas-ifc -Vaxlib $(QMMM_LIBS) JS> LFLAGS = -L/net/pro/opt/intel/mkl/lib/32 -L. -lmkl_lapack -lmkl_def -Vaxlib JS> -lpthread $(QMMM_LIBS) -static hmmm, why are you using the generic mkl instead of the P4 version? especially in the BLAS/LAPACK subroutines you benefit the most from the SIMD instructions of the P4 cpu. if you don't have a P4/Xeon cpu, however, you should not use the P4 tuning flags (-tpp7 -tune pn4 -arch pn4) since the architecture of most other x86 cpus (pentium 3, AMD athlon, opteron) works better with -tpp6 -tune pn3 -arch pn3. the p4 architecture is quite different from those and thus requires different optimizations. next, on newer linux machines, you should not use -static, since even static binaries will read some shared objects. in case of static linking, they have to be _binary_ identical. so your binary may fail in strange ways, when you update the system, or want to run it on different installations. if libc, libm, and libpthread are linked dynamically, however, the dynamic linker can utilize the backward compatibility layer of newer library versions and the binary will continue to work correctly. for the very latest version of their fortran compiler, intel introduced the -i-static flag, which will have only the intel provided libraries linked statically. for older versions you can use -static-libcxa instead. you can run 'ifort -help' to see if -i-static is supported. JS> CFLAGS = -c -O2 -Wall -I/net/pro/opt/intel/mkl/include JS> CPP = /net/pro/lib/cpp -P -C -traditional JS> CPPFLAGS = -D__Linux -D__PGI -DLAPACK -DFFT_DEFAULT -DLINUX_IFC \ JS> -DPARALLEL -DMP_LIBRARY=__MPI JS> NOOPT_FLAG = JS> CC = gcc-3.2 JS> FC = $(HOME)/mpi/bin/mpif77 JS> LD = $(HOME)/mpi/bin/mpif77 JS> ************************************* JS> Thanks again, and best regards, your are welcome. best regards, axel. JS> Jernej JS> _______________________________________________ JS> CPMD-list mailing list JS> CPMD-list at cpmd.org JS> http://cpmd.org/mailman/listinfo/cpmd-list -- ======================================================================= 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 ct at chemie.hu-berlin.de Wed Feb 16 12:20:37 2005 From: ct at chemie.hu-berlin.de (Christian Tuma) Date: Wed, 16 Feb 2005 12:20:37 +0100 Subject: [CPMD-list] density calculation: memory errors In-Reply-To: <421316C9.5090502@hmi.de> References: <421316C9.5090502@hmi.de> Message-ID: <20050216112037.GA158675@uranus.chemie.hu-berlin.de> Hello Wolfgang, it is just this combination of cell size and planewave cutoff that requires some more memory. As you can see from the last line, a little more than 2 GB of memory have been requested for allocation. BTW, you need to compile a 64bit-binary (-DPOINTER8 -DMALLOC8) to get such a large chunk of memory within a single memory request. However, if there are only 500 MB available, you should try to run the job in parallel on at least 5-6 CPUs to reduce the memory needed in the individual MPI threads. Good luck, Christian. -- Christian Tuma Humboldt-Universitaet zu Berlin ct at chemie.hu-berlin.de Arbeitsgruppe Quantenchemie (Prof. Sauer) phone: +49-30-20937140 Brook-Taylor-Str. 2, 12489 Berlin, GERMANY fax: +49-30-20937136 http://www.chemie.hu-berlin.de/ag_sauer From axel.kohlmeyer at theochem.ruhr-uni-bochum.de Wed Feb 16 12:50:11 2005 From: axel.kohlmeyer at theochem.ruhr-uni-bochum.de (Axel Kohlmeyer) Date: Wed, 16 Feb 2005 12:50:11 +0100 Subject: [CPMD-list] density calculation: memory errors In-Reply-To: Your message of "Wed, 16 Feb 2005 10:47:53 +0100." <421316C9.5090502@hmi.de> Message-ID: <200502161150.j1GBoBP09882@yello.theochem.ruhr-uni-bochum.de> >>> "WF" == Wolfgang Fritsch writes: WF> Dear in the knowing, dear wolfgang, WF> for a band structure calculation I have used an input file for the WF> density, that I have modified from file found somewhere in this list WF> here and which seems to have run successfully. Now this does not run at WF> all here. I am getting memory errors. No matter what I can think of in WF> terms of cutting back on the size of the calculation, the memory errors WF> stay on though the details of the complaints look different. WF> Surely something I am doing is odd? you may be simply tripping over some bugs, i found some time ago in the swapfile handling for k-points. please try the following changes (you may have to apply them manually, since they are extracted from the cvs but may overlap with other changes, that don't matter here). -begin-patch--begin-patch--begin-patch--begin-patch--begin-patch- Index: rwswap.F =================================================================== RCS file: /users/cvsroot/cpmd/CPMD/rwswap.F,v retrieving revision 1.12.4.1 retrieving revision 1.12.4.3 diff -u -r1.12.4.1 -r1.12.4.3 --- rwswap.F 17 Aug 2004 09:19:29 -0000 1.12.4.1 +++ rwswap.F 19 Sep 2004 16:37:56 -0000 1.12.4.3 @@ -408,12 +408,14 @@ ENDDO TSWCALC=TKBCALC TSWAPC0=.NOT.TKNOSWAP + CALL TIHALT(' INI_SWAP',ISUB) RETURN ENDIF C ==--------------------------------------------------------------== C Special case if we want to calculate k point per k point(TKNOSWAP) IF(IFILE.EQ.1.AND.(.NOT.TSWAPC0)) THEN C Do nothing (no swap file for C0). + CALL TIHALT(' INI_SWAP',ISUB) RETURN ENDIF C ==--------------------------------------------------------------== @@ -427,11 +429,18 @@ IF(ISIDISW(IFILE).LE.0) & CALL STOPGM('INI_SWAP','SIZE OF RECORD NOT SPECIFIED') IUNITSW(IFILE)=50+IFILE -C WARNING: THE RECORDSIZE IS PROCESSOR-DEPENDENT!! -#if defined(__SUN) -C You must specify the length in number of characters. +C WARNING: THE RECORDSIZE IS MACHINE/COMPILER-DEPENDENT!! +#if defined(__SUN) || defined(__PGI) || defined(LINUX_IFC) || defined(__IBM) || defined(__HP) || defined(__OSX) +C these machines specify the length in number of characters. IRECORDSIZE=8*ISIDISW(IFILE) +#elif defined(__SGI) || defined(__alpha) +C these machines specify the length in number of real words. + IRECORDSIZE=2*ISIDISW(IFILE) #else + WRITE(*,'(/,2A,/,2A)') ' WARNING| RECORDSIZE MULTIPLIER FOR', + & ' DIRECT ACCESS SWAP FILES NOT KNOWN.', + & ' WARNING| USING DEFAULT OF 2.', + & ' CHANGE IN FILE rwswap.F IF NEEDED.' IRECORDSIZE=2*ISIDISW(IFILE) #endif OPEN(UNIT=IUNITSW(IFILE),FILE=SWFILE(IFILE), @@ -635,6 +644,7 @@ ENDIF 30 CONTINUE IF(ISWMEM(INDI,IFILE).EQ.IKPT) THEN + CALL TIHALT(' REA_SWAP',ISUB) RETURN ENDIF IPOSITION=ISWPOS(IKPT,INDI,IFILE) Index: sysin.F =================================================================== RCS file: /users/cvsroot/cpmd/CPMD/sysin.F,v retrieving revision 1.42.2.3 retrieving revision 1.42.2.5 diff -u -r1.42.2.3 -r1.42.2.5 --- sysin.F 18 May 2004 13:40:39 -0000 1.42.2.3 +++ sysin.F 19 Sep 2004 16:37:58 -0000 1.42.2.5 @@ -574,6 +574,7 @@ I1=INDEX(LINE,'BLOCK') IF(I1.NE.0) THEN TKBLOCK=.TRUE. + TKBCALC=.TRUE. I2=I1+INDEX(LINE(I1:80),'=') CALL READSI(LINE(I2:80),1,IOUT,NKPNT,ERREAD) IF(ERREAD) THEN @@ -585,7 +586,9 @@ & CALL STOPGM(' SYSIN','WRONG NUMBER OF K POINTS BLOCK') IF(INDEX(LINE,'ALL').NE.0) THEN TKALL=.TRUE. + TKBCALC=.FALSE. ELSEIF(INDEX(LINE,'CALC').NE.0) THEN + TKALL=.FALSE. TKBCALC=.TRUE. ENDIF IF(INDEX(LINE,'NOSWAP').NE.0) THEN Index: sysin.F =================================================================== RCS file: /users/cvsroot/cpmd/CPMD/sysin.F,v retrieving revision 1.42.2.6 retrieving revision 1.42.2.7 diff -u -r1.42.2.6 -r1.42.2.7 --- sysin.F 22 Sep 2004 09:18:27 -0000 1.42.2.6 +++ sysin.F 18 Dec 2004 12:40:31 -0000 1.