From eric... at gmail.com Sat Jul 1 03:51:10 2017 From: eric... at gmail.com (Chanwoo Noh) Date: Fri, 30 Jun 2017 20:51:10 -0700 (PDT) Subject: Mulliken charge of atoms in mos2 monolayer is wrong In-Reply-To: <31626c4f-2dfb-4e84-a6dd-f75cbacbd0ea@googlegroups.com> References: <33125662-f42f-4113-910a-4da7f5405af7@googlegroups.com> <31626c4f-2dfb-4e84-a6dd-f75cbacbd0ea@googlegroups.com> Message-ID: Dear Matt W, I really appreciate your suggestion and comment. As you suggest, I calculated again with TZV basis sets. The BASIS_MOLOPT_UCL file only contains the TZVP basis for Mo atoms, so I used the Mo TZVP-SR(short range) basis in BASIS_MOLOPT_UCL file, and S TZVP basis in BASIS_MOLOPT file. Is it ok to use data in both files at the same time? If it is right, when using TZVP basis, the absolute value of Mulliken charges are decreased, but their signs are still opposite. As you said, I should check other properties such as band gap and geometry. I attach the input and output data for calculation using TZVP basis. I thank for your comment again. Chanwoo -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: MoS2.inp Type: chemical/x-gamess-input Size: 1469 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: MoS2.out Type: application/octet-stream Size: 414573 bytes Desc: not available URL: From mattwa... at gmail.com Sat Jul 1 07:59:41 2017 From: mattwa... at gmail.com (Matt W) Date: Sat, 1 Jul 2017 00:59:41 -0700 (PDT) Subject: Mulliken charge of atoms in mos2 monolayer is wrong In-Reply-To: References: <33125662-f42f-4113-910a-4da7f5405af7@googlegroups.com> <31626c4f-2dfb-4e84-a6dd-f75cbacbd0ea@googlegroups.com> Message-ID: It is fine to mix the basis sets from those two files (they are compatible) Matt On Saturday, July 1, 2017 at 4:51:11 AM UTC+1, Chanwoo Noh wrote: > > Dear Matt W, > > I really appreciate your suggestion and comment. > As you suggest, I calculated again with TZV basis sets. > The BASIS_MOLOPT_UCL file only contains the TZVP basis for Mo atoms, so I > used the Mo TZVP-SR(short range) basis in BASIS_MOLOPT_UCL file, and S TZVP > basis in BASIS_MOLOPT file. > Is it ok to use data in both files at the same time? > If it is right, when using TZVP basis, the absolute value of Mulliken > charges are decreased, but their signs are still opposite. > As you said, I should check other properties such as band gap and geometry. > I attach the input and output data for calculation using TZVP basis. > I thank for your comment again. > > Chanwoo > -------------- next part -------------- An HTML attachment was scrubbed... URL: From eric... at gmail.com Sat Jul 1 09:56:39 2017 From: eric... at gmail.com (Chanwoo Noh) Date: Sat, 1 Jul 2017 02:56:39 -0700 (PDT) Subject: I have a question for calculating band structure. Message-ID: <4872e5c4-0115-4fc3-8835-fdb52bd00e39@googlegroups.com> Dear all, I want to calculate the band structure of the MoS2 layer. I made the input file by referring to the following exercise: https://www.cp2k.org/exercises:2016_uzh_cmest:band_structure_calculation In this exercise, BAND_STRUCTURE option is in the FORCE_EVAL -- DFT-- KPOINTS section, but I found that BAND_STRUCTURE option is in the FORCE_EVAL-- DFT--PRINT section in the manual homepage https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/PRINT/BAND_STRUCTURE.html . Which one is right?? Where can I get any example input file for band structure calculation? Chanwoo Noh, -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Sun Jul 2 20:57:30 2017 From: mattwa... at gmail.com (Matt W) Date: Sun, 2 Jul 2017 13:57:30 -0700 (PDT) Subject: I have a question for calculating band structure. In-Reply-To: <4872e5c4-0115-4fc3-8835-fdb52bd00e39@googlegroups.com> References: <4872e5c4-0115-4fc3-8835-fdb52bd00e39@googlegroups.com> Message-ID: <19c5ce69-48bf-4951-a8b5-4580fc7d4622@googlegroups.com> I think the exercise is probably based on an early version of the code and some things have been rearranged. The manual for the version of the code you are using is correct for you, best to update if you can. Should be some examples in the $CP2K/tests/QS/regtest-kp directories, I'd expect. Matt On Saturday, July 1, 2017 at 10:56:39 AM UTC+1, Chanwoo Noh wrote: > > Dear all, > > I want to calculate the band structure of the MoS2 layer. > I made the input file by referring to the following exercise: > https://www.cp2k.org/exercises:2016_uzh_cmest:band_structure_calculation > In this exercise, BAND_STRUCTURE option is in the FORCE_EVAL -- DFT-- > KPOINTS section, > but I found that BAND_STRUCTURE option is in the FORCE_EVAL-- DFT--PRINT > section in the manual homepage > https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/PRINT/BAND_STRUCTURE.html > . > Which one is right?? > Where can I get any example input file for band structure calculation? > > > Chanwoo Noh, > -------------- next part -------------- An HTML attachment was scrubbed... URL: From anirudd... at gmail.com Mon Jul 3 16:24:36 2017 From: anirudd... at gmail.com (Aniruddha Dive) Date: Mon, 3 Jul 2017 09:24:36 -0700 (PDT) Subject: [CP2K:9096] Simulating of pure sodium metal using CP2K In-Reply-To: References: Message-ID: <553c37c8-5a46-4d84-848e-dbafa30f2b9c@googlegroups.com> Thanks Juerg, I have used EPS_DEFAULT as suggested. Attached is input file for cell optimization of single unit cell of sodium BCC with a lattice constant of 4.29. After performing cell optimization I obtained a lattice parameter of 3.599 which is significantly lower than 4.29. Also the fermi energy value is three times higher than the reported value for pure sodium. Kindly let me know why I am getting such varied results. I am using fermi dirac smearing at 300K and also using diagonalization methods as suggested earlier. Regards, Aniruddha M Dive PhD Candidate, Washington State University On Wednesday, June 14, 2017 at 12:45:21 AM UTC-7, jgh wrote: > > Hi > > your EPS_DEFAULT value is too small (EPS_DEFAULT 5.0E-5). > You should use at least the default value (1.E-10). > > regards > > Juerg > -------------------------------------------------------------- > Juerg Hutter Phone : ++41 44 635 4491 > Institut f?r Chemie C FAX : ++41 44 635 6838 > Universit?t Z?rich E-mail: hut... at chem.uzh.ch > > Winterthurerstrasse 190 > CH-8057 Z?rich, Switzerland > --------------------------------------------------------------- > > -----cp... at googlegroups.com wrote: -----To: cp2k < > cp... at googlegroups.com > > From: Aniruddha Dive > Sent by: cp... at googlegroups.com > Date: 06/13/2017 07:55PM > Subject: [CP2K:9096] Simulating of pure sodium metal using CP2K > > Hi All, > > I am trying to perform ab initio MD of pure sodium metal using CP2K. > Attached is my input and output files for the same. Since for pure sodium I > am using a semi core PP with nine valance electrons the required energy cut > off is relatively higher (~ 1200 Ry). I am using diagonalization and fermi > smearing at 300 K which is recommended for metallic systems. But still > after a few fs of the run I find that the sodium atoms begin to lose its > shape within the crystal. Also the fermi energy shown in the output while > performing a geometry optimization is only 1/3 of that reported for pure > sodium. > > Kindly let me know what parameters I am setting wrong in doing these > calculations? > > Regards, > Aniruddha M Dive > PhD Candidate > Washington State University > > > > > -- > > You received this message because you are subscribed to the Google Groups > "cp2k" group. > > To unsubscribe from this group and stop receiving emails from it, send an > email to cp2k+... at googlegroups.com . > > To post to this group, send email to cp... at googlegroups.com . > > > Visit this group at https://groups.google.com/group/cp2k. > > For more options, visit https://groups.google.com/d/optout. > > > > [attachment "md_output.log" removed by J?rg Hutter/at/UZH] > [attachment "md-pos-1.xyz" removed by J?rg Hutter/at/UZH] > [attachment "pure_sodium.inp" removed by J?rg Hutter/at/UZH] > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: pure_sodium.inp Type: chemical/x-gamess-input Size: 4286 bytes Desc: not available URL: From hut... at chem.uzh.ch Tue Jul 4 07:45:48 2017 From: hut... at chem.uzh.ch (hut... at chem.uzh.ch) Date: Tue, 4 Jul 2017 09:45:48 +0200 Subject: [CP2K:9191] Simulating of pure sodium metal using CP2K In-Reply-To: <553c37c8-5a46-4d84-848e-dbafa30f2b9c@googlegroups.com> References: <553c37c8-5a46-4d84-848e-dbafa30f2b9c@googlegroups.com>, Message-ID: Hi for such a small cell you need to use an adequate k-point grid. regards Juerg -------------------------------------------------------------- Juerg Hutter Phone : ++41 44 635 4491 Institut f?r Chemie C FAX : ++41 44 635 6838 Universit?t Z?rich E-mail: hut... at chem.uzh.ch Winterthurerstrasse 190 CH-8057 Z?rich, Switzerland --------------------------------------------------------------- -----cp... at googlegroups.com wrote: -----To: cp2k From: Aniruddha Dive Sent by: cp... at googlegroups.com Date: 07/03/2017 06:24PM Subject: Re: [CP2K:9191] Simulating of pure sodium metal using CP2K Thanks Juerg, I have used EPS_DEFAULT as suggested. Attached is input file for cell optimization of single unit cell of sodium BCC with a lattice constant of 4.29. After performing cell optimization I obtained a lattice parameter of 3.599 which is significantly lower than 4.29. Also the fermi energy value is three times higher than the reported value for pure sodium. Kindly let me know why I am getting such varied results. I am using fermi dirac smearing at 300K and also using diagonalization methods as suggested earlier. Regards, Aniruddha M Dive PhD Candidate, Washington State University On Wednesday, June 14, 2017 at 12:45:21 AM UTC-7, jgh wrote:Hi your EPS_DEFAULT value is too small (EPS_DEFAULT 5.0E-5). You should use at least the default value (1.E-10). regards Juerg -------------------------------------------------------------- Juerg Hutter Phone : ++41 44 635 4491 Institut f?r Chemie C FAX : ++41 44 635 6838 Universit?t Z?rich E-mail: hut... at chem.uzh.ch Winterthurerstrasse 190 CH-8057 Z?rich, Switzerland --------------------------------------------------------------- -----cp... at googlegroups.com wrote: -----To: cp2k From: Aniruddha Dive Sent by: cp... at googlegroups.com Date: 06/13/2017 07:55PM Subject: [CP2K:9096] Simulating of pure sodium metal using CP2K Hi All, I am trying to perform ab initio MD of pure sodium metal using CP2K. Attached is my input and output files for the same. Since for pure sodium I am using a semi core PP with nine valance electrons the required energy cut off is relatively higher (~ 1200 Ry). I am using diagonalization and fermi smearing at 300 K which is recommended for metallic systems. But still after a few fs of the run I find that the sodium atoms begin to lose its shape within the crystal. Also the fermi energy shown in the output while performing a geometry optimization is only 1/3 of that reported for pure sodium. Kindly let me know what parameters I am setting wrong in doing these calculations? Regards, Aniruddha M Dive PhD Candidate Washington State University -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. [attachment "md_output.log" removed by J?rg Hutter/at/UZH] [attachment "md-pos-1.xyz" removed by J?rg Hutter/at/UZH] [attachment "pure_sodium.inp" removed by J?rg Hutter/at/UZH] -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. [attachment "pure_sodium.inp" removed by J?rg Hutter/at/UZH] From taosun... at gmail.com Tue Jul 4 18:43:32 2017 From: taosun... at gmail.com (Sun Tao) Date: Tue, 4 Jul 2017 11:43:32 -0700 (PDT) Subject: cp2k qmmm simulation -- system vibrating Message-ID: Dear experts: I am setting up a qmmm simulation with CP2K. In the setup, I set the QM box to be the same with the whole system box, with periodic boundary condition. The md ensemble is in NVT. The simulations could run, however, the output trajectory was strange. 1. The whole system is vibrating. 2. Some atoms are moving out of the periodic box. Please see the attached input files. The trajectory file and the output file could be find following the google drive link, and you could see the problems I mentioned above. I would be appreciating any suggestions and feedbacks. https://drive.google.com/open?id=0B-YkNOdiWhHReDdHOEdwZEhweE0 Thanks very much for your help. Tao -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: dftd3.dat Type: application/octet-stream Size: 1959318 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: GTH_POTENTIALS Type: application/octet-stream Size: 125422 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: HFX_BASIS Type: application/octet-stream Size: 17275 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: qmmm02_cnt.cp2k.inp Type: chemical/x-gamess-input Size: 6637 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: total_cnt_pr.pdb Type: application/vnd.palm Size: 79761 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: total_cnt_pr.psf Type: application/octet-stream Size: 96348 bytes Desc: not available URL: From 161brun... at gmail.com Thu Jul 6 13:48:08 2017 From: 161brun... at gmail.com (Tobias Kraemer) Date: Thu, 6 Jul 2017 06:48:08 -0700 (PDT) Subject: MD with CP2K combined with PLUMED Message-ID: Dear all, I have a question regarding CP2K MD simulations using PLUMED. This might become lengthy to explain, but I hope that someone can perhaps give some advice. I will make corresponding files available through google drive. The aim is to simulate the rearrangement of a substrate coordinated to a metal centre. Specifically, this involves a "flip" of the coordination site of a double bond. >From previous static calculations I have found out a possible pathway for this, which serves as a guidance for setting up the MD simulations. I have defined 3 structures (start, transitions state, final) in the file *path_flip.pdb**. *This should give you an idea what process I am interested in, essentially the substrate flips around. As a note aside, here I am interested in a molecular system, i.e. these simulations will use a cubic supercell, and perhaps this will introduce some problems (see below). Later on, I wish to return to the actual system, and embed this molecular portion back into it's solid state environment (for various reasons it is good to look at the molecular system first). Now, since this process can be viewed as a conformational change, an approach that appears reasonable here is the sampling method proposed by Parinello (J. Chem. Phys., 2007, 126, 054103), which uses as the collective variable an expression that measures the root-mean-square-deviation (RMSD) of the sampled structures along a defined reaction path (as defined in *path_flip.pdb*). This method appears to work quite nicely with PLUMED. Indeed, a quick MD simulation (500 steps) using the movingconstraints keyword in PLUMED and forcing the trajectory along the defined path shows that the transition is smooth and follows the defined path (see https://drive.google.com/open?id=0B3k7Eu2qqsBBNUl2dFlneWROajA). The reference point is always provided in the *path_flip.pdb* file. Note also that only the atoms of the substrate are taken into account, in order to remove "noise" from the movement of the other atoms (this is specified in the .pdb file). Having established that this is a suitable collective variable describing the rearrangement under study, I then moved on to sample the trajectory using metadynamics (see https://drive.google.com/open?id=0B3k7Eu2qqsBBQ21hczVpbUt1azQ). I have done 5 MD runs so far (4 restarts), and this is where things get confusing. I have uploaded the files corresponding to the last restart. The file *movie.xyz *file contains the complete trajectory, which I have concatenated from the individual simulations. In each case I would follow the following protocol: (i) copy the CP2K .restart file from the previous md run to new folder (ii) copy path_flip.pdb, plumed.dat, HILLS and COLVAR files to this folder as well (iii) resubmit job. *plumed.dat* contains the RESTART keyword, and I hope this is sufficient information to continue with the simulation. The PLUMED.OUT file reports that 192 Gaussians were read, which implies that the restart has worked. Previous runs restarted from fewer numbers of hills, naturally. What are the problems: (i) From the COLVAR file it appears that the collective variable as defined above (=path) is changing and the substrate is undergoing some conformational changes. In the start the value is around 1.0, fluctuates around this value for a while, then changes rapidly at around t=13-14 ps, appears to reach the transition phase at 20 ps and is a t the final structure thereafter (although there is some strong fluctuations going on there). However, none of this is reflected in the actual trajectory, where the substrate moves around but never undergoes the transition. There appears to be a dichotomy between the values of the CV and the actual structure. (ii) It also appears that there are jumps in the sampling, as seen from the COLVAR and HILL file. For example after 8 ps the next hills are only spawned at 9.44 ps. How does this happen? Are my settings not right? (iii) Another point of concern relates to the fact that this simulation is molecular, and uses a cubic box (25 Ang). You can see that the molecule rotates a lot, and I wonder if this is affects the estimation of the RMSD value (i.e. the CV). I believe that this method can deal with translations, but not so sure about rotations. HAs anyone experience with this, when the molecule moves a lot during the simulation. I am also using the center coordinates keyword (WAVELET!), how does this affect the results? Is it a problem that the reference structures are in fact sitting on one corner of the box, while the molecule for simulations is at the center (I think not, since this should be accounted for through transnational symmetry). Anyways, I would appreciate any reassurance here. It appears that this is a neat way of doing these simulations, but it seems to be also quite tricky to get it right. I am fairly new to this kind of simulations, so any help would be very much appreciated. The initial simulations hold promise, now I am trying to get the actual mtd right to get some reliable results. Ultimately I would like to sample the system in the solid state, i.e. using periodicity, in order to model this rearrangement in dependence of the crystal structure. It would be nice to work out the reason for variations in the barriers for this conformational change in different systems, which we also study experimentally. I am also using umbrella sampling as an alternative to the mtd part. I am also aware that other collective variables could be defined to sample this process (even without resorting to PLUMED), but I am simply interested in making this work. Thanks for your help, and apologies for this lengthy post. I hope that there are some experts out there that can help. Tobias -------------- next part -------------- An HTML attachment was scrubbed... URL: From ashishd... at gmail.com Thu Jul 6 16:36:18 2017 From: ashishd... at gmail.com (ashish dabral) Date: Thu, 6 Jul 2017 09:36:18 -0700 (PDT) Subject: Setting correct cutoff/convergence criteria Message-ID: <66c6e786-5aa2-479e-b7f8-529cd5e920b6@googlegroups.com> Hi, I am using cp2k for geometry optimization (GEO_OPT only). I have a system with almost 300 atoms(40x25x11 CELL). Even though cp2k is really fast, it seems like here it is taking too much time as I have to wait almost 2 weeks to get it to converge with 12 cores (it does converge!) and a gamma point calculation. I am wondering if my cut-off/convergence values are ok. I realized that the GEO_OPT parameter section was missing in the MOTION section but I assume then that the default values were used. There is a variety of atoms in the complex (Si, C, Mo, Br etc.). Is it normal then, given the below setup, to consume so much time? A complete input is not possible to attach but I am attaching the file showing different parameters: &FORCE_EVAL METHOD QS &DFT MULTIPLICITY 1 CHARGE 0 BASIS_SET_FILE_NAME /BASIS_MOLOPT POTENTIAL_FILE_NAME /GTH_POTENTIALS * SURFACE_DIPOLE_CORRECTION TRUE* * SURF_DIP_DIR Z* &QS METHOD GPW EXTRAPOLATION USE_GUESS &END QS &MGRID CUTOFF 900 REL_CUTOFF 60 &END MGRID &PRINT &MO OFF &END MO &PDOS DEBUG NLUMO 5 &END PDOS &END PRINT &SCF SCF_GUESS ATOMIC MAX_SCF 400 ADDED_MOS 200 &DIAGONALIZATION T ALGORITHM STANDARD &END DIAGONALIZATION &MIXING T NBUFFER 4 BETA 1.25 ALPHA 0.2 METHOD BROYDEN_MIXING &END MIXING &PRINT &RESTART ON &END RESTART &END PRINT &SMEAR ON METHOD FERMI_DIRAC ELECTRONIC_TEMPERATURE [K] 300 &END SMEAR &END SCF &XC &XC_FUNCTIONAL PBE * &PBE* * PARAMETRIZATION ORIG* * &END PBE* &END XC_FUNCTIONAL *&VDW_POTENTIAL * * POTENTIAL_TYPE PAIR_POTENTIAL * * &PAIR_POTENTIAL * * TYPE DFTD3 * * REFERENCE_FUNCTIONAL PBE * * R_CUTOFF 10.0 * * PARAMETER_FILE_NAME /dftd3.dat * * SCALING 0.0 * * &END PAIR_POTENTIAL * * &END VDW_POTENTIAL * &END XC &END DFT &PRINT &STRESS_TENSOR ON &END STRESS_TENSOR &FORCES ON &END FORCES &END PRINT &END FORCE_EVAL &MOTION &CELL_OPT OPTIMIZER BFGS MAX_DR 0.005 MAX_ITER 100 EXTERNAL_PRESSURE 0 MAX_FORCE 0.0004 RMS_DR 0.002 PRESSURE_TOLERANCE 100 RMS_FORCE 0.001 TYPE DIRECT_CELL_OPT &END CELL_OPT &PRINT &TRAJECTORY FORMAT PDB &END TRAJECTORY &END PRINT &END MOTION Thanks for any inputs. -------------- next part -------------- An HTML attachment was scrubbed... URL: From leili... at gmail.com Sun Jul 9 20:31:43 2017 From: leili... at gmail.com (stclair) Date: Sun, 9 Jul 2017 13:31:43 -0700 (PDT) Subject: Colvar output file in extended Lagrange metadynamics Message-ID: <3b13a26a-5975-4cc9-bc1e-7b34d9cb97e9@googlegroups.com> Dear CP2K Developer, I am trying to understand the Colvar output file in extended Lagrange metadynamics. I quickly went through the source code. It looks that the 2nd and the 3rd column is corresponding to cv%ss0 and cv%ss, respectively. Deos 'ss0' represents auxiliary variable corresponding to friction particle but ss is instant value of CV variable? I will appreciate for your answer. Thanks. -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Mon Jul 10 16:48:55 2017 From: mattwa... at gmail.com (Matt W) Date: Mon, 10 Jul 2017 09:48:55 -0700 (PDT) Subject: Setting correct cutoff/convergence criteria In-Reply-To: <66c6e786-5aa2-479e-b7f8-529cd5e920b6@googlegroups.com> References: <66c6e786-5aa2-479e-b7f8-529cd5e920b6@googlegroups.com> Message-ID: <3acf0ea9-6b29-436e-922d-af950667157b@googlegroups.com> Hi, there are quite a few things that could be improved in your input, but the most significant problem is probably EXTRAPOLATION USE_GUESS This means that after every step it starts again from an atomic guess, which is very inefficient. By default CP2K extrapolates from the density matrix of the previous steps. Also, 900 Ry is a very expensive cutoff (might be needed for some elements with GPW). Change to METHOD GAPW in &QS and reduce the cutoff to ~300 Ry. Mixing and broyden optimizer will be moderately inefficient for insulators, but the above are probably the biggest things. Hope that helps, Matt On Thursday, July 6, 2017 at 6:31:57 PM UTC+2, ashish dabral wrote: > > Hi, > > I am using cp2k for geometry optimization (GEO_OPT only). I have a system > with almost 300 atoms(40x25x11 CELL). Even though cp2k is really fast, it > seems > like here it is taking too much time as I have to wait almost 2 weeks to > get it to converge with 12 cores (it does converge!) and a gamma point > calculation. > I am wondering if my cut-off/convergence values are ok. I realized that > the GEO_OPT parameter section was missing in the MOTION section but I assume > then that the default values were used. There is a variety of atoms in the > complex (Si, C, Mo, Br etc.). Is it normal then, given the below setup, to > consume > so much time? A complete input is not possible to attach but I am > attaching the file showing different parameters: > > &FORCE_EVAL > METHOD QS > &DFT > MULTIPLICITY 1 > CHARGE 0 > BASIS_SET_FILE_NAME /BASIS_MOLOPT > POTENTIAL_FILE_NAME /GTH_POTENTIALS > * SURFACE_DIPOLE_CORRECTION TRUE* > * SURF_DIP_DIR Z* > &QS > METHOD GPW > EXTRAPOLATION USE_GUESS > &END QS > &MGRID > CUTOFF 900 > REL_CUTOFF 60 > &END MGRID > &PRINT > &MO OFF > &END MO > &PDOS DEBUG > NLUMO 5 > &END PDOS > &END PRINT > &SCF > SCF_GUESS ATOMIC > MAX_SCF 400 > ADDED_MOS 200 > &DIAGONALIZATION T > ALGORITHM STANDARD > &END DIAGONALIZATION > &MIXING T > NBUFFER 4 > BETA 1.25 > ALPHA 0.2 > METHOD BROYDEN_MIXING > &END MIXING > &PRINT > &RESTART ON > &END RESTART > &END PRINT > &SMEAR ON > METHOD FERMI_DIRAC > ELECTRONIC_TEMPERATURE [K] 300 > &END SMEAR > &END SCF > &XC > &XC_FUNCTIONAL PBE > * &PBE* > * PARAMETRIZATION ORIG* > * &END PBE* > &END XC_FUNCTIONAL > *&VDW_POTENTIAL > * > * POTENTIAL_TYPE PAIR_POTENTIAL > * > * &PAIR_POTENTIAL > * > * TYPE DFTD3 > * > * REFERENCE_FUNCTIONAL PBE > * > * R_CUTOFF 10.0 > * > * PARAMETER_FILE_NAME /dftd3.dat * > * SCALING 0.0 > * > * &END PAIR_POTENTIAL > * > * &END VDW_POTENTIAL * > &END XC > &END DFT > &PRINT > &STRESS_TENSOR ON > &END STRESS_TENSOR > &FORCES ON > &END FORCES > &END PRINT > &END FORCE_EVAL > &MOTION > &CELL_OPT > OPTIMIZER BFGS > MAX_DR 0.005 > MAX_ITER 100 > EXTERNAL_PRESSURE 0 > MAX_FORCE 0.0004 > RMS_DR 0.002 > PRESSURE_TOLERANCE 100 > RMS_FORCE 0.001 > TYPE DIRECT_CELL_OPT > &END CELL_OPT > &PRINT > &TRAJECTORY > FORMAT PDB > &END TRAJECTORY > &END PRINT > &END MOTION > > Thanks for any inputs. > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ashishd... at gmail.com Mon Jul 10 17:28:41 2017 From: ashishd... at gmail.com (ashish dabral) Date: Mon, 10 Jul 2017 10:28:41 -0700 (PDT) Subject: Setting correct cutoff/convergence criteria In-Reply-To: <3acf0ea9-6b29-436e-922d-af950667157b@googlegroups.com> References: <66c6e786-5aa2-479e-b7f8-529cd5e920b6@googlegroups.com> <3acf0ea9-6b29-436e-922d-af950667157b@googlegroups.com> Message-ID: <076ce3f3-b7d7-403c-96a8-d7b6c3535dc5@googlegroups.com> Hi Matt, Thanks for your suggestions. After playing around with the parameters, I figured the Extrapolation issue and cut-off issue (so fixed that). I already see improvement in time after some trials. I am wondering though which method would be best for extrapolation: ASPC etc. I haven't switched to GAPW though. I will try that. Can you please elaborate on "Mixing and broyden optimizer will be moderately inefficient for insulators" and a possible alternative. Thanks On Monday, 10 July 2017 18:48:55 UTC+2, Matt W wrote: > > Hi, > > there are quite a few things that could be improved in your input, but the > most significant problem is probably > > EXTRAPOLATION USE_GUESS > > This means that after every step it starts again from an atomic guess, > which is very inefficient. By default CP2K extrapolates from the density > matrix of the previous steps. > > Also, 900 Ry is a very expensive cutoff (might be needed for some elements > with GPW). Change to METHOD GAPW in &QS and reduce the cutoff to ~300 Ry. > > Mixing and broyden optimizer will be moderately inefficient for > insulators, but the above are probably the biggest things. > > Hope that helps, > > Matt > > On Thursday, July 6, 2017 at 6:31:57 PM UTC+2, ashish dabral wrote: >> >> Hi, >> >> I am using cp2k for geometry optimization (GEO_OPT only). I have a system >> with almost 300 atoms(40x25x11 CELL). Even though cp2k is really fast, it >> seems >> like here it is taking too much time as I have to wait almost 2 weeks to >> get it to converge with 12 cores (it does converge!) and a gamma point >> calculation. >> I am wondering if my cut-off/convergence values are ok. I realized that >> the GEO_OPT parameter section was missing in the MOTION section but I assume >> then that the default values were used. There is a variety of atoms in >> the complex (Si, C, Mo, Br etc.). Is it normal then, given the below setup, >> to consume >> so much time? A complete input is not possible to attach but I am >> attaching the file showing different parameters: >> >> &FORCE_EVAL >> METHOD QS >> &DFT >> MULTIPLICITY 1 >> CHARGE 0 >> BASIS_SET_FILE_NAME /BASIS_MOLOPT >> POTENTIAL_FILE_NAME /GTH_POTENTIALS >> * SURFACE_DIPOLE_CORRECTION TRUE* >> * SURF_DIP_DIR Z* >> &QS >> METHOD GPW >> EXTRAPOLATION USE_GUESS >> &END QS >> &MGRID >> CUTOFF 900 >> REL_CUTOFF 60 >> &END MGRID >> &PRINT >> &MO OFF >> &END MO >> &PDOS DEBUG >> NLUMO 5 >> &END PDOS >> &END PRINT >> &SCF >> SCF_GUESS ATOMIC >> MAX_SCF 400 >> ADDED_MOS 200 >> &DIAGONALIZATION T >> ALGORITHM STANDARD >> &END DIAGONALIZATION >> &MIXING T >> NBUFFER 4 >> BETA 1.25 >> ALPHA 0.2 >> METHOD BROYDEN_MIXING >> &END MIXING >> &PRINT >> &RESTART ON >> &END RESTART >> &END PRINT >> &SMEAR ON >> METHOD FERMI_DIRAC >> ELECTRONIC_TEMPERATURE [K] 300 >> &END SMEAR >> &END SCF >> &XC >> &XC_FUNCTIONAL PBE >> * &PBE* >> * PARAMETRIZATION ORIG* >> * &END PBE* >> &END XC_FUNCTIONAL >> *&VDW_POTENTIAL >> * >> * POTENTIAL_TYPE PAIR_POTENTIAL >> * >> * &PAIR_POTENTIAL >> * >> * TYPE DFTD3 >> * >> * REFERENCE_FUNCTIONAL PBE >> * >> * R_CUTOFF 10.0 >> * >> * PARAMETER_FILE_NAME /dftd3.dat * >> * SCALING 0.0 >> * >> * &END PAIR_POTENTIAL >> * >> * &END VDW_POTENTIAL * >> &END XC >> &END DFT >> &PRINT >> &STRESS_TENSOR ON >> &END STRESS_TENSOR >> &FORCES ON >> &END FORCES >> &END PRINT >> &END FORCE_EVAL >> &MOTION >> &CELL_OPT >> OPTIMIZER BFGS >> MAX_DR 0.005 >> MAX_ITER 100 >> EXTERNAL_PRESSURE 0 >> MAX_FORCE 0.0004 >> RMS_DR 0.002 >> PRESSURE_TOLERANCE 100 >> RMS_FORCE 0.001 >> TYPE DIRECT_CELL_OPT >> &END CELL_OPT >> &PRINT >> &TRAJECTORY >> FORMAT PDB >> &END TRAJECTORY >> &END PRINT >> &END MOTION >> >> Thanks for any inputs. >> >> -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Mon Jul 10 17:35:44 2017 From: mattwa... at gmail.com (Matt W) Date: Mon, 10 Jul 2017 10:35:44 -0700 (PDT) Subject: Setting correct cutoff/convergence criteria In-Reply-To: <076ce3f3-b7d7-403c-96a8-d7b6c3535dc5@googlegroups.com> References: <66c6e786-5aa2-479e-b7f8-529cd5e920b6@googlegroups.com> <3acf0ea9-6b29-436e-922d-af950667157b@googlegroups.com> <076ce3f3-b7d7-403c-96a8-d7b6c3535dc5@googlegroups.com> Message-ID: <608ed2c9-9836-4e9d-bebf-35fbff0600a3@googlegroups.com> > > > Thanks for your suggestions. After playing around with the parameters, I > figured the Extrapolation issue and cut-off issue (so fixed that). > I already see improvement in time after some trials. I am wondering though > which method would be best for extrapolation: ASPC etc. > I'd stick with ASPC. PS is maybe a bit faster for geo_opt, but ASPC should be good and robust. > I haven't switched to GAPW though. I will try that. > Only needed really when you have some elements that need stupidly large cutoff (F, Na, Mg) with normal pseudos. It does tend to make forces a bit more accurate, so can speed up final convergence a lot too. > Can you please elaborate on "Mixing and broyden optimizer will be > moderately inefficient for insulators" > and a possible alternative. > > The Orbital Transform (OT) method is CP2Ks flagship optimizer for insulators. For a smallish system &SCF &OUTER_SCF EPS_SCF 5.0E-7 MAX_SCF 1 &END OUTER_SCF SCF_GUESS RESTART EPS_SCF 5.0E-7 MAX_SCF 15 &OT MINIMIZER DIIS PRECONDITIONER FULL_ALL ENERGY_GAP 0.001 &END &END SCF would be my default setup. PRECONDITIONER FULL_SINGLE_INVERSE ENERGY_GAP 0.1 for large systems (>500 atoms or so) Matt > Thanks > >> >>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From jts2t... at gmail.com Tue Jul 11 12:17:01 2017 From: jts2t... at gmail.com (jts2t... at gmail.com) Date: Tue, 11 Jul 2017 05:17:01 -0700 (PDT) Subject: How to calculate the HOMO-LUMO gap in a QMMM simulation? Message-ID: Dear researchers and developers, I am a beginner in CP2K, and I major in computational mathematics. So the concepts about chemistry are new to me. I want to know how to get the HOMO-LUMO gap in a QMMM simulation. The setting in my input file is as follows &PRINT &MO_CUBES NHOMO -1 NLUMO -1 WRITE_CUBE F &END MO_CUBES which locates at the &DFT subsection. But the results seems unreasonable, the result of 2-QM-atom is even larger than the situation if I use full QM calculation. The whole system consists of 48 atoms. But if I specify 12 atoms as QM-atom, the result is less than the result of full QM calculation. So, is there any error in the setting ? -------------- next part -------------- An HTML attachment was scrubbed... URL: From ashishd... at gmail.com Tue Jul 11 14:08:26 2017 From: ashishd... at gmail.com (ashish dabral) Date: Tue, 11 Jul 2017 07:08:26 -0700 (PDT) Subject: Setting correct cutoff/convergence criteria In-Reply-To: <608ed2c9-9836-4e9d-bebf-35fbff0600a3@googlegroups.com> References: <66c6e786-5aa2-479e-b7f8-529cd5e920b6@googlegroups.com> <3acf0ea9-6b29-436e-922d-af950667157b@googlegroups.com> <076ce3f3-b7d7-403c-96a8-d7b6c3535dc5@googlegroups.com> <608ed2c9-9836-4e9d-bebf-35fbff0600a3@googlegroups.com> Message-ID: Thanks a lot Matt for your suggestions and additional clarification. On Monday, 10 July 2017 19:35:44 UTC+2, Matt W wrote: > > >> Thanks for your suggestions. After playing around with the parameters, I >> figured the Extrapolation issue and cut-off issue (so fixed that). >> I already see improvement in time after some trials. I am wondering >> though which method would be best for extrapolation: ASPC etc. >> > > I'd stick with ASPC. PS is maybe a bit faster for geo_opt, but ASPC should > be good and robust. > > >> I haven't switched to GAPW though. I will try that. >> > > Only needed really when you have some elements that need stupidly large > cutoff (F, Na, Mg) with normal pseudos. It does tend to make forces a bit > more accurate, so can speed up final convergence a lot too. > > >> Can you please elaborate on "Mixing and broyden optimizer will be >> moderately inefficient for insulators" >> and a possible alternative. >> >> The Orbital Transform (OT) method is CP2Ks flagship optimizer for > insulators. For a smallish system > > &SCF > &OUTER_SCF > EPS_SCF 5.0E-7 > MAX_SCF 1 > &END OUTER_SCF > SCF_GUESS RESTART > EPS_SCF 5.0E-7 > MAX_SCF 15 > &OT > MINIMIZER DIIS > PRECONDITIONER FULL_ALL > ENERGY_GAP 0.001 > &END > &END SCF > > would be my default setup. > > PRECONDITIONER FULL_SINGLE_INVERSE > ENERGY_GAP 0.1 > > for large systems (>500 atoms or so) > > Matt > > > >> Thanks >> >>> >>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Tue Jul 11 16:09:05 2017 From: mattwa... at gmail.com (Matt W) Date: Tue, 11 Jul 2017 09:09:05 -0700 (PDT) Subject: How to calculate the HOMO-LUMO gap in a QMMM simulation? In-Reply-To: References: Message-ID: <3009d1c9-aff6-4f29-bd35-bc2dc731537b@googlegroups.com> Hi, your setting to get the HOMO-LUMO gap is correct. QMMM is not straight forward to use (in any code). Check you are happy with both the QM and the MM bit before trying to combine them. In general the gap will not be a smooth function of the number of atoms in the QM region - too small and you won't have enough states to form a 'band' and it is likely that the gap will be too large - in between it will depend on the quality of the coupling between QM and MM. Matt On Tuesday, July 11, 2017 at 2:17:01 PM UTC+2, jts2... at gmail.com wrote: > > Dear researchers and developers, > I am a beginner in CP2K, and I major in computational mathematics. So the > concepts about chemistry are new to me. I want to know how to get the > HOMO-LUMO gap in a QMMM simulation. The setting in my input file is as > follows > > &PRINT > &MO_CUBES > NHOMO -1 > NLUMO -1 > WRITE_CUBE F > &END MO_CUBES > > which locates at the &DFT subsection. > > But the results seems unreasonable, the result of 2-QM-atom is even larger > than the situation if I use full QM calculation. The whole system consists > of 48 atoms. But if I specify 12 atoms as QM-atom, the result is less than > the result of full QM calculation. So, is there any error in the setting ? > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From adebe... at gmail.com Tue Jul 11 20:21:15 2017 From: adebe... at gmail.com (Anthony Debellis) Date: Tue, 11 Jul 2017 13:21:15 -0700 (PDT) Subject: Unpaired spins Message-ID: <73ab1088-983c-4065-ba49-e6de68adaf71@googlegroups.com> Dear All, I am studying a system which has separate structures containing a Cu atom and an NO molecule. How can I indicate an atomic guess which places an unpaired spin on each of Cu (+2 ion) and NO molecule in singlet and triplet biradical states? Anthony -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Tue Jul 11 20:27:36 2017 From: mattwa... at gmail.com (Matt W) Date: Tue, 11 Jul 2017 13:27:36 -0700 (PDT) Subject: Unpaired spins In-Reply-To: <73ab1088-983c-4065-ba49-e6de68adaf71@googlegroups.com> References: <73ab1088-983c-4065-ba49-e6de68adaf71@googlegroups.com> Message-ID: Hi, you'll want to explore the infamous BS setting. If you search for BS in the forum you should find some ideas. Probably a bit of trial and error needed. Matt On Tuesday, July 11, 2017 at 10:21:16 PM UTC+2, Anthony Debellis wrote: > > Dear All, > I am studying a system which has separate structures containing a Cu atom > and an NO molecule. How can I indicate an atomic guess which places an > unpaired spin on each of Cu (+2 ion) and NO molecule in singlet and triplet > biradical states? > Anthony > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jts2t... at gmail.com Wed Jul 12 04:27:00 2017 From: jts2t... at gmail.com (jts2t... at gmail.com) Date: Tue, 11 Jul 2017 21:27:00 -0700 (PDT) Subject: How to calculate the HOMO-LUMO gap in a QMMM simulation? In-Reply-To: <3009d1c9-aff6-4f29-bd35-bc2dc731537b@googlegroups.com> References: <3009d1c9-aff6-4f29-bd35-bc2dc731537b@googlegroups.com> Message-ID: <3ce99bab-324e-4768-9073-7fbfcb8955ca@googlegroups.com> Hi Matt, thanks for your reply, I follow the tutorial at https://www.cp2k.org/exercises:2017_ethz_mmm:qmmm, and I also look up the original paper. The question is that my consequence don't coincide with the author's simulation. The more QM atoms I use, the smaller the band gap is. The results of full QM simulation is about 4.9eV, however, the 1-QM layer/2-MM layer is about 4.7, the 2-QM layer/1-MM layer is about 4.5. I did not misunderstand the QM atoms and the MM atoms. I use the QM_KIND subsection to specify the QM atoms. So, what is the problem ? ? 2017?7?12???? UTC+8??12:09:05?Matt W??? > > Hi, > > your setting to get the HOMO-LUMO gap is correct. > > QMMM is not straight forward to use (in any code). Check you are happy > with both the QM and the MM bit before trying to combine them. > > In general the gap will not be a smooth function of the number of atoms in > the QM region - too small and you won't have enough states to form a 'band' > and it is likely that the gap will be too large - in between it will depend > on the quality of the coupling between QM and MM. > > Matt > > > > On Tuesday, July 11, 2017 at 2:17:01 PM UTC+2, jts2... at gmail.com wrote: >> >> Dear researchers and developers, >> I am a beginner in CP2K, and I major in computational mathematics. So the >> concepts about chemistry are new to me. I want to know how to get the >> HOMO-LUMO gap in a QMMM simulation. The setting in my input file is as >> follows >> >> &PRINT >> &MO_CUBES >> NHOMO -1 >> NLUMO -1 >> WRITE_CUBE F >> &END MO_CUBES >> >> which locates at the &DFT subsection. >> >> But the results seems unreasonable, the result of 2-QM-atom is even >> larger than the situation if I use full QM calculation. The whole system >> consists of 48 atoms. But if I specify 12 atoms as QM-atom, the result is >> less than the result of full QM calculation. So, is there any error in the >> setting ? >> >> -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Wed Jul 12 13:30:33 2017 From: mattwa... at gmail.com (Matt W) Date: Wed, 12 Jul 2017 06:30:33 -0700 (PDT) Subject: How to calculate the HOMO-LUMO gap in a QMMM simulation? In-Reply-To: <3ce99bab-324e-4768-9073-7fbfcb8955ca@googlegroups.com> References: <3009d1c9-aff6-4f29-bd35-bc2dc731537b@googlegroups.com> <3ce99bab-324e-4768-9073-7fbfcb8955ca@googlegroups.com> Message-ID: <2553d2a4-ee10-4734-8f4d-d6b0e787f07b@googlegroups.com> Hi, I think the original paper uses four layers. I'm not sure if the MM_RADIUS is the same either (in principle a fitting parameter). To me the results of the exercise look reasonable. Matt On Wednesday, July 12, 2017 at 6:27:01 AM UTC+2, jts2... at gmail.com wrote: > > Hi Matt, thanks for your reply, > > I follow the tutorial at https://www.cp2k.org/exercises:2017_ethz_mmm:qmmm, > and I also look up the original paper. The question is that my consequence > don't coincide with the author's simulation. The more QM atoms I use, the > smaller the band gap is. The results of full QM simulation is about 4.9eV, > however, the 1-QM layer/2-MM layer is about 4.7, the 2-QM layer/1-MM layer > is about 4.5. > > I did not misunderstand the QM atoms and the MM atoms. I use the QM_KIND > subsection to specify the QM atoms. So, what is the problem ? > > ? 2017?7?12???? UTC+8??12:09:05?Matt W??? >> >> Hi, >> >> your setting to get the HOMO-LUMO gap is correct. >> >> QMMM is not straight forward to use (in any code). Check you are happy >> with both the QM and the MM bit before trying to combine them. >> >> In general the gap will not be a smooth function of the number of atoms >> in the QM region - too small and you won't have enough states to form a >> 'band' and it is likely that the gap will be too large - in between it will >> depend on the quality of the coupling between QM and MM. >> >> Matt >> >> >> >> On Tuesday, July 11, 2017 at 2:17:01 PM UTC+2, jts2... at gmail.com wrote: >>> >>> Dear researchers and developers, >>> I am a beginner in CP2K, and I major in computational mathematics. So >>> the concepts about chemistry are new to me. I want to know how to get the >>> HOMO-LUMO gap in a QMMM simulation. The setting in my input file is as >>> follows >>> >>> &PRINT >>> &MO_CUBES >>> NHOMO -1 >>> NLUMO -1 >>> WRITE_CUBE F >>> &END MO_CUBES >>> >>> which locates at the &DFT subsection. >>> >>> But the results seems unreasonable, the result of 2-QM-atom is even >>> larger than the situation if I use full QM calculation. The whole system >>> consists of 48 atoms. But if I specify 12 atoms as QM-atom, the result is >>> less than the result of full QM calculation. So, is there any error in the >>> setting ? >>> >>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From yanghon... at gmail.com Wed Jul 12 14:55:12 2017 From: yanghon... at gmail.com (Hongliu) Date: Wed, 12 Jul 2017 07:55:12 -0700 (PDT) Subject: geometry optimization of Na slab Message-ID: <8bf66b27-96d8-4206-aedc-abbf3a0c4164@googlegroups.com> Dear CP2Ker: I was studying the binding energy of a Na atom to a Na slab. I tried first to optimize the structure. When I used the attached input file the structure of the system changed only very slightly, even if I put the last Na atom to a unrealistic position. I am not sure what goes wrong with my files. Please help me out. Thanks! BTW, I am using version 2.6.1. Best wishes, Hongliu -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: pf0-86-re.inp Type: chemical/x-gamess-input Size: 1830 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Na_111_663_slab.xyz Type: chemical/x-xyz Size: 7135 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: pf0-86-re.inp Type: chemical/x-gamess-input Size: 1830 bytes Desc: not available URL: From bharats... at gmail.com Wed Jul 12 17:13:06 2017 From: bharats... at gmail.com (bharat) Date: Wed, 12 Jul 2017 10:13:06 -0700 (PDT) Subject: CP2K parameters change in input file when simulating a larger box Message-ID: Hello everyone, If we were to test the box size whether it is enough or not, what CP2K parameters should change keeping other parameters same. I am thinking of Coulomb CUTOFF (which is usually half of the box) and BOX size. I would like to compare properties such as heat of vaporization and dipole moments. Is there any other parameters I should change while simulating a larger box? Any suggestions are appreciated. I am simulating under periodic boundary conditions. Thank you. Bharat -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Wed Jul 12 20:21:04 2017 From: mattwa... at gmail.com (Matt W) Date: Wed, 12 Jul 2017 13:21:04 -0700 (PDT) Subject: geometry optimization of Na slab In-Reply-To: <8bf66b27-96d8-4206-aedc-abbf3a0c4164@googlegroups.com> References: <8bf66b27-96d8-4206-aedc-abbf3a0c4164@googlegroups.com> Message-ID: <41b6bde6-6748-4f01-8a06-ad2441840387@googlegroups.com> Not quite sure. Maybe try deleting all of your GEO_OPT section (i.e. going back to the defaults) and see if that works better? Matt On Wednesday, July 12, 2017 at 4:55:12 PM UTC+2, Hongliu wrote: > > Dear CP2Ker: > > I was studying the binding energy of a Na atom to a Na slab. I tried first > to optimize the structure. When I used the attached input file the > structure of the system changed only very slightly, even if I put the last > Na atom to a unrealistic position. I am not sure what goes wrong with my > files. Please help me out. Thanks! > > BTW, I am using version 2.6.1. > > Best wishes, > Hongliu > -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Wed Jul 12 20:24:06 2017 From: mattwa... at gmail.com (Matt W) Date: Wed, 12 Jul 2017 13:24:06 -0700 (PDT) Subject: CP2K parameters change in input file when simulating a larger box In-Reply-To: References: Message-ID: <124960c3-c3fd-4e9a-bee6-91ed3fced938@googlegroups.com> Opinions may vary, but I'd keep the cutoff the same - it essentially defines a functional, and only makes sense to compare like with like. I can't think of anything else that needs to be changed as you alter the box size (in DFT calculation). Matt On Wednesday, July 12, 2017 at 7:13:06 PM UTC+2, bharat wrote: > > Hello everyone, > > If we were to test the box size whether it is enough or not, what CP2K > parameters should change keeping other parameters same. I am thinking of > Coulomb CUTOFF (which is usually half of the box) and BOX size. I would > like to compare properties such as heat of vaporization and dipole moments. > > Is there any other parameters I should change while simulating a larger > box? > > Any suggestions are appreciated. I am simulating under periodic boundary > conditions. > > Thank you. > > Bharat > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jts2t... at gmail.com Thu Jul 13 01:20:47 2017 From: jts2t... at gmail.com (jts2t... at gmail.com) Date: Wed, 12 Jul 2017 18:20:47 -0700 (PDT) Subject: How to calculate the HOMO-LUMO gap in a QMMM simulation? In-Reply-To: <2553d2a4-ee10-4734-8f4d-d6b0e787f07b@googlegroups.com> References: <3009d1c9-aff6-4f29-bd35-bc2dc731537b@googlegroups.com> <3ce99bab-324e-4768-9073-7fbfcb8955ca@googlegroups.com> <2553d2a4-ee10-4734-8f4d-d6b0e787f07b@googlegroups.com> Message-ID: I did several simulations in the same system, which consists of 24 K atoms and 24 Cl atoms. The x-axis of the picture is the number of atoms in QM region, and the y-axis is the HOMO-LUMO gap that each simulation finally get. The tendency at the beginning seems strange. ? 2017?7?12???? UTC+8??9:30:33?Matt W??? > > Hi, > > I think the original paper uses four layers. I'm not sure if the MM_RADIUS > is the same either (in principle a fitting parameter). > > To me the results of the exercise look reasonable. > > Matt > > On Wednesday, July 12, 2017 at 6:27:01 AM UTC+2, jts2... at gmail.com wrote: >> >> Hi Matt, thanks for your reply, >> >> I follow the tutorial at >> https://www.cp2k.org/exercises:2017_ethz_mmm:qmmm, and I also look up >> the original paper. The question is that my consequence don't coincide with >> the author's simulation. The more QM atoms I use, the smaller the band gap >> is. The results of full QM simulation is about 4.9eV, however, the 1-QM >> layer/2-MM layer is about 4.7, the 2-QM layer/1-MM layer is about 4.5. >> >> I did not misunderstand the QM atoms and the MM atoms. I use the QM_KIND >> subsection to specify the QM atoms. So, what is the problem ? >> >> ? 2017?7?12???? UTC+8??12:09:05?Matt W??? >>> >>> Hi, >>> >>> your setting to get the HOMO-LUMO gap is correct. >>> >>> QMMM is not straight forward to use (in any code). Check you are happy >>> with both the QM and the MM bit before trying to combine them. >>> >>> In general the gap will not be a smooth function of the number of atoms >>> in the QM region - too small and you won't have enough states to form a >>> 'band' and it is likely that the gap will be too large - in between it will >>> depend on the quality of the coupling between QM and MM. >>> >>> Matt >>> >>> >>> >>> On Tuesday, July 11, 2017 at 2:17:01 PM UTC+2, jts2... at gmail.com wrote: >>>> >>>> Dear researchers and developers, >>>> I am a beginner in CP2K, and I major in computational mathematics. So >>>> the concepts about chemistry are new to me. I want to know how to get the >>>> HOMO-LUMO gap in a QMMM simulation. The setting in my input file is as >>>> follows >>>> >>>> &PRINT >>>> &MO_CUBES >>>> NHOMO -1 >>>> NLUMO -1 >>>> WRITE_CUBE F >>>> &END MO_CUBES >>>> >>>> which locates at the &DFT subsection. >>>> >>>> But the results seems unreasonable, the result of 2-QM-atom is even >>>> larger than the situation if I use full QM calculation. The whole system >>>> consists of 48 atoms. But if I specify 12 atoms as QM-atom, the result is >>>> less than the result of full QM calculation. So, is there any error in the >>>> setting ? >>>> >>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From zuohen... at gmail.com Thu Jul 13 01:41:31 2017 From: zuohen... at gmail.com (zuohen... at gmail.com) Date: Wed, 12 Jul 2017 18:41:31 -0700 (PDT) Subject: Run cp2k test error for ADMM (QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp) In-Reply-To: <14e1e7fd-ad74-4dca-8365-2759b9754341@googlegroups.com> References: <14e1e7fd-ad74-4dca-8365-2759b9754341@googlegroups.com> Message-ID: <2eb28df6-debf-4552-9a0d-4cfb8af599e7@googlegroups.com> Hi, my friend. I meet the same problem now. The calculation of ADMM was interrupted but no error information in .out file. The '*SIGSEGV, segmentation fault occurred' *was shown in the screen. I want to know if you solve this problem and hope get your kind help. best regars, Zuoheng ? 2017?5?20???? UTC+8??12:03:41?zhj... at gmail.com??? > > Hi, Alfio Lazzaro, > Thank you for your response. The angular momentum parameters in libint.h > and libderiv.h are: > > *#define LIBINT_MAX_AM 6* > *#define LIBINT_OPT_AM 5* > > *#define LIBDERIV_MAX_AM1 5#define LIBDERIV_MAX_AM12 4* > > The configure parameters of libint are: > *./configure --prefix=/opt/libint-1.4.4 --with-ar=ar FC=ifort F77=ifort **F90=ifort > FCFLAGS="-O3 -xHost" CC=icc CFLAGS="-O3 -xHost" \* > *CXX=icpc CXXFLAGS="-O3 -xHost" --with-cc-optflags="-O2 -xHost" **--with-cxx-optflags="-O2 > -xHost" --with-libint-max-am=5 --with-libint-opt-am=4 \* > *--with-libderiv-max-am1=4* > > I met some strange things: > 1. when the "*-xHost*" optflag is removed, then the test > "QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp" will run successfully(but has a > warning of "SCF run NOT converged"); > 2. When I use the cp2k version in 1 to calculate my example, then the > error is occurred: *application called MPI_Abort(MPI_COMM_WORLD, -10) - > process 0* > > I will attach some files to you. Thank you for your help!! > > Best regards, > Zhjs > > ? 2017?5?19???? UTC+8??4:01:08?Alfio Lazzaro??? >> >> Hi, >> could you compile the code with the debug symbols (i.e. pdbg versin, or >> alternative with -g)? >> My suspicious is that there is a mismatch in the angular momentum used in >> libderiv and CP2K. >> From your output CP2K uses libderiv_max_am1=5 libint_max_am=6, could you >> check what are the values used to compile the library? >> You can check your include/libint/libint.h in libint to see if it matches. >> >> Cheers, >> >> Alfio >> >> >> >> Il giorno gioved? 18 maggio 2017 17:39:21 UTC+2, zhj... at gmail.com ha >> scritto: >>> >>> Dear all, >>> I am trying to run the test of QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp, >>> but an Error message appeared immediately after starting this calculation: >>> >>> *forrtl: severe (174): SIGSEGV, segmentation fault occurred* >>> *Image PC Routine Line >>> Source* >>> *cp2k.popt 00000000076355A1 Unknown Unknown >>> Unknown* >>> *cp2k.popt 0000000007633CF7 Unknown Unknown >>> Unknown* >>> *cp2k.popt 00000000075EF624 Unknown Unknown >>> Unknown* >>> *cp2k.popt 00000000075EF436 Unknown Unknown >>> Unknown* >>> *cp2k.popt 000000000759E2FF Unknown Unknown >>> Unknown* >>> *cp2k.popt 00000000075A561D Unknown Unknown >>> Unknown* >>> *libpthread.so.0 00007F7F9B26E100 Unknown Unknown >>> Unknown* >>> *cp2k.popt 000000000132A4BE Unknown Unknown >>> Unknown* >>> *cp2k.popt 000000000084877E Unknown Unknown >>> Unknown* >>> *cp2k.popt 0000000000843CA5 Unknown Unknown >>> Unknown* >>> *cp2k.popt 0000000000A33524 Unknown Unknown >>> Unknown* >>> *cp2k.popt 0000000000A31563 Unknown Unknown >>> Unknown* >>> *cp2k.popt 00000000009A9289 Unknown Unknown >>> Unknown* >>> *cp2k.popt 00000000016F6EB8 Unknown Unknown >>> Unknown* >>> *cp2k.popt 00000000016F5619 Unknown Unknown >>> Unknown* >>> *cp2k.popt 0000000000E16383 Unknown Unknown >>> Unknown* >>> *cp2k.popt 000000000041326D Unknown Unknown >>> Unknown* >>> *cp2k.popt 00000000004085B2 Unknown Unknown >>> Unknown* >>> *cp2k.popt 0000000000407B1E Unknown Unknown >>> Unknown* >>> *libc.so.6 00007F7F99A94B15 Unknown Unknown >>> Unknown* >>> *cp2k.popt 0000000000407A29 Unknown Unknown >>> Unknown* >>> >>> If the section of *ADMM* and* AUX_FIT_BASIS_SET* is removed, then the >>> test will run without error occurred. >>> Can anyone tell me how to solve this proplem ? Thanks ! >>> >>> >>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From zhj... at gmail.com Thu Jul 13 12:07:53 2017 From: zhj... at gmail.com (zhj... at gmail.com) Date: Thu, 13 Jul 2017 05:07:53 -0700 (PDT) Subject: Run cp2k test error for ADMM (QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp) In-Reply-To: <2eb28df6-debf-4552-9a0d-4cfb8af599e7@googlegroups.com> References: <14e1e7fd-ad74-4dca-8365-2759b9754341@googlegroups.com> <2eb28df6-debf-4552-9a0d-4cfb8af599e7@googlegroups.com> Message-ID: Hi, you can update the newest cp2k development version(e.g. r1798 version), and this error can be removed. ? 2017?7?13???? UTC+8??9:41:32?zuoh... at gmail.com??? > > Hi, my friend. > > I meet the same problem now. The calculation of ADMM was interrupted but > no error information in .out file. The '*SIGSEGV, segmentation fault > occurred' *was shown in the screen. > I want to know if you solve this problem and hope get your kind help. > > best regars, > Zuoheng > > > ? 2017?5?20???? UTC+8??12:03:41?zhj... at gmail.com??? >> >> Hi, Alfio Lazzaro, >> Thank you for your response. The angular momentum parameters in libint.h >> and libderiv.h are: >> >> *#define LIBINT_MAX_AM 6* >> *#define LIBINT_OPT_AM 5* >> >> *#define LIBDERIV_MAX_AM1 5#define LIBDERIV_MAX_AM12 4* >> >> The configure parameters of libint are: >> *./configure --prefix=/opt/libint-1.4.4 --with-ar=ar FC=ifort F77=ifort **F90=ifort >> FCFLAGS="-O3 -xHost" CC=icc CFLAGS="-O3 -xHost" \* >> *CXX=icpc CXXFLAGS="-O3 -xHost" --with-cc-optflags="-O2 -xHost" **--with-cxx-optflags="-O2 >> -xHost" --with-libint-max-am=5 --with-libint-opt-am=4 \* >> *--with-libderiv-max-am1=4* >> >> I met some strange things: >> 1. when the "*-xHost*" optflag is removed, then the test >> "QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp" will run successfully(but has a >> warning of "SCF run NOT converged"); >> 2. When I use the cp2k version in 1 to calculate my example, then the >> error is occurred: *application called MPI_Abort(MPI_COMM_WORLD, -10) - >> process 0* >> >> I will attach some files to you. Thank you for your help!! >> >> Best regards, >> Zhjs >> >> ? 2017?5?19???? UTC+8??4:01:08?Alfio Lazzaro??? >>> >>> Hi, >>> could you compile the code with the debug symbols (i.e. pdbg versin, or >>> alternative with -g)? >>> My suspicious is that there is a mismatch in the angular momentum used >>> in libderiv and CP2K. >>> From your output CP2K uses libderiv_max_am1=5 libint_max_am=6, could you >>> check what are the values used to compile the library? >>> You can check your include/libint/libint.h in libint to see if it >>> matches. >>> >>> Cheers, >>> >>> Alfio >>> >>> >>> >>> Il giorno gioved? 18 maggio 2017 17:39:21 UTC+2, zhj... at gmail.com ha >>> scritto: >>>> >>>> Dear all, >>>> I am trying to run the test >>>> of QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp, but an Error message appeared >>>> immediately after starting this calculation: >>>> >>>> *forrtl: severe (174): SIGSEGV, segmentation fault occurred* >>>> *Image PC Routine Line >>>> Source* >>>> *cp2k.popt 00000000076355A1 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 0000000007633CF7 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 00000000075EF624 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 00000000075EF436 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 000000000759E2FF Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 00000000075A561D Unknown Unknown >>>> Unknown* >>>> *libpthread.so.0 00007F7F9B26E100 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 000000000132A4BE Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 000000000084877E Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 0000000000843CA5 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 0000000000A33524 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 0000000000A31563 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 00000000009A9289 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 00000000016F6EB8 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 00000000016F5619 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 0000000000E16383 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 000000000041326D Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 00000000004085B2 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 0000000000407B1E Unknown Unknown >>>> Unknown* >>>> *libc.so.6 00007F7F99A94B15 Unknown Unknown >>>> Unknown* >>>> *cp2k.popt 0000000000407A29 Unknown Unknown >>>> Unknown* >>>> >>>> If the section of *ADMM* and* AUX_FIT_BASIS_SET* is removed, then the >>>> test will run without error occurred. >>>> Can anyone tell me how to solve this proplem ? Thanks ! >>>> >>>> >>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From yanghon... at gmail.com Thu Jul 13 14:41:39 2017 From: yanghon... at gmail.com (Hongliu) Date: Thu, 13 Jul 2017 07:41:39 -0700 (PDT) Subject: geometry optimization of Na slab In-Reply-To: <41b6bde6-6748-4f01-8a06-ad2441840387@googlegroups.com> References: <8bf66b27-96d8-4206-aedc-abbf3a0c4164@googlegroups.com> <41b6bde6-6748-4f01-8a06-ad2441840387@googlegroups.com> Message-ID: <9d3e0298-89c5-4300-a0bc-cc0250ab2ebb@googlegroups.com> Dear Matt, it really helps! Now the structure changes indeed. What could be the reason for that? But it changes too much :(. For a 3-layer (111) slab I fixed the first layer and the upper layer becomes a mass after a while. BTW, for the cell optimization of Na bcc crystal structure I got always a disordered state. This was also reported by Aniruddha Dive at this post: https://groups.google.com/forum/#!topic/cp2k/aHejgB31V7s. My input files for the cell optimization are attached. Any suggestions for that? Thanks! Best wishes, Hongliu On Wednesday, July 12, 2017 at 10:21:04 PM UTC+2, Matt W wrote: > > Not quite sure. > > Maybe try deleting all of your GEO_OPT section (i.e. going back to the > defaults) and see if that works better? > > Matt > > On Wednesday, July 12, 2017 at 4:55:12 PM UTC+2, Hongliu wrote: >> >> Dear CP2Ker: >> >> I was studying the binding energy of a Na atom to a Na slab. I tried >> first to optimize the structure. When I used the attached input file the >> structure of the system changed only very slightly, even if I put the last >> Na atom to a unrealistic position. I am not sure what goes wrong with my >> files. Please help me out. Thanks! >> >> BTW, I am using version 2.6.1. >> >> Best wishes, >> Hongliu >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: pf0-86-re.inp Type: chemical/x-gamess-input Size: 1829 bytes Desc: not available URL: From mattwa... at gmail.com Thu Jul 13 14:49:17 2017 From: mattwa... at gmail.com (Matt W) Date: Thu, 13 Jul 2017 07:49:17 -0700 (PDT) Subject: geometry optimization of Na slab In-Reply-To: <9d3e0298-89c5-4300-a0bc-cc0250ab2ebb@googlegroups.com> References: <8bf66b27-96d8-4206-aedc-abbf3a0c4164@googlegroups.com> <41b6bde6-6748-4f01-8a06-ad2441840387@googlegroups.com> <9d3e0298-89c5-4300-a0bc-cc0250ab2ebb@googlegroups.com> Message-ID: <6dc2bc14-757a-482d-862f-dc310c60dc8b@googlegroups.com> Try using the GAPW method (change GPW to GAPW in &QS section ). Na is problematic, the pseudopotential is extremely hard. Matt On Thursday, July 13, 2017 at 4:41:39 PM UTC+2, Hongliu wrote: > > Dear Matt, > > it really helps! Now the structure changes indeed. What could be the > reason for that? > > But it changes too much :(. For a 3-layer (111) slab I fixed the first > layer and the upper layer becomes a mass after a while. BTW, for the cell > optimization of Na bcc crystal structure I got always a disordered state. > This was also reported by Aniruddha Dive at this post: > https://groups.google.com/forum/#!topic/cp2k/aHejgB31V7s. > > My input files for the cell optimization are attached. Any suggestions for > that? Thanks! > > Best wishes, > Hongliu > > On Wednesday, July 12, 2017 at 10:21:04 PM UTC+2, Matt W wrote: >> >> Not quite sure. >> >> Maybe try deleting all of your GEO_OPT section (i.e. going back to the >> defaults) and see if that works better? >> >> Matt >> >> On Wednesday, July 12, 2017 at 4:55:12 PM UTC+2, Hongliu wrote: >>> >>> Dear CP2Ker: >>> >>> I was studying the binding energy of a Na atom to a Na slab. I tried >>> first to optimize the structure. When I used the attached input file the >>> structure of the system changed only very slightly, even if I put the last >>> Na atom to a unrealistic position. I am not sure what goes wrong with my >>> files. Please help me out. Thanks! >>> >>> BTW, I am using version 2.6.1. >>> >>> Best wishes, >>> Hongliu >>> >> -------------- next part -------------- An HTML attachment was scrubbed... URL: From taosun... at gmail.com Thu Jul 13 15:22:13 2017 From: taosun... at gmail.com (Sun Tao) Date: Thu, 13 Jul 2017 08:22:13 -0700 (PDT) Subject: confusion about QM/MM setup in CP2k In-Reply-To: References: <988d807c-e79f-414f-a380-ecb8b3833088@googlegroups.com> <6e81e560-704b-4bea-bbf9-f3142f813a0e@googlegroups.com> <870e5215-dee6-40bf-82f1-27ea922cb2aa@googlegroups.com> Message-ID: Dear Matt: I am trying to carry out some simulations with periodic boundary condition. However, no matter ab initio MD or QM/MM simulations gave me some configurations with atoms moving out of the cell box. Please find the attached snapshots for the simulation of water inside carbon nanotube. The initial setup is in picture 0.0ps.png with the rectangular simulation box. The configuration after 1.2 picoseconds, the atoms are moving out of the simulation box. Both the snapshots were taken from the trajectory file *-pos-1.xyz file. I am also attaching the input files. In FORCE_EVAL/DFT/POISSON section, the PERIODIC tag was set to XYZ, and in FORCE_EVAL/SUBSYS/CELL section, the PERIODIC tag was not set and the default should also be XYZ. Then I don't know why the atoms are moving out of the cell box. Sorry to bother you again. Thanks very much for your help. Tao ? 2017?5?26???? UTC-5??6:07:41?Matt W??? > > >> Thanks very much for your help! I checked the 2 papers, here is my >> understanding. The 2005 paper about QM/MM electrostatic coupling does allow >> MM atoms to be inside the QM box. I guess the QM box setup in CP2K is just >> to apply the scheme in the 2006 paper to calculate the electrostatic >> interaction within periodic boundary condition. I have some more questions: >> >> 1. Does the size of the QM box affect the simulation speed a lot? I could >> just set up the QM box to be the same as the overall system box, then >> specify some important part to be treated in QM fasion while other atoms in >> MM. Is it necessary to set up a smaller QM box? >> >> Depends on the size/density of your system, for various reasons. You > might pay quite a lot for empty space using CP2K in a small system, but > negligible for a larger one. > > > >> 2. If I set up a smaller QM box, the the box boundary cuts through a >> covalent bond, do I need to cap the dangling bond with hydrogen atoms as >> what is usually done for QM/MM? >> >> Yes, you'd need to cap the dangling bonds, entirely as you'd expect. > > > >> 3. Inside a QM box, for example there is a covalent bond A-B. Now I set >> atom A to be treated in QM, while atom B to be treated in MM, does this >> seem to be a correct way? Further more the covalent bond A-B is actually >> within a piece of solid, and the solid is kept fixed (does not move). >> >> As long as you have some forcefield that can describe the interaction. > Again, the boxes are pretty much for electrostatics, what really matters is > whether there are changes in the bonding of the system. > > >> Thanks again for your help! >> >> Tao >> >> ? 2017?5?25???? UTC-5??4:25:27?Matt W??? >>> >>> Hi Tao, >>> >>> The boxes are used essentially for electrostatic calculations. The >>> interface between the QM and MM atoms is the real region where they >>> connect, not really anything to do with the cells. >>> >>> You should see a list of relevant references at the end of the output >>> file - there are two particular to CP2K by Laio and co-workers that you >>> should try and read. >>> >>> Matt >>> >>> On Thursday, May 25, 2017 at 6:09:31 PM UTC+1, Sun Tao wrote: >>>> >>>> Dear experts: >>>> >>>> I am confused about some settings of the QMMM portion of CP2K. In the >>>> section QMMM/CELL we need to set up the QM region. Then in QMMM/QM_KIND >>>> section, there is a keyword MM_INDEX, for which my understanding is to tell >>>> the program the index of the MM atoms that would be treated in QM fashion. >>>> >>>> Here is the question: does CP2K allow MM atoms inside QM regions? Or >>>> say in the QM region some atoms could be treated in QM and others could >>>> still be treated in MM? If CP2K does allow this, how do these two different >>>> kinds of atoms (QM atoms and MM atoms) interact with each other inside the >>>> QM box? Is there any papers/books I could refer to if this is the case? >>>> Before I only know that the coupling happens between QM region and MM >>>> region and the boundary between QM and MM is important to treat. >>>> >>>> Thanks very much for your help. Look forward to your reply! >>>> >>>> Tao >>>> >>> -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: 0.0ps.png Type: image/png Size: 661443 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: 1.2ps.png Type: image/png Size: 688352 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cnt.cp2k.inp Type: chemical/x-gamess-input Size: 3355 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: dftd3.dat Type: application/octet-stream Size: 1959318 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: GTH_POTENTIALS Type: application/octet-stream Size: 125422 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: HFX_BASIS Type: application/octet-stream Size: 17275 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: total_cnt_water.xyz Type: chemical/x-xyz Size: 33855 bytes Desc: not available URL: From mattwa... at gmail.com Thu Jul 13 15:42:56 2017 From: mattwa... at gmail.com (Matt W) Date: Thu, 13 Jul 2017 08:42:56 -0700 (PDT) Subject: confusion about QM/MM setup in CP2k In-Reply-To: References: <988d807c-e79f-414f-a380-ecb8b3833088@googlegroups.com> <6e81e560-704b-4bea-bbf9-f3142f813a0e@googlegroups.com> <870e5215-dee6-40bf-82f1-27ea922cb2aa@googlegroups.com> Message-ID: <8606497d-fcbd-42b3-a494-7e245381b446@googlegroups.com> Hi, there is no outside a periodic box; if you leave one side you come back in the other. In VMD or most visualisers you can wrap atoms back into the box. Matt On Thursday, July 13, 2017 at 5:22:13 PM UTC+2, Sun Tao wrote: > > Dear Matt: > > I am trying to carry out some simulations with periodic boundary > condition. However, no matter ab initio MD or QM/MM simulations gave me > some configurations with atoms moving out of the cell box. Please find the > attached snapshots for the simulation of water inside carbon nanotube. The > initial setup is in picture 0.0ps.png with the rectangular simulation box. > The configuration after 1.2 picoseconds, the atoms are moving out of the > simulation box. Both the snapshots were taken from the trajectory file *- > pos-1.xyz file. > > I am also attaching the input files. In FORCE_EVAL/DFT/POISSON section, > the PERIODIC tag was set to XYZ, and in FORCE_EVAL/SUBSYS/CELL section, > the PERIODIC tag was not set and the default should also be XYZ. Then I > don't know why the atoms are moving out of the cell box. > > Sorry to bother you again. Thanks very much for your help. > > Tao > > > ? 2017?5?26???? UTC-5??6:07:41?Matt W??? >> >> >>> Thanks very much for your help! I checked the 2 papers, here is my >>> understanding. The 2005 paper about QM/MM electrostatic coupling does allow >>> MM atoms to be inside the QM box. I guess the QM box setup in CP2K is just >>> to apply the scheme in the 2006 paper to calculate the electrostatic >>> interaction within periodic boundary condition. I have some more questions: >>> >>> 1. Does the size of the QM box affect the simulation speed a lot? I >>> could just set up the QM box to be the same as the overall system box, then >>> specify some important part to be treated in QM fasion while other atoms in >>> MM. Is it necessary to set up a smaller QM box? >>> >>> Depends on the size/density of your system, for various reasons. You >> might pay quite a lot for empty space using CP2K in a small system, but >> negligible for a larger one. >> >> >> >>> 2. If I set up a smaller QM box, the the box boundary cuts through a >>> covalent bond, do I need to cap the dangling bond with hydrogen atoms as >>> what is usually done for QM/MM? >>> >>> Yes, you'd need to cap the dangling bonds, entirely as you'd expect. >> >> >> >>> 3. Inside a QM box, for example there is a covalent bond A-B. Now I set >>> atom A to be treated in QM, while atom B to be treated in MM, does this >>> seem to be a correct way? Further more the covalent bond A-B is actually >>> within a piece of solid, and the solid is kept fixed (does not move). >>> >>> As long as you have some forcefield that can describe the interaction. >> Again, the boxes are pretty much for electrostatics, what really matters is >> whether there are changes in the bonding of the system. >> >> >>> Thanks again for your help! >>> >>> Tao >>> >>> ? 2017?5?25???? UTC-5??4:25:27?Matt W??? >>>> >>>> Hi Tao, >>>> >>>> The boxes are used essentially for electrostatic calculations. The >>>> interface between the QM and MM atoms is the real region where they >>>> connect, not really anything to do with the cells. >>>> >>>> You should see a list of relevant references at the end of the output >>>> file - there are two particular to CP2K by Laio and co-workers that you >>>> should try and read. >>>> >>>> Matt >>>> >>>> On Thursday, May 25, 2017 at 6:09:31 PM UTC+1, Sun Tao wrote: >>>>> >>>>> Dear experts: >>>>> >>>>> I am confused about some settings of the QMMM portion of CP2K. In the >>>>> section QMMM/CELL we need to set up the QM region. Then in QMMM/QM_KIND >>>>> section, there is a keyword MM_INDEX, for which my understanding is to tell >>>>> the program the index of the MM atoms that would be treated in QM fashion. >>>>> >>>>> Here is the question: does CP2K allow MM atoms inside QM regions? Or >>>>> say in the QM region some atoms could be treated in QM and others could >>>>> still be treated in MM? If CP2K does allow this, how do these two different >>>>> kinds of atoms (QM atoms and MM atoms) interact with each other inside the >>>>> QM box? Is there any papers/books I could refer to if this is the case? >>>>> Before I only know that the coupling happens between QM region and MM >>>>> region and the boundary between QM and MM is important to treat. >>>>> >>>>> Thanks very much for your help. Look forward to your reply! >>>>> >>>>> Tao >>>>> >>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From wei... at gmail.com Fri Jul 14 02:32:25 2017 From: wei... at gmail.com (wei... at gmail.com) Date: Thu, 13 Jul 2017 19:32:25 -0700 (PDT) Subject: hybrid functional and symmetry with k-point Message-ID: <8d468847-237b-4a5b-a816-c8cf9935af0b@googlegroups.com> Dear CP2K developers, May I ask if doing hybrid functional calculations with k-point has been implemented in the latest trunk of CP2K? There was a post in this forum in 2015 asking this question (https://groups.google.com/forum/#!topic/cp2k/2XxNUt8tML8) and Professor Hutter answer at that time it was not yet implemented. How about now? Also, according to this FAQ item https://www.cp2k.org/faq:kpoints, reduction of k-points using symmetry is not supported. Is there any development on that planned? Thanks, Wei -------------- next part -------------- An HTML attachment was scrubbed... URL: From zuohen... at gmail.com Fri Jul 14 03:03:33 2017 From: zuohen... at gmail.com (zuohen... at gmail.com) Date: Thu, 13 Jul 2017 20:03:33 -0700 (PDT) Subject: Run cp2k test error for ADMM (QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp) In-Reply-To: References: <14e1e7fd-ad74-4dca-8365-2759b9754341@googlegroups.com> <2eb28df6-debf-4552-9a0d-4cfb8af599e7@googlegroups.com> Message-ID: Thank you very much! ? 2017?7?13???? UTC+8??8:07:53?zhj... at gmail.com??? > > Hi, you can update the newest cp2k development version(e.g. r1798 > version), and this error can be removed. > > ? 2017?7?13???? UTC+8??9:41:32?zuoh... at gmail.com??? >> >> Hi, my friend. >> >> I meet the same problem now. The calculation of ADMM was interrupted but >> no error information in .out file. The '*SIGSEGV, segmentation fault >> occurred' *was shown in the screen. >> I want to know if you solve this problem and hope get your kind help. >> >> best regars, >> Zuoheng >> >> >> ? 2017?5?20???? UTC+8??12:03:41?zhj... at gmail.com??? >>> >>> Hi, Alfio Lazzaro, >>> Thank you for your response. The angular momentum parameters in libint.h >>> and libderiv.h are: >>> >>> *#define LIBINT_MAX_AM 6* >>> *#define LIBINT_OPT_AM 5* >>> >>> *#define LIBDERIV_MAX_AM1 5#define LIBDERIV_MAX_AM12 4* >>> >>> The configure parameters of libint are: >>> *./configure --prefix=/opt/libint-1.4.4 --with-ar=ar FC=ifort F77=ifort **F90=ifort >>> FCFLAGS="-O3 -xHost" CC=icc CFLAGS="-O3 -xHost" \* >>> *CXX=icpc CXXFLAGS="-O3 -xHost" --with-cc-optflags="-O2 -xHost" **--with-cxx-optflags="-O2 >>> -xHost" --with-libint-max-am=5 --with-libint-opt-am=4 \* >>> *--with-libderiv-max-am1=4* >>> >>> I met some strange things: >>> 1. when the "*-xHost*" optflag is removed, then the test >>> "QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp" will run successfully(but has a >>> warning of "SCF run NOT converged"); >>> 2. When I use the cp2k version in 1 to calculate my example, then the >>> error is occurred: *application called MPI_Abort(MPI_COMM_WORLD, -10) - >>> process 0* >>> >>> I will attach some files to you. Thank you for your help!! >>> >>> Best regards, >>> Zhjs >>> >>> ? 2017?5?19???? UTC+8??4:01:08?Alfio Lazzaro??? >>>> >>>> Hi, >>>> could you compile the code with the debug symbols (i.e. pdbg versin, or >>>> alternative with -g)? >>>> My suspicious is that there is a mismatch in the angular momentum used >>>> in libderiv and CP2K. >>>> From your output CP2K uses libderiv_max_am1=5 libint_max_am=6, could >>>> you check what are the values used to compile the library? >>>> You can check your include/libint/libint.h in libint to see if it >>>> matches. >>>> >>>> Cheers, >>>> >>>> Alfio >>>> >>>> >>>> >>>> Il giorno gioved? 18 maggio 2017 17:39:21 UTC+2, zhj... at gmail.com ha >>>> scritto: >>>>> >>>>> Dear all, >>>>> I am trying to run the test >>>>> of QS/regtest-hfx-wfn-fitting/CH4-HSE06.inp, but an Error message appeared >>>>> immediately after starting this calculation: >>>>> >>>>> *forrtl: severe (174): SIGSEGV, segmentation fault occurred* >>>>> *Image PC Routine Line >>>>> Source* >>>>> *cp2k.popt 00000000076355A1 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 0000000007633CF7 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 00000000075EF624 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 00000000075EF436 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 000000000759E2FF Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 00000000075A561D Unknown Unknown >>>>> Unknown* >>>>> *libpthread.so.0 00007F7F9B26E100 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 000000000132A4BE Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 000000000084877E Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 0000000000843CA5 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 0000000000A33524 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 0000000000A31563 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 00000000009A9289 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 00000000016F6EB8 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 00000000016F5619 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 0000000000E16383 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 000000000041326D Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 00000000004085B2 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 0000000000407B1E Unknown Unknown >>>>> Unknown* >>>>> *libc.so.6 00007F7F99A94B15 Unknown Unknown >>>>> Unknown* >>>>> *cp2k.popt 0000000000407A29 Unknown Unknown >>>>> Unknown* >>>>> >>>>> If the section of *ADMM* and* AUX_FIT_BASIS_SET* is removed, then the >>>>> test will run without error occurred. >>>>> Can anyone tell me how to solve this proplem ? Thanks ! >>>>> >>>>> >>>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From hut... at chem.uzh.ch Fri Jul 14 06:55:42 2017 From: hut... at chem.uzh.ch (hut... at chem.uzh.ch) Date: Fri, 14 Jul 2017 08:55:42 +0200 Subject: [CP2K:9218] hybrid functional and symmetry with k-point In-Reply-To: <8d468847-237b-4a5b-a816-c8cf9935af0b@googlegroups.com> References: <8d468847-237b-4a5b-a816-c8cf9935af0b@googlegroups.com> Message-ID: Hi both of these options are still not implemented in CP2K. Symmetry reduction of k-points might be available soon. HFX for k-point is planned but nobody is working on it right now. regards Juerg -------------------------------------------------------------- Juerg Hutter Phone : ++41 44 635 4491 Institut f?r Chemie C FAX : ++41 44 635 6838 Universit?t Z?rich E-mail: hut... at chem.uzh.ch Winterthurerstrasse 190 CH-8057 Z?rich, Switzerland --------------------------------------------------------------- -----cp... at googlegroups.com wrote: -----To: cp2k From: wei... at gmail.com Sent by: cp... at googlegroups.com Date: 07/14/2017 04:32AM Subject: [CP2K:9218] hybrid functional and symmetry with k-point Dear CP2K developers, May I ask if doing hybrid functional calculations with k-point has been implemented in the latest trunk of CP2K? There was a post in this forum in 2015 asking this question (https://groups.google.com/forum/#!topic/cp2k/2XxNUt8tML8) and Professor Hutter answer at that time it was not yet implemented. How about now? Also, according to this FAQ item https://www.cp2k.org/faq:kpoints, reduction of k-points using symmetry is not supported. Is there any development on that planned? Thanks,Wei -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. From yanghon... at gmail.com Fri Jul 14 12:24:06 2017 From: yanghon... at gmail.com (Hongliu) Date: Fri, 14 Jul 2017 05:24:06 -0700 (PDT) Subject: geometry optimization of Na slab In-Reply-To: <6dc2bc14-757a-482d-862f-dc310c60dc8b@googlegroups.com> References: <8bf66b27-96d8-4206-aedc-abbf3a0c4164@googlegroups.com> <41b6bde6-6748-4f01-8a06-ad2441840387@googlegroups.com> <9d3e0298-89c5-4300-a0bc-cc0250ab2ebb@googlegroups.com> <6dc2bc14-757a-482d-862f-dc310c60dc8b@googlegroups.com> Message-ID: <279e10de-64de-437c-8e3e-61af3bc11242@googlegroups.com> Dear Matt, your suggestion helps indeed! With GAPW the cell optimization of a 6x6x6 cell finished successfully with the bcc symmetry being kept quite well. The obtained lattice constant is about 4.20 Angstrom, close to that obtained by using other methods, for stance https://wiki.fysik.dtu.dk/gpaw/setups/bulk_tests.html. For the geometry optimization of Na-111-slabs the 3-fold symmetry is also kept quite well. Great! Thanks for the insightful hints. But some new issues come out now: For a 3-layer slab, the inter-layer distance changes from 1.2 Angstrom to 0.8 Angstrom. This seems fine, more or less. For a 4-layer slab, the two inner layers merge and the inter-layer distance of the resultant 3-layer slab becomes 1.9 Angstrom. Strange?! For a 6-layer slab, the inter-layer distances are 1.15, 1.0, 1.55, 1.0, 1.15, i.e. a wider gap appears between the upper 3-layer and the lower 3-layer. The bottom layer atoms are fixed. The input files and results for the 4-layer case are attached. Could you and other experts comment on that? Thanks! Best wishes, Hongliu -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: vmdscene.tga Type: application/octet-stream Size: 3817986 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: slab-111-4.mpg Type: video/mpeg Size: 63916 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: pf0-86-re.inp Type: chemical/x-gamess-input Size: 1612 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Na_111_663_slab.xyz Type: chemical/x-xyz Size: 4562 bytes Desc: not available URL: From anirudd... at gmail.com Fri Jul 14 18:20:33 2017 From: anirudd... at gmail.com (Aniruddha Dive) Date: Fri, 14 Jul 2017 11:20:33 -0700 (PDT) Subject: Compiling CP2K wih Intel 2016 Message-ID: <599af1a8-c6be-4068-b881-7ab27ece0c3a@googlegroups.com> Hi All, I have trying to compile CP2K with intel MKL 2016, LIBINIT 1.1.4, LIBXC 3.0, FFTW 3.3.4 and ELPA 2016.05.03 but I am getting the below errors: /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_terminate_libint_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:181: undefined reference to `free_libint' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_terminate_libderiv_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:191: undefined reference to `free_libderiv' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_initialize_libderiv_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:160: undefined reference to `init_libderiv_base' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:164: undefined reference to `libderiv1_storage_required' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:169: undefined reference to `init_libderiv1' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_initialize_libint_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:128: undefined reference to `init_libint_base' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:132: undefined reference to `libint_storage_required' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:136: undefined reference to `init_libint' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): In function `cp_fm_diag_mp_cp_fm_power_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: undefined reference to `elpa2_mp_solve_evp_real_2stage_' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): In function `cp_fm_diag_mp_cp_fm_elpa_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: undefined reference to `elpa2_mp_solve_evp_real_2stage_' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_terminate_libint_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:181: undefined reference to `free_libint' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_terminate_libderiv_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:191: undefined reference to `free_libderiv' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_initialize_libderiv_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:160: undefined reference to `init_libderiv_base' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:164: undefined reference to `libderiv1_storage_required' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:169: undefined reference to `init_libderiv1' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_initialize_libint_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:128: undefined reference to `init_libint_base' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:132: undefined reference to `libint_storage_required' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:136: undefined reference to `init_libint' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_terminate_libint_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:181: undefined reference to `free_libint' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_terminate_libderiv_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:191: undefined reference to `free_libderiv' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_initialize_libderiv_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:160: undefined reference to `init_libderiv_base' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:164: undefined reference to `libderiv1_storage_required' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:169: undefined reference to `init_libderiv1' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): In function `hfx_libint_wrapper_mp_initialize_libint_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:128: undefined reference to `init_libint_base' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:132: undefined reference to `libint_storage_required' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:136: undefined reference to `init_libint' make[3]: *** [/home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/exe/Linux-x86-64-intel-host/cp2k_shell.popt] Error 1 make[3]: *** Waiting for unfinished jobs.... /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): In function `cp_fm_diag_mp_cp_fm_power_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: undefined reference to `elpa2_mp_solve_evp_real_2stage_' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): In function `cp_fm_diag_mp_cp_fm_elpa_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: undefined reference to `elpa2_mp_solve_evp_real_2stage_' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): In function `cp_fm_diag_mp_cp_fm_power_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: undefined reference to `elpa2_mp_solve_evp_real_2stage_' /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): In function `cp_fm_diag_mp_cp_fm_elpa_': /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: undefined reference to `elpa2_mp_solve_evp_real_2stage_' make[3]: *** [/home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/exe/Linux-x86-64-intel-host/cp2k.popt] Error 1 make[3]: *** [/home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/exe/Linux-x86-64-intel-host/libcp2k_unittest.popt] Error 1 Here is my .popt arch file: CC = icc CPP = FC = mpiifort LD = mpiifort AR = ar -r LIBXC_DIR = /home/aniruddha.dive/CP2K/LIBXC/install LIBINT_DIR = /opt/apps/libint/1.1.4_2017-1-4 ELPA_LIB = /opt/apps/elpa/2016.05.003/lib ELPA_INC = /opt/apps/elpa/2016.05.003/include/elpa_openmp-2016.05.003 CPPFLAGS = DFLAGS = -D__MKL -D__parallel -D__SCALAPACK -D__FFTW3 -D__LIBXC -D__ELPA3 \ -D__LIBINT \ -I$(LIBXC_DIR)/include CFLAGS = $(DFLAGS) MKLROOT = /opt/apps/intel/xe_2016_update3/compilers_and_libraries_2016.3.210/linux/mkl FCFLAGS = $(DFLAGS) -O2 -g -traceback -fpp -free \ -I$(MKLROOT)/include -I$(MKLROOT)/include/fftw -I$(ELPA_INC)/modules -I$(ELPA_INC)/elpa FCFLAGS2 = $(DFLAGS) -O0 -g -traceback -fpp -free \ -I$(MKLROOT)/include -I$(MKLROOT)/include/fftw -I$(ELPA_INC)/modules -I$(ELPA_INC)/elpa LDFLAGS = $(FCFLAGS) -static-intel LDFLAGS_C = $(FCFLAGS) -static-intel -nofor_main LIBS += $(MKLROOT)/lib/intel64/libmkl_scalapack_lp64.a \ -Wl,--start-group $(MKLROOT)/lib/intel64/libmkl_intel_lp64.a \ $(MKLROOT)/lib/intel64/libmkl_sequential.a \ $(MKLROOT)/lib/intel64/libmkl_core.a \ $(MKLROOT)/lib/intel64/libmkl_blacs_intelmpi_lp64.a -Wl,--end-group \ -lpthread -lm \ -L$(LIBXC_DIR)/lib -lxcf90 -lxc \ -L$(LIBINT_DIR)/lib/libderiv.a -L$(LIBINT_DIR)/lib/libint.a -lstdc++ \ -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.la -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.a \ -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so.4 -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so.4.2.0 # Required due to memory leak that occurs if high optimisations are used mp2_optimize_ri_basis.o: mp2_optimize_ri_basis.F $(FC) -c $(FCFLAGS2) $< hfx_contraction_methods.o: hfx_contraction_methods.F $(FC) -c $(FCFLAGS2) $< Kindly let me know where am I going wrong. Regards, Aniruddha M Dive PhD Candidate Washington State University -------------- next part -------------- An HTML attachment was scrubbed... URL: From wei... at gmail.com Sat Jul 15 02:42:01 2017 From: wei... at gmail.com (wei... at gmail.com) Date: Fri, 14 Jul 2017 19:42:01 -0700 (PDT) Subject: [CP2K:9218] hybrid functional and symmetry with k-point In-Reply-To: References: <8d468847-237b-4a5b-a816-c8cf9935af0b@googlegroups.com> Message-ID: <76f087f2-7070-4997-b281-0254a0fb3875@googlegroups.com> Thank you for the kind reply, Professor Hutter. Wei On Friday, July 14, 2017 at 2:55:48 AM UTC-4, jgh wrote: > > Hi > > both of these options are still not implemented in CP2K. > Symmetry reduction of k-points might be available soon. > HFX for k-point is planned but nobody is working on it right now. > > regards > > Juerg > -------------------------------------------------------------- > Juerg Hutter Phone : ++41 44 635 4491 > Institut f?r Chemie C FAX : ++41 44 635 6838 > Universit?t Z?rich E-mail: hut... at chem.uzh.ch > > Winterthurerstrasse 190 > CH-8057 Z?rich, Switzerland > --------------------------------------------------------------- > > -----cp... at googlegroups.com wrote: -----To: cp2k < > cp... at googlegroups.com > > From: wei... at gmail.com > Sent by: cp... at googlegroups.com > Date: 07/14/2017 04:32AM > Subject: [CP2K:9218] hybrid functional and symmetry with k-point > > Dear CP2K developers, > May I ask if doing hybrid functional calculations with k-point has been > implemented in the latest trunk of CP2K? > There was a post in this forum in 2015 asking this question ( > https://groups.google.com/forum/#!topic/cp2k/2XxNUt8tML8) and Professor > Hutter answer at that time it was not yet implemented. How about now? > Also, according to this FAQ item https://www.cp2k.org/faq:kpoints, > reduction of k-points using symmetry is not supported. Is there any > development on that planned? > Thanks,Wei > > > > > -- > > You received this message because you are subscribed to the Google Groups > "cp2k" group. > > To unsubscribe from this group and stop receiving emails from it, send an > email to cp2k+... at googlegroups.com . > > To post to this group, send email to cp... at googlegroups.com . > > > Visit this group at https://groups.google.com/group/cp2k. > > For more options, visit https://groups.google.com/d/optout. > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From grsm... at gmail.com Mon Jul 17 03:17:51 2017 From: grsm... at gmail.com (Grigory Smirnov) Date: Sun, 16 Jul 2017 20:17:51 -0700 (PDT) Subject: Wrong results during regtesting Message-ID: Dear developers, Recently I have noticed on dashboard that CP2K failed during regtesting on PizDaint. Actually, I've got the same problem on the different cluster with the different compilers (Intel MPI Build 20161016, GNU compilers 6.2, OpenBLAS 0.2.19, SCALAPACK 2.0.2, FFTW 3.3.6, Libxsmm 1.8.1). My version works fine with mpiranks=1 or 2, but produces wrong results with mpiranks=4 or 8. The wrong tests are the same as on PizDaint. Is it a bug in CP2K or somewhere else? Best regards, Grigory. -------------- next part -------------- An HTML attachment was scrubbed... URL: From tiziano... at chem.uzh.ch Mon Jul 17 08:33:52 2017 From: tiziano... at chem.uzh.ch (=?UTF-8?Q?Tiziano_M=c3=bcller?=) Date: Mon, 17 Jul 2017 10:33:52 +0200 Subject: [CP2K:9222] Compiling CP2K wih Intel 2016 In-Reply-To: <599af1a8-c6be-4068-b881-7ab27ece0c3a@googlegroups.com> References: <599af1a8-c6be-4068-b881-7ab27ece0c3a@googlegroups.com> Message-ID: <868ff808-be5d-81ce-9122-615fde1e14ab@chem.uzh.ch> Hi Aniruddha My guess is that the following linker/LIBS flags are invalid: -L$(LIBINT_DIR)/lib/libderiv.a -L$(LIBINT_DIR)/lib/libint.a -lstdc++ \ -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.la -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.a \ -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so.4 -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so.4.2.0 which should be something like instead: -L$(LIBINT_DIR)/lib -lderiv -lint -lstdc++ \ -L/opt/apps/elpa/2016.05.003/lib -lelpa_openmp Please read `man ld` for more information about the difference between -L and -l Best regards, Tiziano On 14.07.2017 20:20, Aniruddha Dive wrote: > Hi All, > > I have trying to compile CP2K with intel MKL 2016, LIBINIT 1.1.4, LIBXC > 3.0, FFTW 3.3.4 and ELPA 2016.05.03 but I am getting the below errors: > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_terminate_libint_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:181: > undefined reference to `free_libint' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_terminate_libderiv_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:191: > undefined reference to `free_libderiv' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_initialize_libderiv_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:160: > undefined reference to `init_libderiv_base' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:164: > undefined reference to `libderiv1_storage_required' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:169: > undefined reference to `init_libderiv1' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_initialize_libint_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:128: > undefined reference to `init_libint_base' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:132: > undefined reference to `libint_storage_required' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:136: > undefined reference to `init_libint' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): > In function `cp_fm_diag_mp_cp_fm_power_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: > undefined reference to `elpa2_mp_solve_evp_real_2stage_' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): > In function `cp_fm_diag_mp_cp_fm_elpa_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: > undefined reference to `elpa2_mp_solve_evp_real_2stage_' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_terminate_libint_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:181: > undefined reference to `free_libint' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_terminate_libderiv_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:191: > undefined reference to `free_libderiv' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_initialize_libderiv_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:160: > undefined reference to `init_libderiv_base' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:164: > undefined reference to `libderiv1_storage_required' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:169: > undefined reference to `init_libderiv1' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_initialize_libint_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:128: > undefined reference to `init_libint_base' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:132: > undefined reference to `libint_storage_required' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:136: > undefined reference to `init_libint' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_terminate_libint_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:181: > undefined reference to `free_libint' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_terminate_libderiv_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:191: > undefined reference to `free_libderiv' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_initialize_libderiv_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:160: > undefined reference to `init_libderiv_base' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:164: > undefined reference to `libderiv1_storage_required' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:169: > undefined reference to `init_libderiv1' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kmain.a(hfx_libint_wrapper.o): > In function `hfx_libint_wrapper_mp_initialize_libint_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:128: > undefined reference to `init_libint_base' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:132: > undefined reference to `libint_storage_required' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/hfx_libint_wrapper.F:136: > undefined reference to `init_libint' > make[3]: *** > [/home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/exe/Linux-x86-64-intel-host/cp2k_shell.popt] > Error 1 > make[3]: *** Waiting for unfinished jobs.... > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): > In function `cp_fm_diag_mp_cp_fm_power_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: > undefined reference to `elpa2_mp_solve_evp_real_2stage_' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): > In function `cp_fm_diag_mp_cp_fm_elpa_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: > undefined reference to `elpa2_mp_solve_evp_real_2stage_' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): > In function `cp_fm_diag_mp_cp_fm_power_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: > undefined reference to `elpa2_mp_solve_evp_real_2stage_' > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/lib/Linux-x86-64-intel-host/popt/libcp2kfm.a(cp_fm_diag.o): > In function `cp_fm_diag_mp_cp_fm_elpa_': > /home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/src/fm/cp_fm_diag.F:971: > undefined reference to `elpa2_mp_solve_evp_real_2stage_' > make[3]: *** > [/home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/exe/Linux-x86-64-intel-host/cp2k.popt] > Error 1 > make[3]: *** > [/home/aniruddha.dive/CP2K/CP2K_INTEL/cp2k-4.1/exe/Linux-x86-64-intel-host/libcp2k_unittest.popt] > Error 1 > > Here is my .popt arch file: > CC = icc > CPP = > FC = mpiifort > LD = mpiifort > AR = ar -r > > LIBXC_DIR = /home/aniruddha.dive/CP2K/LIBXC/install > LIBINT_DIR = /opt/apps/libint/1.1.4_2017-1-4 > ELPA_LIB = /opt/apps/elpa/2016.05.003/lib > ELPA_INC = /opt/apps/elpa/2016.05.003/include/elpa_openmp-2016.05.003 > > CPPFLAGS = > DFLAGS = -D__MKL -D__parallel -D__SCALAPACK -D__FFTW3 -D__LIBXC > -D__ELPA3 \ > -D__LIBINT \ > -I$(LIBXC_DIR)/include > CFLAGS = $(DFLAGS) > MKLROOT = > /opt/apps/intel/xe_2016_update3/compilers_and_libraries_2016.3.210/linux/mkl > FCFLAGS = $(DFLAGS) -O2 -g -traceback -fpp -free \ > -I$(MKLROOT)/include -I$(MKLROOT)/include/fftw > -I$(ELPA_INC)/modules -I$(ELPA_INC)/elpa > FCFLAGS2 = $(DFLAGS) -O0 -g -traceback -fpp -free \ > -I$(MKLROOT)/include -I$(MKLROOT)/include/fftw > -I$(ELPA_INC)/modules -I$(ELPA_INC)/elpa > LDFLAGS = $(FCFLAGS) -static-intel > LDFLAGS_C = $(FCFLAGS) -static-intel -nofor_main > LIBS += $(MKLROOT)/lib/intel64/libmkl_scalapack_lp64.a \ > -Wl,--start-group $(MKLROOT)/lib/intel64/libmkl_intel_lp64.a \ > $(MKLROOT)/lib/intel64/libmkl_sequential.a \ > $(MKLROOT)/lib/intel64/libmkl_core.a \ > $(MKLROOT)/lib/intel64/libmkl_blacs_intelmpi_lp64.a > -Wl,--end-group \ > -lpthread -lm \ > -L$(LIBXC_DIR)/lib -lxcf90 -lxc \ > -L$(LIBINT_DIR)/lib/libderiv.a -L$(LIBINT_DIR)/lib/libint.a > -lstdc++ \ > -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.la > -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.a \ > -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so > -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so.4 > -L/opt/apps/elpa/2016.05.003/lib/libelpa_openmp.so.4.2.0 > > # Required due to memory leak that occurs if high optimisations are used > mp2_optimize_ri_basis.o: mp2_optimize_ri_basis.F > $(FC) -c $(FCFLAGS2) $< > > hfx_contraction_methods.o: hfx_contraction_methods.F > $(FC) -c $(FCFLAGS2) $< > > Kindly let me know where am I going wrong. > > Regards, > Aniruddha M Dive > PhD Candidate > Washington State University > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com > . > To post to this group, send email to cp... at googlegroups.com > . > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. -- Tiziano M?ller University of Zurich Department of Chemistry Winterthurerstrasse 190 CH-8057 Z?rich Tel: +41 44 63 54234 www.chem.uzh.ch tiziano... at chem.uzh.ch From ganta.pra... at gmail.com Mon Jul 17 16:20:23 2017 From: ganta.pra... at gmail.com (ganta.pra... at gmail.com) Date: Mon, 17 Jul 2017 09:20:23 -0700 (PDT) Subject: Distorted optimized structure (FeOOH) while GEO_OPT. Message-ID: I am doing geometry optimization of a slab (FeOOH), with 30 Angstroms vacuum in Z direction. I have done bulk optimization successfully and I am using optimized bulk structure to do slab optimization (with 30 Angstroms vacuum). The slab is not optimizing and it is heavily distorted towards vacuum region. Please let me know any ideas to avoid this distortion. I am planning to use more layers (to see if that avoids distortion), but increasing more layers increases more atoms and it is computationally expensive. The distortion can be seen in the attached surf001-pos-1.xyz file. I tried to see old posts about this and the structures used were comparatively small (my model 382 atoms). Please let me know if you have any suggestions. Thanks and Regards, Prasanth. -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: surf001-pos-1.xyz Type: chemical/x-xyz Size: 4951782 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: input_slab_001.inp Type: chemical/x-gamess-input Size: 2200 bytes Desc: not available URL: From mattwa... at gmail.com Mon Jul 17 16:59:34 2017 From: mattwa... at gmail.com (Matt W) Date: Mon, 17 Jul 2017 09:59:34 -0700 (PDT) Subject: Distorted optimized structure (FeOOH) while GEO_OPT. In-Reply-To: References: Message-ID: <504e24b3-6b92-4a54-b39d-de386f729969@googlegroups.com> Hi, have you just cut a bulk slab to generate your surface? Please check the literature - do you expect this to be a stable structure. It looks like the optimization is trying to do sensible things, but the start point is highly unstable (polar?). Matt On Monday, July 17, 2017 at 5:20:23 PM UTC+1, ganta.... at gmail.com wrote: > > I am doing geometry optimization of a slab (FeOOH), with 30 Angstroms > vacuum in Z direction. I have done bulk optimization successfully and I am > using optimized bulk structure to do slab optimization (with 30 Angstroms > vacuum). The slab is not optimizing and it is heavily distorted towards > vacuum region. Please let me know any ideas to avoid this distortion. > > I am planning to use more layers (to see if that avoids distortion), but > increasing more layers increases more atoms and it is computationally > expensive. > > The distortion can be seen in the attached surf001-pos-1.xyz file. I > tried to see old posts about this and the structures used were > comparatively small (my model 382 atoms). > > Please let me know if you have any suggestions. > > Thanks and Regards, > Prasanth. > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ganta.pra... at gmail.com Mon Jul 17 17:31:02 2017 From: ganta.pra... at gmail.com (ganta.pra... at gmail.com) Date: Mon, 17 Jul 2017 10:31:02 -0700 (PDT) Subject: Distorted optimized structure (FeOOH) while GEO_OPT. In-Reply-To: <504e24b3-6b92-4a54-b39d-de386f729969@googlegroups.com> References: <504e24b3-6b92-4a54-b39d-de386f729969@googlegroups.com> Message-ID: <4f436d69-ebd2-4881-837f-97d32b86c818@googlegroups.com> Dear Matt W, It is complete structure, I did not cut bulk to generate my surface. My initial point is the optimized bulk structure. I have also built structure from the FeOOH unit cell and I have done slab simulation. The same distortion is happening. I dont think that this is due to symmetry or polarity (I am not 100% sure :) ). I am new to CP2K and QM and if you suggest some literature on how to check the stability of the structure before proceeding with simulations, that will be great for me. Thank you for the reply. If you have any other suggestions, do let me know. Regards, Prasanth. On Monday, July 17, 2017 at 6:59:35 PM UTC+2, Matt W wrote: > > Hi, > > have you just cut a bulk slab to generate your surface? > > Please check the literature - do you expect this to be a stable structure. > It looks like the optimization is trying to do sensible things, but the > start point is highly unstable (polar?). > > Matt > > On Monday, July 17, 2017 at 5:20:23 PM UTC+1, ganta.... at gmail.com > wrote: >> >> I am doing geometry optimization of a slab (FeOOH), with 30 Angstroms >> vacuum in Z direction. I have done bulk optimization successfully and I am >> using optimized bulk structure to do slab optimization (with 30 Angstroms >> vacuum). The slab is not optimizing and it is heavily distorted towards >> vacuum region. Please let me know any ideas to avoid this distortion. >> >> I am planning to use more layers (to see if that avoids distortion), but >> increasing more layers increases more atoms and it is computationally >> expensive. >> >> The distortion can be seen in the attached surf001-pos-1.xyz file. I >> tried to see old posts about this and the structures used were >> comparatively small (my model 382 atoms). >> >> Please let me know if you have any suggestions. >> >> Thanks and Regards, >> Prasanth. >> >> >> -------------- next part -------------- An HTML attachment was scrubbed... URL: From chot... at gmail.com Tue Jul 18 06:17:50 2017 From: chot... at gmail.com (Cho S) Date: Mon, 17 Jul 2017 23:17:50 -0700 (PDT) Subject: Bug Report: COMBINE_COLVAR Message-ID: <35aad8c9-3a7f-4acd-b865-834c6289f03b@googlegroups.com> Hello, I have noticed a bug where COMBINE_COLVAR incorrectly reports COLVAR values. For example, one would expect two COLVARs below to be the same: &COLVAR &COORDINATION ATOMS_FROM 1 KINDS_TO H NN 9 ND 14 R0 [angstrom] 1.5 &END COORDINATION &END COLVAR and &COLVAR &COMBINE_COLVAR &COLVAR &COORDINATION ATOMS_FROM 1 KINDS_TO H NN 9 ND 14 R0 [angstrom] 1.5 &END COORDINATION &END COLVAR FUNCTION CV_combined VARIABLES CV_combined &END COMBINE_COLVAR &END COLVAR However, metadynLog reports correct CV1, while CV2 (combined CV) is listed as 0. I think this bug was reported earlier at https://groups.google.com/forum/#!searchin/cp2k/COMBINE_COLVAR%7Csort:relevance/cp2k/_uFtNhs8ptY/la3vSdiX7I4J but apparently it is not fixed yet. I would like to confirm that this bug is (at least) specific to COMBINE_COLVAR, where COORDINATION is used together with ATOMS_FROM and KINDS_TO. I have attached an input file that would reproduce this result. It should take less than a minute to run on a single core. -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: cp2k.inp Type: chemical/x-gamess-input Size: 2210 bytes Desc: not available URL: From o... at schuett.name Tue Jul 18 11:14:16 2017 From: o... at schuett.name (=?UTF-8?Q?Ole_Sch=C3=BCtt?=) Date: Tue, 18 Jul 2017 13:14:16 +0200 Subject: [CP2K:9230] CP2K tutorial for biochemical systems In-Reply-To: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> Message-ID: <2ad368658595f9b8e5b266bd6c91b0e1@schuett.name> Hi Dries, your tutorials are indeed very nice and they fill an important gap our documentation concerning QM/MM. Usually, we collect tutorials centrally on our wiki: https://www.cp2k.org/howto . Could I interest you in moving them there? The conversion from Markdown to DokuWiki should be straight forward. In any case, thanks a lot for the work you already put in! Cheers, Ole On 2017-07-18 08:49, Dries Van Rompaey wrote: > Hi cp2k-users, > > After experimenting with cp2k for a while I have written down a short > tutorial on how to use cp2k for biochemical systems. The tutorial > covers the setup of an enzyme system with ambertools followed by > equilibration with cp2k at the MM level, after which we move the > system to QM/MM for metadynamics simulations of an enzymatic reaction. > I am far from an expert myself, but I thought this might be useful to > people starting out. If you spot any mistakes or something that could > be improved, please let me know. The tutorial can be found at > https://driesvr.github.io/Tutorials/. > > Kind regards > > Dries > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com. > To post to this group, send email to cp... at googlegroups.com. > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. From dries.va... at gmail.com Tue Jul 18 11:40:46 2017 From: dries.va... at gmail.com (Dries Van Rompaey) Date: Tue, 18 Jul 2017 04:40:46 -0700 (PDT) Subject: [CP2K:9230] CP2K tutorial for biochemical systems In-Reply-To: <2ad368658595f9b8e5b266bd6c91b0e1@schuett.name> References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> <2ad368658595f9b8e5b266bd6c91b0e1@schuett.name> Message-ID: <800e4c60-33bb-41ef-a5bd-a2212084a2e3@googlegroups.com> Hi Ole, Sure, I'll put them on the wiki as well. Thanks for the comments! Dries Op dinsdag 18 juli 2017 13:14:20 UTC+2 schreef Ole Sch?tt: > > Hi Dries, > > your tutorials are indeed very nice and they fill an important gap our > documentation concerning QM/MM. Usually, we collect tutorials centrally > on our wiki: https://www.cp2k.org/howto . Could I interest you in moving > them there? The conversion from Markdown to DokuWiki should be straight > forward. > > In any case, thanks a lot for the work you already put in! > > Cheers, > Ole > > On 2017-07-18 08:49, Dries Van Rompaey wrote: > > Hi cp2k-users, > > > > After experimenting with cp2k for a while I have written down a short > > tutorial on how to use cp2k for biochemical systems. The tutorial > > covers the setup of an enzyme system with ambertools followed by > > equilibration with cp2k at the MM level, after which we move the > > system to QM/MM for metadynamics simulations of an enzymatic reaction. > > I am far from an expert myself, but I thought this might be useful to > > people starting out. If you spot any mistakes or something that could > > be improved, please let me know. The tutorial can be found at > > https://driesvr.github.io/Tutorials/. > > > > Kind regards > > > > Dries > > > > -- > > You received this message because you are subscribed to the Google > > Groups "cp2k" group. > > To unsubscribe from this group and stop receiving emails from it, send > > an email to cp2k+... at googlegroups.com . > > To post to this group, send email to cp... at googlegroups.com > . > > Visit this group at https://groups.google.com/group/cp2k. > > For more options, visit https://groups.google.com/d/optout. > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From adebe... at gmail.com Tue Jul 18 20:38:57 2017 From: adebe... at gmail.com (Anthony Debellis) Date: Tue, 18 Jul 2017 13:38:57 -0700 (PDT) Subject: Confusion about &BS Message-ID: <83da531c-f97b-4ec8-8103-26969b4de98a@googlegroups.com> Dear cp2k users, After reading several posts on this board I have included in my input file the following: &KIND Cu BASIS_SET DZVP-MOLOPT-SR-GTH POTENTIAL GTH-PBE-q11 &BS &ALPHA NEL -1 0 L 0 2 N 4 3 &END &BETA NEL -1 -2 L 0 2 N 4 3 &END &END BS &END KIND In this calculation I want to compare a triplet to a singlet biradical state, so there is another electron on another molecule (nitric oxide) that can be singlet or triplet coupled to the Cu atom. The total charge on the system is zero. I believe this to set the configuration of Cu+2 to 4s-0 3d-9 with 1 extra alpha electron in the d level. Cu is now +2 with 27 electrons. Is this incorrect? What confuses me is the following section in the output file that prints the guess for Cu: ========================================================================= Guess for atomic kind: Cu Electronic structure Total number of core electrons 18.00 Total number of valence electrons 10.00 Total number of electrons 28.00 Multiplicity not specified S [ 2.00 2.00 2.00] P [ 6.00 6.00] D 10.00 ******************************************************************************* Iteration Convergence Energy [au] ******************************************************************************* 1 1.45663 -46.971858672390 2 8.20012 -45.115629076527 3 0.959876 -47.359129413493 4 0.415414 -47.404183185274 5 0.240868 -47.411929209168 6 0.163028 -47.414155982591 7 0.103505 -47.415261301259 8 0.488716E-02 -47.416106927225 9 0.956072E-03 -47.416108693194 10 0.932233E-03 -47.416108696648 11 0.603325E-03 -47.416108735409 12 0.466873E-05 -47.416108763346 13 0.298676E-06 -47.416108763348 Energy components [Hartree] Total Energy :: -47.416108763348 Band Energy :: -4.785483365049 Kinetic Energy :: 71.754778900740 Potential Energy :: -119.170887664088 Virial (-V/T) :: 1.660807677060 Core Energy :: -85.003591522950 XC Energy :: -7.357859065342 Coulomb Energy :: 44.945341824943 Total Pseudopotential Energy :: -156.812157479736 Local Pseudopotential Energy :: -107.094489614039 Nonlocal Pseudopotential Energy :: -49.717667865697 Confinement :: 0.537870560470 Orbital energies State L Occupation Energy[a.u.] Energy[eV] 1 2 10.000 -0.478548 -13.021962 ======================================================================= Why does Cu have a total of 28 electrons with 10 in the d level? What am I missing? Anthony -------------- next part -------------- An HTML attachment was scrubbed... URL: From feliped... at gmail.com Wed Jul 19 09:04:16 2017 From: feliped... at gmail.com (fquesada) Date: Wed, 19 Jul 2017 02:04:16 -0700 (PDT) Subject: =?UTF-8?Q?Error_with_=E2=80=9CADDED=5FMOS=E2=80=9D_keyword_?= =?UTF-8?Q?for_Mg_Energy-Force_Calculation?= Message-ID: Dear cp2k users, I am relatively new to the program and I am interested in performing static DFT energy and force calculations for Magnesium. The code I wrote to perform the calculation runs smoothly in general until I add a ?SMEAR? section and ?ADDED_MOS?, which then yields the following error: ________________________________________________________________ *** 16:57:26 WARNING in qs_environment:qs_init_subsys :: More added MOs *** *** requested than available. The full set of unoccupied MOs will be used. *** ******************************************************* *** ERROR in qs_init_subsys (MODULE qs_environment) *** ******************************************************* *** Extra MOs (ADDED_MOS) are required for smearing *** *** Program stopped at line number 1190 of MODULE qs_environment *** ===== Routine Calling Stack ===== 3 qs_init_subsys 2 quickstep_create_force_env 1 CP2K ________________________________________________________________ Despite trying to increase/decrease the number of added molecular orbitals for ?ADDED_MOS?, I still get the error and the simulation can?t run. I would very much appreciate somebody?s help in trying to understand why the code is not working properly and how I could improve it. I attached my code below this message for reference. Thank you in advance for your comments, Cheers, Felipe ________________________________________________________________ &GLOBAL PROJECT magnesium_10 RUN_TYPE ENERGY_FORCE PRINT_LEVEL LOW &END GLOBAL &FORCE_EVAL METHOD Quickstep &SUBSYS &KIND Mg ELEMENT Mg BASIS_SET SZV-GTH POTENTIAL GTH-PBE-q10 &END KIND &CELL A 3.209400000 0.000000000 0.000000000 B 0.000000000 3.209400000 0.000000000 C 0.000000000 0.000000000 5.210800000 ALPHA_BETA_GAMMA 90.0000000 90.0000000 120.0000000 &END CELL &COORD Mg 0.000000000 0.000000000 0.000000000 Mg 0.333333333 0.666666666 0.500000000 &END COORD &END SUBSYS &DFT BASIS_SET_FILE_NAME GTH_BASIS_SETS POTENTIAL_FILE_NAME GTH_POTENTIALS &QS EPS_DEFAULT 1.0E-10 &END QS &MGRID NGRIDS 4 CUTOFF 300 REL_CUTOFF 60 &END MGRID &XC &XC_FUNCTIONAL PBE &END XC_FUNCTIONAL &END XC &SCF SCF_GUESS ATOMIC EPS_SCF 1.0E-7 MAX_SCF 300 ADDED_MOS 10 &SMEAR ON METHOD FERMI_DIRAC ELECTRONIC_TEMPERATURE [K] 300 &END SMEAR &DIAGONALIZATION ON ALGORITHM STANDARD &END DIAGONALIZATION &MIXING T METHOD BROYDEN_MIXING ALPHA 0.4 NBROYDEN 8 &END MIXING &END SCF &END DFT &PRINT &FORCES ON &END FORCES &END PRINT &END FORCE_EVAL -------------- next part -------------- An HTML attachment was scrubbed... URL: From mattwa... at gmail.com Wed Jul 19 09:14:52 2017 From: mattwa... at gmail.com (Matt W) Date: Wed, 19 Jul 2017 02:14:52 -0700 (PDT) Subject: =?UTF-8?Q?Re:_Error_with_=E2=80=9CADDED=5FMOS=E2=80=9D_keywor?= =?UTF-8?Q?d_for_Mg_Energy-Force_Calculation?= In-Reply-To: References: Message-ID: Hi, with your minimal basis set you do not have any extra orbitals to add. You need a bigger basis set (and unit cell, or k-points). Matt On Wednesday, July 19, 2017 at 10:04:16 AM UTC+1, fquesada wrote: > > Dear cp2k users, > > I am relatively new to the program and I am interested in performing > static DFT energy and force calculations for Magnesium. The code I wrote to > perform the calculation runs smoothly in general until I add a ?SMEAR? > section and ?ADDED_MOS?, which then yields the following error: > > ________________________________________________________________ > > > > *** 16:57:26 WARNING in qs_environment:qs_init_subsys :: More added MOs > *** > > *** requested than available. The full set of unoccupied MOs will be > used. *** > > > > > > > > ******************************************************* > > *** ERROR in qs_init_subsys (MODULE qs_environment) *** > > ******************************************************* > > > > *** Extra MOs (ADDED_MOS) are required for smearing *** > > > > *** Program stopped at line number 1190 of MODULE qs_environment *** > > > > ===== Routine Calling Stack ===== > > > > 3 qs_init_subsys > > 2 quickstep_create_force_env > > 1 CP2K > > ________________________________________________________________ > > > > Despite trying to increase/decrease the number of added molecular orbitals > for ?ADDED_MOS?, I still get the error and the simulation can?t run. I > would very much appreciate somebody?s help in trying to understand why the > code is not working properly and how I could improve it. I attached my code > below this message for reference. > > Thank you in advance for your comments, > > Cheers, > > Felipe > > > > ________________________________________________________________ > > &GLOBAL > > PROJECT magnesium_10 > > RUN_TYPE ENERGY_FORCE > > PRINT_LEVEL LOW > > &END GLOBAL > > &FORCE_EVAL > > METHOD Quickstep > > &SUBSYS > > &KIND Mg > > ELEMENT Mg > > BASIS_SET SZV-GTH > > POTENTIAL GTH-PBE-q10 > > &END KIND > > &CELL > > A 3.209400000 0.000000000 0.000000000 > > B 0.000000000 3.209400000 0.000000000 > > C 0.000000000 0.000000000 5.210800000 > > ALPHA_BETA_GAMMA 90.0000000 90.0000000 120.0000000 > > &END CELL > > &COORD > > Mg 0.000000000 0.000000000 0.000000000 > > Mg 0.333333333 0.666666666 0.500000000 > > &END COORD > > &END SUBSYS > > &DFT > > BASIS_SET_FILE_NAME GTH_BASIS_SETS > > POTENTIAL_FILE_NAME GTH_POTENTIALS > > &QS > > EPS_DEFAULT 1.0E-10 > > &END QS > > &MGRID > > NGRIDS 4 > > CUTOFF 300 > > REL_CUTOFF 60 > > &END MGRID > > &XC > > &XC_FUNCTIONAL PBE > > &END XC_FUNCTIONAL > > &END XC > > &SCF > > SCF_GUESS ATOMIC > > EPS_SCF 1.0E-7 > > MAX_SCF 300 > > ADDED_MOS 10 > > &SMEAR ON > > METHOD FERMI_DIRAC > > ELECTRONIC_TEMPERATURE [K] 300 > > &END SMEAR > > &DIAGONALIZATION ON > > ALGORITHM STANDARD > > &END DIAGONALIZATION > > &MIXING T > > METHOD BROYDEN_MIXING > > ALPHA 0.4 > > NBROYDEN 8 > > &END MIXING > > &END SCF > > &END DFT > > &PRINT > > &FORCES ON > > &END FORCES > > &END PRINT > > &END FORCE_EVAL > -------------- next part -------------- An HTML attachment was scrubbed... URL: From feliped... at gmail.com Wed Jul 19 09:24:50 2017 From: feliped... at gmail.com (fquesada) Date: Wed, 19 Jul 2017 02:24:50 -0700 (PDT) Subject: =?UTF-8?Q?Re:_Error_with_=E2=80=9CADDED=5FMOS=E2=80=9D_keywor?= =?UTF-8?Q?d_for_Mg_Energy-Force_Calculation?= In-Reply-To: References: Message-ID: <2e954aca-d621-401a-ba13-2d87a98f11c6@googlegroups.com> Dear Matt, thank you for your response. I included the "KPOINTS" section under "DFT" in my code: ______________________________________________________________________________ &KPOINTS SCHEME MONKHORST-PACK 12 12 8 &END KPOINTS ______________________________________________________________________________ It seems like the problem is still persisting. Do you have any further suggestions how to work around this issue? Best, Felipe On Wednesday, July 19, 2017 at 10:14:52 AM UTC+1, Matt W wrote: > > Hi, > > with your minimal basis set you do not have any extra orbitals to add. You > need a bigger basis set (and unit cell, or k-points). > > Matt > -------------- next part -------------- An HTML attachment was scrubbed... URL: From 2922... at qq.com Wed Jul 19 10:20:54 2017 From: 2922... at qq.com (Tonys_Yu) Date: Wed, 19 Jul 2017 03:20:54 -0700 (PDT) Subject: Confusion: cation and anion move towards the same direction under an external electric field Message-ID: <9b168bad-d662-4fd5-a156-38cfb58d1bf5@googlegroups.com> Hello, I want to study a ionic displacement in a bulk water box with 30 water molecules. External electric field is 0.1V/A, and added by EXTERNAL_POTENTIAL command as below: &EXTERNAL_POTENTIAL FUNCTION (A/B)*X PARAMETERS A B VALUES [eV] 1.0 [angstrom] 10.0 &END EXTERNAL_POTENTIAL What makes me confused is that both water molecules and K-ion are moving along the negative direction of X. Moreover, when I changed the K-ion by a Cl-ion or reduced the electric field to a smaller value (0.001V/A), the same results was observed. It seems contrary to the fact that, with the same electric field (value and direction), K-ion and Cl-ion would move towards opposite directions. I beg someone could shed light on my confusion. Necessary files have been attached. Tonys Yu -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: BASIS_MOLOPT Type: application/octet-stream Size: 126074 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: dftd3.dat Type: application/octet-stream Size: 1959318 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: GTH_POTENTIALS Type: application/octet-stream Size: 109253 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: K_Water.inp Type: chemical/x-gamess-input Size: 2459 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: wholemodel-modyfied.xyz Type: chemical/x-xyz Size: 8511 bytes Desc: not available URL: From marci... at gmail.com Wed Jul 19 10:25:10 2017 From: marci... at gmail.com (Marcella Iannuzzi) Date: Wed, 19 Jul 2017 03:25:10 -0700 (PDT) Subject: =?UTF-8?Q?Re:_Error_with_=E2=80=9CADDED=5FMOS=E2=80=9D_keywor?= =?UTF-8?Q?d_for_Mg_Energy-Force_Calculation?= In-Reply-To: <2e954aca-d621-401a-ba13-2d87a98f11c6@googlegroups.com> References: <2e954aca-d621-401a-ba13-2d87a98f11c6@googlegroups.com> Message-ID: <054c2dd6-948e-4bba-a42d-679d1ee7fc26@googlegroups.com> Hi, using k-points is for sure necessary with such a small CELL, however the problem with the BASIS SET is not solved. As Matt pointed out, you need a BASIS SET with more functions, a SZV-GTH is not large enough. Regards Marcella On Wednesday, July 19, 2017 at 11:24:50 AM UTC+2, fquesada wrote: > > Dear Matt, > thank you for your response. I included the "KPOINTS" section under "DFT" > in my code: > > ______________________________________________________________________________ > &KPOINTS > SCHEME MONKHORST-PACK 12 12 8 > &END KPOINTS > > ______________________________________________________________________________ > > It seems like the problem is still persisting. Do you have any further > suggestions how to work around this issue? > Best, > Felipe > > On Wednesday, July 19, 2017 at 10:14:52 AM UTC+1, Matt W wrote: >> >> Hi, >> >> with your minimal basis set you do not have any extra orbitals to add. >> You need a bigger basis set (and unit cell, or k-points). >> >> Matt >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From marci... at gmail.com Wed Jul 19 10:32:48 2017 From: marci... at gmail.com (Marcella Iannuzzi) Date: Wed, 19 Jul 2017 03:32:48 -0700 (PDT) Subject: Confusion: cation and anion move towards the same direction under an external electric field In-Reply-To: <9b168bad-d662-4fd5-a156-38cfb58d1bf5@googlegroups.com> References: <9b168bad-d662-4fd5-a156-38cfb58d1bf5@googlegroups.com> Message-ID: <17b8b336-c676-4cd5-8ea7-21e10a57b69d@googlegroups.com> Hi, this external potential is not an electric field, it is just a linear potential. Check the keywords EFIELD and PERIODIC_EFIELD. Marcella On Wednesday, July 19, 2017 at 12:20:54 PM UTC+2, Tonys_Yu wrote: > > Hello, > I want to study a ionic displacement in a bulk water box with 30 > water molecules. External electric field is 0.1V/A, and added by > EXTERNAL_POTENTIAL command as below: > > &EXTERNAL_POTENTIAL > FUNCTION (A/B)*X > PARAMETERS A B > VALUES [eV] 1.0 [angstrom] 10.0 > &END EXTERNAL_POTENTIAL > > What makes me confused is that both water molecules and K-ion are > moving along the negative direction of X. Moreover, when I changed the > K-ion by a Cl-ion or reduced the electric field to a smaller value > (0.001V/A), the same results was observed. It seems contrary to the fact > that, with the same electric field (value and direction), K-ion and Cl-ion > would move towards opposite directions. > I beg someone could shed light on my confusion. Necessary files have > been attached. > > Tonys Yu > -------------- next part -------------- An HTML attachment was scrubbed... URL: From feliped... at gmail.com Wed Jul 19 10:39:07 2017 From: feliped... at gmail.com (fquesada) Date: Wed, 19 Jul 2017 03:39:07 -0700 (PDT) Subject: =?UTF-8?Q?Re:_Error_with_=E2=80=9CADDED=5FMOS=E2=80=9D_keywor?= =?UTF-8?Q?d_for_Mg_Energy-Force_Calculation?= In-Reply-To: <054c2dd6-948e-4bba-a42d-679d1ee7fc26@googlegroups.com> References: <2e954aca-d621-401a-ba13-2d87a98f11c6@googlegroups.com> <054c2dd6-948e-4bba-a42d-679d1ee7fc26@googlegroups.com> Message-ID: <60de8b48-2036-4aca-9157-d55cb916c728@googlegroups.com> Dear Marcella, thank you for the advice! I tried running the calculation now with a larger basis set and the problem was resolved. However, I got another error associated with the subsection "KPOINTS". I copy below the error message that I get when trying to perform the calculation: ______________________________________________________________________________________ ************************************************************************* *** 11:33:19 ERRORL2 in kpoint_transitional processor 0 :: err=-1000 *** *** UNIMPLEMENTED, Method not implemented for k-points *** ************************************************************************* ===== Routine Calling Stack ===== 6 qs_rho_update_rho 5 scf_env_initial_rho_setup 4 init_scf_run 3 qs_energies_scf 2 qs_forces 1 CP2K ______________________________________________________________________________________ I would very much appreciate your help on this issue (and sorry for the many questions). Best, Felipe On Wednesday, July 19, 2017 at 11:25:10 AM UTC+1, Marcella Iannuzzi wrote: > > Hi, > > using k-points is for sure necessary with such a small CELL, however the > problem with the BASIS SET is not solved. > As Matt pointed out, you need a BASIS SET with more functions, a SZV-GTH > is not large enough. > Regards > Marcella > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jts2t... at gmail.com Wed Jul 19 11:34:52 2017 From: jts2t... at gmail.com (jts2t... at gmail.com) Date: Wed, 19 Jul 2017 04:34:52 -0700 (PDT) Subject: CP2K tutorial for biochemical systems In-Reply-To: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> Message-ID: <2f53575f-67c5-4b47-b448-32baf622bf7a@googlegroups.com> Hi Dries, I have read the tutorial and wonder how could I show the advantage of QMMM calculation over MM calculation. Should I run a complete MM simulation and compare the two results ? I notice that you said in the tutorial that the reaction is not achieved in the simulation. Does that mean QMMM method doesn't work in this simulation ? ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? > > Hi cp2k-users, > > After experimenting with cp2k for a while I have written down a short > tutorial on how to use cp2k for biochemical systems. The tutorial covers > the setup of an enzyme system with ambertools followed by equilibration > with cp2k at the MM level, after which we move the system to QM/MM for > metadynamics simulations of an enzymatic reaction. I am far from an expert > myself, but I thought this might be useful to people starting out. If you > spot any mistakes or something that could be improved, please let me know. > The tutorial can be found at https://driesvr.github.io/Tutorials/. > > Kind regards > > Dries > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From dries.va... at gmail.com Wed Jul 19 12:00:58 2017 From: dries.va... at gmail.com (Dries Van Rompaey) Date: Wed, 19 Jul 2017 05:00:58 -0700 (PDT) Subject: CP2K tutorial for biochemical systems In-Reply-To: <2f53575f-67c5-4b47-b448-32baf622bf7a@googlegroups.com> References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> <2f53575f-67c5-4b47-b448-32baf622bf7a@googlegroups.com> Message-ID: Hi, The MM simulations in the tutorial are intended to allow the system to relax and to find suitable cell dimensions. We do this at the MM level because it's rather fast when compared to QMMM. The relaxed system is then moved to the QMMM level in order to simulate bond breaking. The reaction is not sampled in the *unbiased* QMMM, which is why we add metadynamics in the second part. Once we bias the collective variable, you can see the reaction occurring at the QMMM level. Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: > > Hi Dries, > > I have read the tutorial and wonder how could I show the advantage of QMMM > calculation over MM calculation. Should I run a complete MM simulation and > compare the two results ? I notice that you said in the tutorial that the > reaction is not achieved in the simulation. Does that mean QMMM method > doesn't work in this simulation ? > > ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >> >> Hi cp2k-users, >> >> After experimenting with cp2k for a while I have written down a short >> tutorial on how to use cp2k for biochemical systems. The tutorial covers >> the setup of an enzyme system with ambertools followed by equilibration >> with cp2k at the MM level, after which we move the system to QM/MM for >> metadynamics simulations of an enzymatic reaction. I am far from an expert >> myself, but I thought this might be useful to people starting out. If you >> spot any mistakes or something that could be improved, please let me know. >> The tutorial can be found at https://driesvr.github.io/Tutorials/. >> >> Kind regards >> >> Dries >> >> >> -------------- next part -------------- An HTML attachment was scrubbed... URL: From dries.va... at gmail.com Wed Jul 19 12:19:38 2017 From: dries.va... at gmail.com (Dries Van Rompaey) Date: Wed, 19 Jul 2017 05:19:38 -0700 (PDT) Subject: CP2K tutorial for biochemical systems Message-ID: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> Hi cp2k-users, After experimenting with cp2k for a while I have written down a short tutorial on how to use cp2k for biochemical systems. The tutorial covers the setup of an enzyme system with ambertools followed by equilibration with cp2k at the MM level, after which we move the system to QM/MM for metadynamics simulations of an enzymatic reaction. I am far from an expert myself, but I thought this might be useful to people starting out. If you spot any mistakes or something that could be improved, please let me know. The tutorial can be found at https://driesvr.github.io/Tutorials/ as well as on the main cp2k wiki: https://www.cp2k.org/howto:biochem_qmmm. Kind regards Dries -------------- next part -------------- An HTML attachment was scrubbed... URL: From 2922... at qq.com Wed Jul 19 13:09:52 2017 From: 2922... at qq.com (Tonys_Yu) Date: Wed, 19 Jul 2017 06:09:52 -0700 (PDT) Subject: Confusion: cation and anion move towards the same direction under an external electric field In-Reply-To: <17b8b336-c676-4cd5-8ea7-21e10a57b69d@googlegroups.com> References: <9b168bad-d662-4fd5-a156-38cfb58d1bf5@googlegroups.com> <17b8b336-c676-4cd5-8ea7-21e10a57b69d@googlegroups.com> Message-ID: <07672bba-d5e2-49c6-acf4-43a4ea49b77a@googlegroups.com> Hello Marcella, Thanks for your advice. Actually, I also tried the keywords EFIELD and PERIODIC_EFIELD as shown below: &EFIELD ENVELOP CONSTANT &CONSTANT_ENV &END CONSTANT_ENV INTENSITY 0.00194468915 #1V/A=0.0194468915a.u. POLARISATION 1 0 0 &END EFIELD and &PERIODIC_EFIELD INTENSITY 0.00194468915 #1V/A=0.0194468915a.u. POLARISATION 1 0 0 &END PERIODIC_EFIELD but I get an error "On entry to DLASCL parameter number 4 had an illegal value" and the same results that cation and anion move towards the same direction,respectively. Both are strange results. I think some inappropriate keywords and values may exist. Could you have any suggestion of my situation? By the way, my cp2k package is version 2.5. I'm checking for using the latest version. Tonys Yu ? 2017?7?19???? UTC+8??6:32:48?Marcella Iannuzzi??? > > Hi, > > this external potential is not an electric field, it is just a linear > potential. > Check the keywords EFIELD and PERIODIC_EFIELD. > > Marcella > > On Wednesday, July 19, 2017 at 12:20:54 PM UTC+2, Tonys_Yu wrote: >> >> Hello, >> I want to study a ionic displacement in a bulk water box with 30 >> water molecules. External electric field is 0.1V/A, and added by >> EXTERNAL_POTENTIAL command as below: >> >> &EXTERNAL_POTENTIAL >> FUNCTION (A/B)*X >> PARAMETERS A B >> VALUES [eV] 1.0 [angstrom] 10.0 >> &END EXTERNAL_POTENTIAL >> >> What makes me confused is that both water molecules and K-ion are >> moving along the negative direction of X. Moreover, when I changed the >> K-ion by a Cl-ion or reduced the electric field to a smaller value >> (0.001V/A), the same results was observed. It seems contrary to the fact >> that, with the same electric field (value and direction), K-ion and Cl-ion >> would move towards opposite directions. >> I beg someone could shed light on my confusion. Necessary files have >> been attached. >> >> Tonys Yu >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From anu... at gmail.com Wed Jul 19 13:11:59 2017 From: anu... at gmail.com (anurag vohra) Date: Wed, 19 Jul 2017 06:11:59 -0700 (PDT) Subject: positron densities In-Reply-To: <3f982256-2365-4d89-9419-2268bb81ef6e@googlegroups.com> References: <3f982256-2365-4d89-9419-2268bb81ef6e@googlegroups.com> Message-ID: <83819b32-4506-4eae-9cce-856dd8df44bb@googlegroups.com> Hi Asier, did you managed to do something for positron calculation in CP2K? Regards, Anurag On Wednesday, 3 February 2016 22:58:41 UTC+1, Asier Zubiaga wrote: > > Hi, > > I am using cp2k to obtain optimized structures and electron densities for > another code which calculates the positron distribution inside my system. I > am finding quite inefficient this approach and I would prefer to calculate > the positron density using the same program. Typical systems to be studied > are crystalline solids like metals and semiconductors or porous materials > like zeolites. > > Does anybody know if it is possible to include the calculation of a > positron (with the ions and the electron density frozen) inside cp2k? In > case it is not, I would try to write it myself. Is that allowed? > > Btw, I would thank any suggestion about where to start from. > > Cheers, > > Asier > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jts2t... at gmail.com Wed Jul 19 13:12:14 2017 From: jts2t... at gmail.com (jts2t... at gmail.com) Date: Wed, 19 Jul 2017 06:12:14 -0700 (PDT) Subject: CP2K tutorial for biochemical systems In-Reply-To: References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> <2f53575f-67c5-4b47-b448-32baf622bf7a@googlegroups.com> Message-ID: <8b4a499c-ac3e-4886-b49e-3718d351a632@googlegroups.com> Thanks for your reply, Dries ! The first paragraph you said is understood. I also wonder how does one know the reaction really occurs in the simulation ? i.e. Which part of the output file is important to know the information about the reaction ? I major in calculation mathematics, therefore the easy thing for you may seems hard for me. I'll appreciate for your reply ! ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? > > Hi, > > The MM simulations in the tutorial are intended to allow the system to > relax and to find suitable cell dimensions. We do this at the MM level > because it's rather fast when compared to QMMM. The relaxed system is then > moved to the QMMM level in order to simulate bond breaking. > > The reaction is not sampled in the *unbiased* QMMM, which is why we add > metadynamics in the second part. Once we bias the collective variable, you > can see the reaction occurring at the QMMM level. > > > Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >> >> Hi Dries, >> >> I have read the tutorial and wonder how could I show the advantage of >> QMMM calculation over MM calculation. Should I run a complete MM simulation >> and compare the two results ? I notice that you said in the tutorial that >> the reaction is not achieved in the simulation. Does that mean QMMM method >> doesn't work in this simulation ? >> >> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>> >>> Hi cp2k-users, >>> >>> After experimenting with cp2k for a while I have written down a short >>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>> the setup of an enzyme system with ambertools followed by equilibration >>> with cp2k at the MM level, after which we move the system to QM/MM for >>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>> myself, but I thought this might be useful to people starting out. If you >>> spot any mistakes or something that could be improved, please let me know. >>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>> >>> Kind regards >>> >>> Dries >>> >>> >>> ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? > > Hi, > > The MM simulations in the tutorial are intended to allow the system to > relax and to find suitable cell dimensions. We do this at the MM level > because it's rather fast when compared to QMMM. The relaxed system is then > moved to the QMMM level in order to simulate bond breaking. > > The reaction is not sampled in the *unbiased* QMMM, which is why we add > metadynamics in the second part. Once we bias the collective variable, you > can see the reaction occurring at the QMMM level. > > > Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >> >> Hi Dries, >> >> I have read the tutorial and wonder how could I show the advantage of >> QMMM calculation over MM calculation. Should I run a complete MM simulation >> and compare the two results ? I notice that you said in the tutorial that >> the reaction is not achieved in the simulation. Does that mean QMMM method >> doesn't work in this simulation ? >> >> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>> >>> Hi cp2k-users, >>> >>> After experimenting with cp2k for a while I have written down a short >>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>> the setup of an enzyme system with ambertools followed by equilibration >>> with cp2k at the MM level, after which we move the system to QM/MM for >>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>> myself, but I thought this might be useful to people starting out. If you >>> spot any mistakes or something that could be improved, please let me know. >>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>> >>> Kind regards >>> >>> Dries >>> >>> >>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From dries.va... at gmail.com Wed Jul 19 13:40:39 2017 From: dries.va... at gmail.com (Dries Van Rompaey) Date: Wed, 19 Jul 2017 06:40:39 -0700 (PDT) Subject: CP2K tutorial for biochemical systems In-Reply-To: <8b4a499c-ac3e-4886-b49e-3718d351a632@googlegroups.com> References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> <2f53575f-67c5-4b47-b448-32baf622bf7a@googlegroups.com> <8b4a499c-ac3e-4886-b49e-3718d351a632@googlegroups.com> Message-ID: <9f9bedd2-c08f-46bf-a428-2192b882b199@googlegroups.com> Hi, You can observe the reaction occurring by visualising the xyz files of metadynamics qmmm simulation - you will see the C-O bond breaking and the C-C bond forming. You can also plot the collective variable in the metadynlog file: the 3 to 8 bohr region is occupied by the substrate and the -3 to -8 region is occupied by the product. Hope this helps, Dries Op woensdag 19 juli 2017 15:12:14 UTC+2 schreef jts2... at gmail.com: > > Thanks for your reply, Dries ! > > The first paragraph you said is understood. I also wonder how does one > know the reaction really occurs in the simulation ? i.e. Which part of the > output file is important to know the information about the reaction ? > > I major in calculation mathematics, therefore the easy thing for you may > seems hard for me. > I'll appreciate for your reply ! > > > > ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? >> >> Hi, >> >> The MM simulations in the tutorial are intended to allow the system to >> relax and to find suitable cell dimensions. We do this at the MM level >> because it's rather fast when compared to QMMM. The relaxed system is then >> moved to the QMMM level in order to simulate bond breaking. >> >> The reaction is not sampled in the *unbiased* QMMM, which is why we add >> metadynamics in the second part. Once we bias the collective variable, you >> can see the reaction occurring at the QMMM level. >> >> >> Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >>> >>> Hi Dries, >>> >>> I have read the tutorial and wonder how could I show the advantage of >>> QMMM calculation over MM calculation. Should I run a complete MM simulation >>> and compare the two results ? I notice that you said in the tutorial that >>> the reaction is not achieved in the simulation. Does that mean QMMM method >>> doesn't work in this simulation ? >>> >>> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>>> >>>> Hi cp2k-users, >>>> >>>> After experimenting with cp2k for a while I have written down a short >>>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>>> the setup of an enzyme system with ambertools followed by equilibration >>>> with cp2k at the MM level, after which we move the system to QM/MM for >>>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>>> myself, but I thought this might be useful to people starting out. If you >>>> spot any mistakes or something that could be improved, please let me know. >>>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>>> >>>> Kind regards >>>> >>>> Dries >>>> >>>> >>>> > ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? >> >> Hi, >> >> The MM simulations in the tutorial are intended to allow the system to >> relax and to find suitable cell dimensions. We do this at the MM level >> because it's rather fast when compared to QMMM. The relaxed system is then >> moved to the QMMM level in order to simulate bond breaking. >> >> The reaction is not sampled in the *unbiased* QMMM, which is why we add >> metadynamics in the second part. Once we bias the collective variable, you >> can see the reaction occurring at the QMMM level. >> >> >> Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >>> >>> Hi Dries, >>> >>> I have read the tutorial and wonder how could I show the advantage of >>> QMMM calculation over MM calculation. Should I run a complete MM simulation >>> and compare the two results ? I notice that you said in the tutorial that >>> the reaction is not achieved in the simulation. Does that mean QMMM method >>> doesn't work in this simulation ? >>> >>> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>>> >>>> Hi cp2k-users, >>>> >>>> After experimenting with cp2k for a while I have written down a short >>>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>>> the setup of an enzyme system with ambertools followed by equilibration >>>> with cp2k at the MM level, after which we move the system to QM/MM for >>>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>>> myself, but I thought this might be useful to people starting out. If you >>>> spot any mistakes or something that could be improved, please let me know. >>>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>>> >>>> Kind regards >>>> >>>> Dries >>>> >>>> >>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From jts2t... at gmail.com Thu Jul 20 01:15:40 2017 From: jts2t... at gmail.com (jts2t... at gmail.com) Date: Wed, 19 Jul 2017 18:15:40 -0700 (PDT) Subject: CP2K tutorial for biochemical systems In-Reply-To: <9f9bedd2-c08f-46bf-a428-2192b882b199@googlegroups.com> References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> <2f53575f-67c5-4b47-b448-32baf622bf7a@googlegroups.com> <8b4a499c-ac3e-4886-b49e-3718d351a632@googlegroups.com> <9f9bedd2-c08f-46bf-a428-2192b882b199@googlegroups.com> Message-ID: <7cd3d3e0-e91c-4c0c-8567-198732222394@googlegroups.com> Thanks for your patience ! ? 2017?7?19???? UTC+8??9:40:39?Dries Van Rompaey??? > > Hi, > > You can observe the reaction occurring by visualising the xyz files of > metadynamics qmmm simulation - you will see the C-O bond breaking and the > C-C bond forming. You can also plot the collective variable in the > metadynlog file: the 3 to 8 bohr region is occupied by the substrate and > the -3 to -8 region is occupied by the product. > > Hope this helps, > > Dries > > Op woensdag 19 juli 2017 15:12:14 UTC+2 schreef jts2... at gmail.com: >> >> Thanks for your reply, Dries ! >> >> The first paragraph you said is understood. I also wonder how does one >> know the reaction really occurs in the simulation ? i.e. Which part of the >> output file is important to know the information about the reaction ? >> >> I major in calculation mathematics, therefore the easy thing for you may >> seems hard for me. >> I'll appreciate for your reply ! >> >> >> >> ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? >>> >>> Hi, >>> >>> The MM simulations in the tutorial are intended to allow the system to >>> relax and to find suitable cell dimensions. We do this at the MM level >>> because it's rather fast when compared to QMMM. The relaxed system is then >>> moved to the QMMM level in order to simulate bond breaking. >>> >>> The reaction is not sampled in the *unbiased* QMMM, which is why we add >>> metadynamics in the second part. Once we bias the collective variable, you >>> can see the reaction occurring at the QMMM level. >>> >>> >>> Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >>>> >>>> Hi Dries, >>>> >>>> I have read the tutorial and wonder how could I show the advantage of >>>> QMMM calculation over MM calculation. Should I run a complete MM simulation >>>> and compare the two results ? I notice that you said in the tutorial that >>>> the reaction is not achieved in the simulation. Does that mean QMMM method >>>> doesn't work in this simulation ? >>>> >>>> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>>>> >>>>> Hi cp2k-users, >>>>> >>>>> After experimenting with cp2k for a while I have written down a short >>>>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>>>> the setup of an enzyme system with ambertools followed by equilibration >>>>> with cp2k at the MM level, after which we move the system to QM/MM for >>>>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>>>> myself, but I thought this might be useful to people starting out. If you >>>>> spot any mistakes or something that could be improved, please let me know. >>>>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>>>> >>>>> Kind regards >>>>> >>>>> Dries >>>>> >>>>> >>>>> >> ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? >>> >>> Hi, >>> >>> The MM simulations in the tutorial are intended to allow the system to >>> relax and to find suitable cell dimensions. We do this at the MM level >>> because it's rather fast when compared to QMMM. The relaxed system is then >>> moved to the QMMM level in order to simulate bond breaking. >>> >>> The reaction is not sampled in the *unbiased* QMMM, which is why we add >>> metadynamics in the second part. Once we bias the collective variable, you >>> can see the reaction occurring at the QMMM level. >>> >>> >>> Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >>>> >>>> Hi Dries, >>>> >>>> I have read the tutorial and wonder how could I show the advantage of >>>> QMMM calculation over MM calculation. Should I run a complete MM simulation >>>> and compare the two results ? I notice that you said in the tutorial that >>>> the reaction is not achieved in the simulation. Does that mean QMMM method >>>> doesn't work in this simulation ? >>>> >>>> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>>>> >>>>> Hi cp2k-users, >>>>> >>>>> After experimenting with cp2k for a while I have written down a short >>>>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>>>> the setup of an enzyme system with ambertools followed by equilibration >>>>> with cp2k at the MM level, after which we move the system to QM/MM for >>>>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>>>> myself, but I thought this might be useful to people starting out. If you >>>>> spot any mistakes or something that could be improved, please let me know. >>>>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>>>> >>>>> Kind regards >>>>> >>>>> Dries >>>>> >>>>> >>>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From 2922... at qq.com Thu Jul 20 12:11:41 2017 From: 2922... at qq.com (Tonys_Yu) Date: Thu, 20 Jul 2017 05:11:41 -0700 (PDT) Subject: Confusion: cation and anion move towards the same direction under an external electric field In-Reply-To: <17b8b336-c676-4cd5-8ea7-21e10a57b69d@googlegroups.com> References: <9b168bad-d662-4fd5-a156-38cfb58d1bf5@googlegroups.com> <17b8b336-c676-4cd5-8ea7-21e10a57b69d@googlegroups.com> Message-ID: Dear Marcella, I'm sorry to bother you again for my new question. I have checked the keyword of PERIODIC_EFIELD to apply an external electric field 0.1V/A, which seems work well, as shown below. The model contains 30 water molecules and a ion (chloride ion or potassium ion). But I found that, in the two independent calculations, the chloride ion and potassium ion moved towards the direction of electric field and the opposite direction, respectively, which didn't match the reality as far as I know. Am I missing something important? &PERIODIC_EFIELD INTENSITY 0.00194468915 #1V/A=0.0194468915 a.u. POLARISATION 1 0 0 &END PERIODIC_EFIELD Tonys Yu ? 2017?7?19???? UTC+8??6:32:48?Marcella Iannuzzi??? > > Hi, > > this external potential is not an electric field, it is just a linear > potential. > Check the keywords EFIELD and PERIODIC_EFIELD. > > Marcella > > On Wednesday, July 19, 2017 at 12:20:54 PM UTC+2, Tonys_Yu wrote: >> >> Hello, >> I want to study a ionic displacement in a bulk water box with 30 >> water molecules. External electric field is 0.1V/A, and added by >> EXTERNAL_POTENTIAL command as below: >> >> &EXTERNAL_POTENTIAL >> FUNCTION (A/B)*X >> PARAMETERS A B >> VALUES [eV] 1.0 [angstrom] 10.0 >> &END EXTERNAL_POTENTIAL >> >> What makes me confused is that both water molecules and K-ion are >> moving along the negative direction of X. Moreover, when I changed the >> K-ion by a Cl-ion or reduced the electric field to a smaller value >> (0.001V/A), the same results was observed. It seems contrary to the fact >> that, with the same electric field (value and direction), K-ion and Cl-ion >> would move towards opposite directions. >> I beg someone could shed light on my confusion. Necessary files have >> been attached. >> >> Tonys Yu >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From patrick... at gmail.com Thu Jul 20 16:29:12 2017 From: patrick... at gmail.com (Patrick Seewald) Date: Thu, 20 Jul 2017 09:29:12 -0700 (PDT) Subject: cp2k revision 17894 compile fatal error In-Reply-To: <43993c55-6e33-4e65-80c9-879bd969d56a@googlegroups.com> References: <43993c55-6e33-4e65-80c9-879bd969d56a@googlegroups.com> Message-ID: <72bdecab-f4ce-47a3-a58b-f8647bc7ab9a@googlegroups.com> I'm the author of dbcsr_tensor_test module but it is far from obvious to me what may have caused this failure (especially since it is a bug in the compiler, not in the code). Apparently, the bug has been fixed in the intel compiler, so I'd recommend to use a newer version of the compiler. As a quick and dirty solution, you could remove the two files dbcsr_tensor/dbcsr_tensor_test.F dbcsr_tensor/dbcsr_tensor_unittest.F and CP2K will probably compile fine even with the 2015 version of the intel compiler, since these 2 files are compiled into a separate executable and not part of CP2K executable. I'm sorry for not having seen this post sooner. Am Dienstag, 9. Mai 2017 10:29:15 UTC+2 schrieb zhj... at gmail.com: > > Dear CP2K Developers and Users, > > When I compile cp2k revision 17894 by using > *parallel_studio_xe_2015_update2* and *parallel_studio_xe_2016_update4*, > there was a fatal error: > *catastrophic error: **Internal compiler error: segmentation violation > signal raised** Please report this error along with the circumstances in > which it occurred in a Software Problem Report. Note: File and line given > may not be explicit cause of this error.* > *compilation aborted for dbcsr_tensor_test.F90 (code 1)make[3]: *** > [dbcsr_tensor_test.o] Error 1make[3]: *** Waiting for unfinished > jobs....make[2]: *** [all] Error 2make[1]: *** [popt] Error 2make: *** > [all] Error 2* > > But when I use *parallel_studio_xe_2017_update3* instead, this error was > gone, but the MKL lib have some errors. Using cp2k compiled by > *parallel_studio_xe_2017_update3*, the energy can not converge. > > The compiled version is *Linux-x86-64-intel-host.popt*. > I will upload some files about compiled. > -------------- next part -------------- An HTML attachment was scrubbed... URL: From zhj... at gmail.com Fri Jul 21 11:06:47 2017 From: zhj... at gmail.com (zhj... at gmail.com) Date: Fri, 21 Jul 2017 04:06:47 -0700 (PDT) Subject: cp2k revision 17894 compile fatal error In-Reply-To: <72bdecab-f4ce-47a3-a58b-f8647bc7ab9a@googlegroups.com> References: <43993c55-6e33-4e65-80c9-879bd969d56a@googlegroups.com> <72bdecab-f4ce-47a3-a58b-f8647bc7ab9a@googlegroups.com> Message-ID: <6bd18573-4346-4e62-ae1c-094d3b171d70@googlegroups.com> Thanks for your reply, Patrick I will have a try! Zhjs ? 2017?7?21???? UTC+8??12:29:12?Patrick Seewald??? > > I'm the author of dbcsr_tensor_test module but it is far from obvious to > me what may have caused this failure (especially since it is a bug in the > compiler, not in the code). Apparently, the bug has been fixed in the intel > compiler, so I'd recommend to use a newer version of the compiler. > > As a quick and dirty solution, you could remove the two files > dbcsr_tensor/dbcsr_tensor_test.F > dbcsr_tensor/dbcsr_tensor_unittest.F > and CP2K will probably compile fine even with the 2015 version of the > intel compiler, since these 2 files are compiled into a separate executable > and not part of CP2K executable. > > I'm sorry for not having seen this post sooner. > > Am Dienstag, 9. Mai 2017 10:29:15 UTC+2 schrieb zhj... at gmail.com: >> >> Dear CP2K Developers and Users, >> >> When I compile cp2k revision 17894 by using >> *parallel_studio_xe_2015_update2* and *parallel_studio_xe_2016_update4*, >> there was a fatal error: >> *catastrophic error: **Internal compiler error: segmentation violation >> signal raised** Please report this error along with the circumstances in >> which it occurred in a Software Problem Report. Note: File and line given >> may not be explicit cause of this error.* >> *compilation aborted for dbcsr_tensor_test.F90 (code 1)make[3]: *** >> [dbcsr_tensor_test.o] Error 1make[3]: *** Waiting for unfinished >> jobs....make[2]: *** [all] Error 2make[1]: *** [popt] Error 2make: *** >> [all] Error 2* >> >> But when I use *parallel_studio_xe_2017_update3* instead, this error was >> gone, but the MKL lib have some errors. Using cp2k compiled by >> *parallel_studio_xe_2017_update3*, the energy can not converge. >> >> The compiled version is *Linux-x86-64-intel-host.popt*. >> I will upload some files about compiled. >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From fa... at squashclub.org Fri Jul 21 13:56:06 2017 From: fa... at squashclub.org (Faraz H) Date: Fri, 21 Jul 2017 06:56:06 -0700 (PDT) Subject: mpi job runs fine when submitted via scheduler, but fails when run by hand Message-ID: <8035a1ca-9295-462f-a761-583cab886b10@googlegroups.com> I have a simple job submission script which runs fine. However, when I try and run the same script by hand ( outside the scheduler), it fails with the error below. By hand I mean that I take the nodes offline, login to the first node, load the cp2k/open mpi modules, keep same machine file, then run the command: mpirun -machinefile machines -np 40 --mca btl openib,self cp2k.popt test.inp mpirun was unable to find the specified executable file, and therefore did not launch the job. This error was first reported for process rank 20; it may have occurred for other processes as well. NOTE: A common cause for this error is misspelling a mpirun command line parameter option (remember that mpirun interprets the first unrecognized command line token as the executable). Node: lustwzb5 Executable: /gpfs/app/software/cp2k/3.0.0/exe/Linux_x86-64-intel/cp2k.popt -------------- next part -------------- An HTML attachment was scrubbed... URL: From dries.va... at gmail.com Fri Jul 21 15:14:28 2017 From: dries.va... at gmail.com (Dries Van Rompaey) Date: Fri, 21 Jul 2017 08:14:28 -0700 (PDT) Subject: CP2K tutorial for biochemical systems In-Reply-To: <7cd3d3e0-e91c-4c0c-8567-198732222394@googlegroups.com> References: <4c21c035-5da7-448c-a215-154efba53d4e@googlegroups.com> <2f53575f-67c5-4b47-b448-32baf622bf7a@googlegroups.com> <8b4a499c-ac3e-4886-b49e-3718d351a632@googlegroups.com> <9f9bedd2-c08f-46bf-a428-2192b882b199@googlegroups.com> <7cd3d3e0-e91c-4c0c-8567-198732222394@googlegroups.com> Message-ID: You're welcome. The tutorial has been updated with a brief section on how to include enzyme residues in the QM region. As always, comments and improvements are welcome. Op donderdag 20 juli 2017 03:15:40 UTC+2 schreef jts2... at gmail.com: > > Thanks for your patience ! > > > > ? 2017?7?19???? UTC+8??9:40:39?Dries Van Rompaey??? >> >> Hi, >> >> You can observe the reaction occurring by visualising the xyz files of >> metadynamics qmmm simulation - you will see the C-O bond breaking and the >> C-C bond forming. You can also plot the collective variable in the >> metadynlog file: the 3 to 8 bohr region is occupied by the substrate and >> the -3 to -8 region is occupied by the product. >> >> Hope this helps, >> >> Dries >> >> Op woensdag 19 juli 2017 15:12:14 UTC+2 schreef jts2... at gmail.com: >>> >>> Thanks for your reply, Dries ! >>> >>> The first paragraph you said is understood. I also wonder how does one >>> know the reaction really occurs in the simulation ? i.e. Which part of the >>> output file is important to know the information about the reaction ? >>> >>> I major in calculation mathematics, therefore the easy thing for you >>> may seems hard for me. >>> I'll appreciate for your reply ! >>> >>> >>> >>> ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? >>>> >>>> Hi, >>>> >>>> The MM simulations in the tutorial are intended to allow the system to >>>> relax and to find suitable cell dimensions. We do this at the MM level >>>> because it's rather fast when compared to QMMM. The relaxed system is then >>>> moved to the QMMM level in order to simulate bond breaking. >>>> >>>> The reaction is not sampled in the *unbiased* QMMM, which is why we >>>> add metadynamics in the second part. Once we bias the collective variable, >>>> you can see the reaction occurring at the QMMM level. >>>> >>>> >>>> Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >>>>> >>>>> Hi Dries, >>>>> >>>>> I have read the tutorial and wonder how could I show the advantage of >>>>> QMMM calculation over MM calculation. Should I run a complete MM simulation >>>>> and compare the two results ? I notice that you said in the tutorial that >>>>> the reaction is not achieved in the simulation. Does that mean QMMM method >>>>> doesn't work in this simulation ? >>>>> >>>>> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>>>>> >>>>>> Hi cp2k-users, >>>>>> >>>>>> After experimenting with cp2k for a while I have written down a short >>>>>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>>>>> the setup of an enzyme system with ambertools followed by equilibration >>>>>> with cp2k at the MM level, after which we move the system to QM/MM for >>>>>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>>>>> myself, but I thought this might be useful to people starting out. If you >>>>>> spot any mistakes or something that could be improved, please let me know. >>>>>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>>>>> >>>>>> Kind regards >>>>>> >>>>>> Dries >>>>>> >>>>>> >>>>>> >>> ? 2017?7?19???? UTC+8??8:00:59?Dries Van Rompaey??? >>>> >>>> Hi, >>>> >>>> The MM simulations in the tutorial are intended to allow the system to >>>> relax and to find suitable cell dimensions. We do this at the MM level >>>> because it's rather fast when compared to QMMM. The relaxed system is then >>>> moved to the QMMM level in order to simulate bond breaking. >>>> >>>> The reaction is not sampled in the *unbiased* QMMM, which is why we >>>> add metadynamics in the second part. Once we bias the collective variable, >>>> you can see the reaction occurring at the QMMM level. >>>> >>>> >>>> Op woensdag 19 juli 2017 13:34:52 UTC+2 schreef jts2... at gmail.com: >>>>> >>>>> Hi Dries, >>>>> >>>>> I have read the tutorial and wonder how could I show the advantage of >>>>> QMMM calculation over MM calculation. Should I run a complete MM simulation >>>>> and compare the two results ? I notice that you said in the tutorial that >>>>> the reaction is not achieved in the simulation. Does that mean QMMM method >>>>> doesn't work in this simulation ? >>>>> >>>>> ? 2017?7?18???? UTC+8??2:49:11?Dries Van Rompaey??? >>>>>> >>>>>> Hi cp2k-users, >>>>>> >>>>>> After experimenting with cp2k for a while I have written down a short >>>>>> tutorial on how to use cp2k for biochemical systems. The tutorial covers >>>>>> the setup of an enzyme system with ambertools followed by equilibration >>>>>> with cp2k at the MM level, after which we move the system to QM/MM for >>>>>> metadynamics simulations of an enzymatic reaction. I am far from an expert >>>>>> myself, but I thought this might be useful to people starting out. If you >>>>>> spot any mistakes or something that could be improved, please let me know. >>>>>> The tutorial can be found at https://driesvr.github.io/Tutorials/. >>>>>> >>>>>> Kind regards >>>>>> >>>>>> Dries >>>>>> >>>>>> >>>>>> -------------- next part -------------- An HTML attachment was scrubbed... URL: From lksz... at gmail.com Fri Jul 21 23:31:52 2017 From: lksz... at gmail.com (Chris K. Lee) Date: Fri, 21 Jul 2017 16:31:52 -0700 (PDT) Subject: Some questions in the NEB example. In-Reply-To: References: Message-ID: Hello, anybody home? Is there nobody to explain the keywords of the inputs? 2017? 6? 29? ??? ?? 3? 4? 37? UTC+9, Chris K. Lee ?? ?: > Dear CP2K users, > > > > I have some questions in the NEB example. > > (*https://www.cp2k.org/exercises:2015_cecam_tutorial:neb* > ) > > It?s so hard to understand only with the Input reference. > > (*https://manual.cp2k.org* ) > > > > > > At ./NEB/EX2/neb_GPW.inp > > > > > > 1. line 2 > > &MULTIPLE_FORCE_EVALS > > FORCE_EVAL_ORDER 2 3 > > > > ?FORCE_EVAL_ORDER {Integer} ...? : Specify the orders of the different > force_eval. > > > > What does ?orders? mean? What I guess for ?2 3? is > > There are three FORCE_EVAL sections and 1st one only contains the > information. > > The 2nd and 3rd ones act as usual ?FORCE_EVAL?. Am I right? > > > > 2. line 9 > > &FORCE_EVAL / &MIXED > > GROUP_PARTITION 2 16 > > > > ?GROUP_PARTITION {Integer} ...? : gives the exact number of processors for > each group > > > > Is this mean that using 2 cpu processors for the 2nd FORCE_EVAL and 16 for > the 3rd one? > > At ./NEB/EX2/neb.inp ?GROUP_PARTITION 1 1? means it only use total 2 > processors? > > > > 3. line 12, 13 > > &FORCE_EVAL / &MIXED / &GENERIC > > MIXING_FUNCTION E1+E2 > > VARIABLES E1 E2 > > > > What does these two terms mean and how it works? > > Is this ?VARIABLES? only for energy? > > Does the CP2K understands that the variables as energy of 2nd and 3rd > FORCE_EVAL? > > Which energy would be printed on the output? > > > > 4. line 24, 27 > > &FORCE_EVAL / &MIXED / &MAPPING > > &FORCE_EVAL 1 , &FORCE_EVAL 2 > > > > Does it mean that the 2nd FORCE_EVAL would treated as ?1? and the 3rd one > as ?2? ? > > > > 5. line 33 > > &FORCE_EVAL / &SUBSYS > > &COLVAR > > > > Does this ?COLVAR? make automatically its own output without PRINT options? > > > > 6. line 72 > > &FORCE_EVAL / &MM / &FORCEFIELD / &NONBONDED / &GENPOT > > atoms Cu C > > FUNCTION A*exp(-av*r)+B*exp(-ac*r)-C/(r^6) > > VARIABLES r > > > > ?VARIABLES {Word} ... ? : Defines the variable of the functional form. > > > > It shows they are trying to use ?r? as variable but that?s all. > > Does CP2K understand that ?r? as distance between Cu and C? > > Only the distance should be the ?variables? here? > > > > 7. line 213 > > &MOTION / &BAND > > NPROC_REP 18 > > > > Is this option useless for the serial run like ssmp? > > > > 8. REPLICAs > > What I under stand is that they used the intermediate coordinates as > replica. > > They already optimized geometry for starting materials and final products. > > How they get the optimized intermediate coordinates of the replica? > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From christia... at gmail.com Fri Jul 21 23:42:33 2017 From: christia... at gmail.com (Christian Leitold) Date: Fri, 21 Jul 2017 16:42:33 -0700 (PDT) Subject: Basic tutorial for all-electron calculations (HF, MP2, ...) Message-ID: <9ef46dff-8f71-49b0-8bdd-1ad2470d6f30@googlegroups.com> Hi, I am new in the field of electronic calculations, and I am looking for some basic tutorials. In particular, right now I am interested in doing HF and MP2 calculations for simple molecules in the gas phase, no PBCs, and no DFT. The tutorials I could find so far look really good, but I could only find them for DFT calculations. Could you maybe point me to some examples of Hartree-Fock-only calculations? Thanks, Christian -------------- next part -------------- An HTML attachment was scrubbed... URL: From zuohen... at gmail.com Tue Jul 25 06:20:11 2017 From: zuohen... at gmail.com (zuohen... at gmail.com) Date: Mon, 24 Jul 2017 23:20:11 -0700 (PDT) Subject: Problem with Singlet and Triplet spin state calculation In-Reply-To: <6ede00bf-e412-4e43-915a-edbaf9892955@googlegroups.com> References: <6ede00bf-e412-4e43-915a-edbaf9892955@googlegroups.com> Message-ID: Hello, my friend. Do you solve this problem now? I meet this problem too. Thank you very much. ? 2012?7?4???? UTC+8??8:23:33?bharat??? > > Dear Experts, > > I am new to CP2K. I am testing my calculation accuracy in the CP2K for the > small system. I calculated the singlet and triplet for ZnO. Finally I have > found the SAME ENERGY AND BOND LENGTH. It seems that there is no difference > between MULTIPLICITY 1 and 3 for this system. I have posted my INPUT file. > Can you please give any IDEA why I got the same result. > > Thank you in advance. > > Sincerey, > Bharat Sharma > > > &FORCE_EVAL > METHOD Quickstep > &DFT > MULTIPLICITY 1* # I have used 3 for triplet state* > &QS > METHOD MNDO > &SE > &END > > &END QS > &SCF > SCF_GUESS ATOMIC > &END SCF > &END DFT > &SUBSYS > &CELL > ABC 20.0 20.0 20.0 > PERIODIC NONE > &END CELL > &COORD > Zn -1.47619650 0.72649572 0.00000000 > O 0.22380350 0.72649572 0.00000000 > &END COORD > &END SUBSYS > &END FORCE_EVAL > &GLOBAL > > PROJECT zno > RUN_TYPE GEO_OPT > &END GLOBAL > -------------- next part -------------- An HTML attachment was scrubbed... URL: From hut... at chem.uzh.ch Tue Jul 25 09:11:10 2017 From: hut... at chem.uzh.ch (hut... at chem.uzh.ch) Date: Tue, 25 Jul 2017 11:11:10 +0200 Subject: [CP2K:9255] Re: Problem with Singlet and Triplet spin state calculation In-Reply-To: References: , <6ede00bf-e412-4e43-915a-edbaf9892955@googlegroups.com> Message-ID: Hi you need to set the Keyword UKS or LSD at CP2K_INPUT / FORCE_EVAL / DFT for Triplet calculations. regards Juerg -------------------------------------------------------------- Juerg Hutter Phone : ++41 44 635 4491 Institut f?r Chemie C FAX : ++41 44 635 6838 Universit?t Z?rich E-mail: hut... at chem.uzh.ch Winterthurerstrasse 190 CH-8057 Z?rich, Switzerland --------------------------------------------------------------- -----cp... at googlegroups.com wrote: -----To: cp2k From: zuohen... at gmail.com Sent by: cp... at googlegroups.com Date: 07/25/2017 08:20AM Subject: [CP2K:9255] Re: Problem with Singlet and Triplet spin state calculation Hello, my friend. Do you solve this problem now?I meet this problem too. Thank you very much. ? 2012?7?4???? UTC+8??8:23:33?bharat???Dear Experts, I am new to CP2K. I am testing my calculation accuracy in the CP2K for the small system. I calculated the singlet and triplet for ZnO. Finally I have found the SAME ENERGY AND BOND LENGTH. It seems that there is no difference between MULTIPLICITY 1 and 3 for this system. I have posted my INPUT file. Can you please give any IDEA why I got the same result. Thank you in advance. Sincerey,Bharat Sharma &FORCE_EVAL METHOD Quickstep &DFT MULTIPLICITY 1 # I have used 3 for triplet state &QS METHOD MNDO &SE &END &END QS &SCF SCF_GUESS ATOMIC &END SCF &END DFT &SUBSYS &CELL ABC 20.0 20.0 20.0 PERIODIC NONE &END CELL &COORD Zn -1.47619650 0.72649572 0.00000000 O 0.22380350 0.72649572 0.00000000 &END COORD &END SUBSYS&END FORCE_EVAL&GLOBAL PROJECT zno RUN_TYPE GEO_OPT&END GLOBAL -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. From natalie... at gmail.com Tue Jul 25 16:17:42 2017 From: natalie... at gmail.com (Natalie Austin) Date: Tue, 25 Jul 2017 09:17:42 -0700 (PDT) Subject: Solvation model SCCS convergence issues Message-ID: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> Hello, I am having issues converging SCCS iterations and my SCF cycles are taking long to converge as well, when applying the SCCS solvation model in CP2K. I incorporated the suggesting given in this previous posting: https://groups.google.com/forum/#!searchin/cp2k/sccs%7Csort:relevance/cp2k/8VHkm_Qi5Fo/3Je5sy8YFzsJ but my calculations are still not converging. Any help with this would be appreciated. My input file is attached as au.inp. Thanks, Natalie -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: au.inp.docx Type: application/vnd.openxmlformats-officedocument.wordprocessingml.document Size: 13100 bytes Desc: not available URL: From natalie... at gmail.com Tue Jul 25 16:28:27 2017 From: natalie... at gmail.com (Natalie Austin) Date: Tue, 25 Jul 2017 09:28:27 -0700 (PDT) Subject: Solvation model SCCS convergence issues In-Reply-To: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> References: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> Message-ID: <0aab8946-157f-4109-abf4-9ed950f43d6a@googlegroups.com> I have also attached the current output file (au.out). I am now getting an error that states: *** WARNING in pw/ps_wavelet_methods.F:236 :: Density non-zero on the *** *** edges of the unit cell: wrong results in WAVELET solver *** Natalie -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: au.out Type: application/octet-stream Size: 201737 bytes Desc: not available URL: From matthia... at psi.ch Wed Jul 26 08:39:03 2017 From: matthia... at psi.ch (Matthias Krack) Date: Wed, 26 Jul 2017 01:39:03 -0700 (PDT) Subject: Solvation model SCCS convergence issues In-Reply-To: <0aab8946-157f-4109-abf4-9ed950f43d6a@googlegroups.com> References: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> <0aab8946-157f-4109-abf4-9ed950f43d6a@googlegroups.com> Message-ID: Dear Natalie To the best of my knowledge, SCCS has never been validated to work with the Wavelet solver properly. Matthias From natalie... at gmail.com Wed Jul 26 16:07:42 2017 From: natalie... at gmail.com (Natalie Austin) Date: Wed, 26 Jul 2017 09:07:42 -0700 (PDT) Subject: Solvation model SCCS convergence issues In-Reply-To: References: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> <0aab8946-157f-4109-abf4-9ed950f43d6a@googlegroups.com> Message-ID: <20ef8d34-f876-4bdf-9735-f8e8c4ef85c9@googlegroups.com> So is it only validated to work properly for inputs that have been used in the regtests? like : https://sourceforge.net/p/cp2k/code/HEAD/tree/trunk/cp2k/tests/QS/regtest-sccs-1/ On Wednesday, July 26, 2017 at 4:39:03 AM UTC-4, Matthias Krack wrote: > > Dear Natalie > > To the best of my knowledge, SCCS has never been validated to work with > the Wavelet solver properly. > > Matthias > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From matthia... at psi.ch Wed Jul 26 18:13:04 2017 From: matthia... at psi.ch (Matthias Krack) Date: Wed, 26 Jul 2017 11:13:04 -0700 (PDT) Subject: Solvation model SCCS convergence issues In-Reply-To: <20ef8d34-f876-4bdf-9735-f8e8c4ef85c9@googlegroups.com> References: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> <0aab8946-157f-4109-abf4-9ed950f43d6a@googlegroups.com> <20ef8d34-f876-4bdf-9735-f8e8c4ef85c9@googlegroups.com> Message-ID: <82285a65-0984-45a9-9a50-2305be8db16a@googlegroups.com> Correct. Is it mandatory to employ the wavelet solver? From christia... at gmail.com Wed Jul 26 23:17:56 2017 From: christia... at gmail.com (Christian Leitold) Date: Wed, 26 Jul 2017 16:17:56 -0700 (PDT) Subject: MD simulation: print contributions to total energy (bonds, angles, etc.) Message-ID: <00fbe17d-83f1-4452-9237-4ba7c30b5f68@googlegroups.com> Hi, Once again I have a very simple question, namely, how to output the different energy contributions in a classical MD simulation. I have just set up a very simple simulation containing two water molecules (TIP4P). Since this is a rigid molecule, only non-bonded LJ and electrostatic interactions should contribute to the total potential energy, but not bonds, angles, and dihedrals (at the very least, all these interactions should be constants for a rigid molecule). My question is, how can I check the different contributions to the potential energy? Unfortunately, I could not find this neither in the documentation nor when googling it. Thanks, Christian -------------- next part -------------- An HTML attachment was scrubbed... URL: From hqz... at nju.edu.cn Thu Jul 27 06:22:22 2017 From: hqz... at nju.edu.cn (=?utf-8?B?aHF6aG91?=) Date: Thu, 27 Jul 2017 14:22:22 +0800 Subject: Unddefined reference when compiling cp2k 4.1 with CUDA 8.0 Message-ID: Hi, I'm compiling cp2k 4.1 with CUDA 8.0 using gnu compiler gfortran 4.8.5 on a CentOS 7.3 machine. For the compilation going on smoothly, I have made some modifications to the relevant source code: 1. change syncthreads() to __syncthreads() 2. added condition to support native atomicAdd() in CUDA 8.0 3. change FFT_ALIGNMENT to CUFFT_COMPATIBILITY_FFTW_PADDING The compilation went almost OK, except the last step of creating cp2k.ssmp. I got following error. Actually, all undefined references are defined in pw/pw_cuda.F, am I wrong somewhere in my configuration (refer to the attached)? Thanks in advance! PS: I can run dbcsr_unittest.ssmp, libcusmm_unittest.ssmp without any problem. Huiqun Zhou @Earth Sciences, Nanjing University, China ======================= snippet of screen output ========================================== gfortran -I/usr/local/include -I/usr/local/cuda/include -I/usr/local/libint/include -I/usr/local/libxc/include -D__GRID_CORE=2 -D__PILAENV_BLOCKSIZE=1024 -D__ACC -D__DBCSR_ACC -D__PW_CUDA -D__LIBXC2 -D__LIBINT -D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5 -D__MAX_CONTR=4 -O2 -ffast-math -ffree-form -ffree-line-length-none -fopenmp -ftree-vectorize -funroll-loops -mtune=native -I/usr/local/include -I/usr/local/cuda/include -I/usr/local/libint/include -I/usr/local/libxc/include -D__COMPILE_ARCH="\"unclecuda-gnu\"" -D__COMPILE_DATE="\"Thu Jul 27 14:08:24 CST 2017\"" -D__COMPILE_HOST="\"unclecuda\"" -D__COMPILE_REVISION="\"svn:17462\"" -D__DATA_DIR="\"/home/zhou_huiqun/Cookplate/cp2k-4.1/data\"" -L/home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp -o /home/zhou_huiqun/Cookplate/cp2k-4.1/exe/unclecuda-gnu/cp2k.ssmp cp2k.o -lcp2kstart -lcp2kmc -lcp2kswarm -lcp2kmotion -lcp2kthermostat -lcp2kemd -lcp2ktmc -lcp2kmain -lcp2keri_mme -lcp2kminimax -lcp2khfxbase -lcp2ksubsys -lcp2kxc -lcp2kao -lcp2kinput -lcp2kpw -lcp2kpwcuda -lcp2kfft -lcp2kdbcsrwrap -ldbcsr -ldbcsrarnoldi -ldbcsrvec -ldbcsrmm -ldbcsrops -ldbcsrbblock -ldbcsrbdist -ldbcsrbase -ldbcsrdata -ldbcsrerr -lmicsmm -lclsmm -lcusmm -lcp2kacc -lcp2kaccmic -lcp2kaccopencl -lcp2kacccuda -lcp2kfm -lcp2kcommon -lcp2kmpiwrap -lcp2kgrid -lcp2kbase -L/usr/local/cuda/lib64 -lcudart -lcublas -lcufft -lrt /usr/lib64/libopenblas.a /usr/local/libxc/lib/libxcf90.a /usr/local/libxc/lib/libxc.a /usr/local/libint/lib/libderiv.a /usr/local/libint/lib/libint.a /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_f': pw_cuda.F:(.text+0x6c): undefined reference to `pw_cuda_f_z_' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_sf': pw_cuda.F:(.text+0x1ef): undefined reference to `pw_cuda_sf_z_' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_fg': pw_cuda.F:(.text+0x36e): undefined reference to `pw_cuda_fg_z_' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d_ps': pw_cuda.F:(.text+0x3673): undefined reference to `pw_cuda_fc_z_' pw_cuda.F:(.text+0x3b6e): undefined reference to `pw_cuda_ffc_z_' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d_ps': pw_cuda.F:(.text+0x6ca5): undefined reference to `pw_cuda_cf_z_' pw_cuda.F:(.text+0x7336): undefined reference to `pw_cuda_cff_z_' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d': pw_cuda.F:(.text+0x779c): undefined reference to `pw_cuda_sfffc_z_' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d': pw_cuda.F:(.text+0x79d0): undefined reference to `pw_cuda_cfffg_z_' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_init': pw_cuda.F:(.text+0x7a55): undefined reference to `pw_cuda_init' /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): In function `__pw_cuda_MOD_pw_cuda_finalize': pw_cuda.F:(.text+0x7a41): undefined reference to `pw_cuda_finalize' collect2: error: ld returned 1 exit status -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: unclecuda-gnu.ssmp Type: application/octet-stream Size: 1128 bytes Desc: not available URL: From tiziano... at chem.uzh.ch Thu Jul 27 09:10:19 2017 From: tiziano... at chem.uzh.ch (=?UTF-8?Q?Tiziano_M=c3=bcller?=) Date: Thu, 27 Jul 2017 11:10:19 +0200 Subject: [CP2K:9262] Unddefined reference when compiling cp2k 4.1 with CUDA 8.0 In-Reply-To: References: Message-ID: <8b269b70-05d0-426a-a32a-8bf778e7792b@chem.uzh.ch> Hi Huiqun Can you please attach the complete build log? It might also be a good idea to do a `make distclean` and rebuild everything from scratch for that. The mentioned symbols should be defined in lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a You can check for their presence using `nm lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a` which should give something like this: [...] pw_cuda_z.o: [...] 00000000000003a0 T pw_cuda_cfffg_z_ 0000000000000f60 T pw_cuda_cff_z_ 0000000000001680 T pw_cuda_cf_z_ 00000000000019e0 T pw_cuda_fc_z_ 00000000000012f0 T pw_cuda_ffc_z_ 0000000000001f00 T pw_cuda_fg_z_ 0000000000001d40 T pw_cuda_f_z_ 0000000000000960 T pw_cuda_sfffc_z_ 0000000000002310 T pw_cuda_sf_z_ [...] If they are not there (after the rebuild), please also attach the complete output of this command. Best regards, Tiziano On 27.07.2017 08:22, hqzhou wrote: > Hi, > > I'm compiling cp2k 4.1 with CUDA 8.0 using gnu compiler gfortran 4.8.5 > on a CentOS 7.3 machine. > For the compilation going on smoothly, I have made some modifications to > the relevant source > code: > 1. change syncthreads() to __syncthreads() > 2. added condition to support native atomicAdd() in CUDA 8.0 > 3. change FFT_ALIGNMENT to CUFFT_COMPATIBILITY_FFTW_PADDING > > The compilation went almost OK, except the last step of creating > cp2k.ssmp. I got following error. > Actually, all undefined references are defined in pw/pw_cuda.F, am I > wrong somewhere in my > configuration (refer to the attached)? > > Thanks in advance! > > PS: I can run dbcsr_unittest.ssmp, libcusmm_unittest.ssmp without any > problem. > > Huiqun Zhou > @Earth Sciences, Nanjing University, China > > ======================= snippet of screen output > ========================================== > gfortran -I/usr/local/include -I/usr/local/cuda/include > -I/usr/local/libint/include -I/usr/local/libxc/include -D__GRID_CORE=2 > -D__PILAENV_BLOCKSIZE=1024 -D__ACC -D__DBCSR_ACC -D__PW_CUDA -D__LIBXC2 > -D__LIBINT -D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5 -D__MAX_CONTR=4 > -O2 -ffast-math -ffree-form -ffree-line-length-none -fopenmp > -ftree-vectorize -funroll-loops -mtune=native -I/usr/local/include > -I/usr/local/cuda/include -I/usr/local/libint/include > -I/usr/local/libxc/include -D__COMPILE_ARCH="\"unclecuda-gnu\"" > -D__COMPILE_DATE="\"Thu Jul 27 14:08:24 CST 2017\"" > -D__COMPILE_HOST="\"unclecuda\"" -D__COMPILE_REVISION="\"svn:17462\"" > -D__DATA_DIR="\"/home/zhou_huiqun/Cookplate/cp2k-4.1/data\"" > -L/home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp -o > /home/zhou_huiqun/Cookplate/cp2k-4.1/exe/unclecuda-gnu/cp2k.ssmp cp2k.o > -lcp2kstart -lcp2kmc -lcp2kswarm -lcp2kmotion -lcp2kthermostat -lcp2kemd > -lcp2ktmc -lcp2kmain -lcp2keri_mme -lcp2kminimax -lcp2khfxbase > -lcp2ksubsys -lcp2kxc -lcp2kao -lcp2kinput -lcp2kpw -lcp2kpwcuda > -lcp2kfft -lcp2kdbcsrwrap -ldbcsr -ldbcsrarnoldi -ldbcsrvec -ldbcsrmm > -ldbcsrops -ldbcsrbblock -ldbcsrbdist -ldbcsrbase -ldbcsrdata -ldbcsrerr > -lmicsmm -lclsmm -lcusmm -lcp2kacc -lcp2kaccmic -lcp2kaccopencl > -lcp2kacccuda -lcp2kfm -lcp2kcommon -lcp2kmpiwrap -lcp2kgrid -lcp2kbase > -L/usr/local/cuda/lib64 -lcudart -lcublas -lcufft -lrt > /usr/lib64/libopenblas.a /usr/local/libxc/lib/libxcf90.a > /usr/local/libxc/lib/libxc.a /usr/local/libint/lib/libderiv.a > /usr/local/libint/lib/libint.a > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_f': > pw_cuda.F:(.text+0x6c): undefined reference to `pw_cuda_f_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_sf': > pw_cuda.F:(.text+0x1ef): undefined reference to `pw_cuda_sf_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_fg': > pw_cuda.F:(.text+0x36e): undefined reference to `pw_cuda_fg_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d_ps': > pw_cuda.F:(.text+0x3673): undefined reference to `pw_cuda_fc_z_' > pw_cuda.F:(.text+0x3b6e): undefined reference to `pw_cuda_ffc_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d_ps': > pw_cuda.F:(.text+0x6ca5): undefined reference to `pw_cuda_cf_z_' > pw_cuda.F:(.text+0x7336): undefined reference to `pw_cuda_cff_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d': > pw_cuda.F:(.text+0x779c): undefined reference to `pw_cuda_sfffc_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d': > pw_cuda.F:(.text+0x79d0): undefined reference to `pw_cuda_cfffg_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_init': > pw_cuda.F:(.text+0x7a55): undefined reference to `pw_cuda_init' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_finalize': > pw_cuda.F:(.text+0x7a41): undefined reference to `pw_cuda_finalize' > collect2: error: ld returned 1 exit status > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com > . > To post to this group, send email to cp... at googlegroups.com > . > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. -- Tiziano M?ller University of Zurich Department of Chemistry Winterthurerstrasse 190 CH-8057 Z?rich Tel: +41 44 63 54234 www.chem.uzh.ch tiziano... at chem.uzh.ch From feliped... at gmail.com Thu Jul 27 10:18:12 2017 From: feliped... at gmail.com (fquesada) Date: Thu, 27 Jul 2017 03:18:12 -0700 (PDT) Subject: Issue with Mg cell optimization Message-ID: Dear CP2K users, I am working on a model of magnesium and I am interested in relaxing the cell parameters of the unit cell. I wrote a code to perform the cell optimization, but I am having the following issues with it: 1) the energy values are not converging properly during the optimization procedure, and 2) the sum of forces on the atoms is never reaches zero. I have tried modifying different parameters in my code (such as the CUTOFF, REL_CUTOFF, KPOINTS, EPS_SCF, ADDED_MOS, and even a different unit cell), but the problems still persist. I attached the code below for reference. I am clueless as to why these issues persist and I would really appreciate any help on the matter. Thank you in advance for the response. Cheers, Felipe -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: Magnesium_Cell_Optimization.pdf Type: application/pdf Size: 17570 bytes Desc: not available URL: From hut... at chem.uzh.ch Thu Jul 27 10:50:55 2017 From: hut... at chem.uzh.ch (hut... at chem.uzh.ch) Date: Thu, 27 Jul 2017 12:50:55 +0200 Subject: [CP2K:9265] Issue with Mg cell optimization In-Reply-To: References: Message-ID: Hi your input file looks OK to me. I would make the following changes EPS_DEFAULT 1.e-12 (or less) REL_CUTOFF 75 CUTOFF 1000 and introduce a reference cell &CELL_REF in &CELL in order to get constant number of grid points during optimization. regards Juerg -------------------------------------------------------------- Juerg Hutter Phone : ++41 44 635 4491 Institut f?r Chemie C FAX : ++41 44 635 6838 Universit?t Z?rich E-mail: hut... at chem.uzh.ch Winterthurerstrasse 190 CH-8057 Z?rich, Switzerland --------------------------------------------------------------- -----cp... at googlegroups.com wrote: -----To: cp2k From: fquesada Sent by: cp... at googlegroups.com Date: 07/27/2017 12:18PM Subject: [CP2K:9265] Issue with Mg cell optimization Dear CP2K users, I am working on a model of magnesium and I am interested in relaxing the cell parameters of the unit cell. I wrote a code to perform the cell optimization, but I am having the following issues with it: 1) the energy values are not converging properly during the optimization procedure, and 2) the sum of forces on the atoms is never reaches zero. I have tried modifying different parameters in my code (such as the CUTOFF, REL_CUTOFF, KPOINTS, EPS_SCF, ADDED_MOS, and even a different unit cell), but the problems still persist. I attached the code below for reference. I am clueless as to why these issues persist and I would really appreciate any help on the matter. Thank you in advance for the response. Cheers, Felipe -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. [attachment "Magnesium_Cell_Optimization.pdf" removed by J?rg Hutter/at/UZH] From hqz... at nju.edu.cn Thu Jul 27 15:06:59 2017 From: hqz... at nju.edu.cn (=?utf-8?B?aHF6aG91?=) Date: Thu, 27 Jul 2017 23:06:59 +0800 Subject: [CP2K:9263] Unddefined reference when compiling cp2k 4.1 withCUDA 8.0 In-Reply-To: <8b269b70-05d0-426a-a32a-8bf778e7792b@chem.uzh.ch> References: <8b269b70-05d0-426a-a32a-8bf778e7792b@chem.uzh.ch> Message-ID: Hi, Tiziano Thank you for your quick response! Here are attached the compile log and output of nm libcp2kpwcuda.a. It seems that no such references exist in libcp2kpwcuda.a. Thanks for resolving the problem! Huiqun ------------------ Original ------------------ From: "Tiziano M?ller"; Date: Thu, Jul 27, 2017 05:10 PM To: "hqzhou"; Cc: "cp2k"; Subject: Re: [CP2K:9263] Unddefined reference when compiling cp2k 4.1 withCUDA 8.0 Hi Huiqun Can you please attach the complete build log? It might also be a good idea to do a `make distclean` and rebuild everything from scratch for that. The mentioned symbols should be defined in lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a You can check for their presence using `nm lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a` which should give something like this: [...] pw_cuda_z.o: [...] 00000000000003a0 T pw_cuda_cfffg_z_ 0000000000000f60 T pw_cuda_cff_z_ 0000000000001680 T pw_cuda_cf_z_ 00000000000019e0 T pw_cuda_fc_z_ 00000000000012f0 T pw_cuda_ffc_z_ 0000000000001f00 T pw_cuda_fg_z_ 0000000000001d40 T pw_cuda_f_z_ 0000000000000960 T pw_cuda_sfffc_z_ 0000000000002310 T pw_cuda_sf_z_ [...] If they are not there (after the rebuild), please also attach the complete output of this command. Best regards, Tiziano On 27.07.2017 08:22, hqzhou wrote: > Hi, > > I'm compiling cp2k 4.1 with CUDA 8.0 using gnu compiler gfortran 4.8.5 > on a CentOS 7.3 machine. > For the compilation going on smoothly, I have made some modifications to > the relevant source > code: > 1. change syncthreads() to __syncthreads() > 2. added condition to support native atomicAdd() in CUDA 8.0 > 3. change FFT_ALIGNMENT to CUFFT_COMPATIBILITY_FFTW_PADDING > > The compilation went almost OK, except the last step of creating > cp2k.ssmp. I got following error. > Actually, all undefined references are defined in pw/pw_cuda.F, am I > wrong somewhere in my > configuration (refer to the attached)? > > Thanks in advance! > > PS: I can run dbcsr_unittest.ssmp, libcusmm_unittest.ssmp without any > problem. > > Huiqun Zhou > @Earth Sciences, Nanjing University, China > > ======================= snippet of screen output > ========================================== > gfortran -I/usr/local/include -I/usr/local/cuda/include > -I/usr/local/libint/include -I/usr/local/libxc/include -D__GRID_CORE=2 > -D__PILAENV_BLOCKSIZE=1024 -D__ACC -D__DBCSR_ACC -D__PW_CUDA -D__LIBXC2 > -D__LIBINT -D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5 -D__MAX_CONTR=4 > -O2 -ffast-math -ffree-form -ffree-line-length-none -fopenmp > -ftree-vectorize -funroll-loops -mtune=native -I/usr/local/include > -I/usr/local/cuda/include -I/usr/local/libint/include > -I/usr/local/libxc/include -D__COMPILE_ARCH="\"unclecuda-gnu\"" > -D__COMPILE_DATE="\"Thu Jul 27 14:08:24 CST 2017\"" > -D__COMPILE_HOST="\"unclecuda\"" -D__COMPILE_REVISION="\"svn:17462\"" > -D__DATA_DIR="\"/home/zhou_huiqun/Cookplate/cp2k-4.1/data\"" > -L/home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp -o > /home/zhou_huiqun/Cookplate/cp2k-4.1/exe/unclecuda-gnu/cp2k.ssmp cp2k.o > -lcp2kstart -lcp2kmc -lcp2kswarm -lcp2kmotion -lcp2kthermostat -lcp2kemd > -lcp2ktmc -lcp2kmain -lcp2keri_mme -lcp2kminimax -lcp2khfxbase > -lcp2ksubsys -lcp2kxc -lcp2kao -lcp2kinput -lcp2kpw -lcp2kpwcuda > -lcp2kfft -lcp2kdbcsrwrap -ldbcsr -ldbcsrarnoldi -ldbcsrvec -ldbcsrmm > -ldbcsrops -ldbcsrbblock -ldbcsrbdist -ldbcsrbase -ldbcsrdata -ldbcsrerr > -lmicsmm -lclsmm -lcusmm -lcp2kacc -lcp2kaccmic -lcp2kaccopencl > -lcp2kacccuda -lcp2kfm -lcp2kcommon -lcp2kmpiwrap -lcp2kgrid -lcp2kbase > -L/usr/local/cuda/lib64 -lcudart -lcublas -lcufft -lrt > /usr/lib64/libopenblas.a /usr/local/libxc/lib/libxcf90.a > /usr/local/libxc/lib/libxc.a /usr/local/libint/lib/libderiv.a > /usr/local/libint/lib/libint.a > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_f': > pw_cuda.F:(.text+0x6c): undefined reference to `pw_cuda_f_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_sf': > pw_cuda.F:(.text+0x1ef): undefined reference to `pw_cuda_sf_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_fg': > pw_cuda.F:(.text+0x36e): undefined reference to `pw_cuda_fg_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d_ps': > pw_cuda.F:(.text+0x3673): undefined reference to `pw_cuda_fc_z_' > pw_cuda.F:(.text+0x3b6e): undefined reference to `pw_cuda_ffc_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d_ps': > pw_cuda.F:(.text+0x6ca5): undefined reference to `pw_cuda_cf_z_' > pw_cuda.F:(.text+0x7336): undefined reference to `pw_cuda_cff_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d': > pw_cuda.F:(.text+0x779c): undefined reference to `pw_cuda_sfffc_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d': > pw_cuda.F:(.text+0x79d0): undefined reference to `pw_cuda_cfffg_z_' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_init': > pw_cuda.F:(.text+0x7a55): undefined reference to `pw_cuda_init' > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > In function `__pw_cuda_MOD_pw_cuda_finalize': > pw_cuda.F:(.text+0x7a41): undefined reference to `pw_cuda_finalize' > collect2: error: ld returned 1 exit status > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com > . > To post to this group, send email to cp... at googlegroups.com > . > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. -- Tiziano M?ller University of Zurich Department of Chemistry Winterthurerstrasse 190 CH-8057 Z?rich Tel: +41 44 63 54234 www.chem.uzh.ch tiziano... at chem.uzh.ch -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. -------------- next part -------------- An HTML attachment was scrubbed... URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: compile-gnu.log Type: application/octet-stream Size: 901981 bytes Desc: not available URL: -------------- next part -------------- A non-text attachment was scrubbed... Name: nm-libcp2kpwcuda.a.out Type: application/octet-stream Size: 3776 bytes Desc: not available URL: From tiziano... at chem.uzh.ch Fri Jul 28 08:16:48 2017 From: tiziano... at chem.uzh.ch (=?UTF-8?Q?Tiziano_M=c3=bcller?=) Date: Fri, 28 Jul 2017 10:16:48 +0200 Subject: [CP2K:9266] Unddefined reference when compiling cp2k 4.1 withCUDA 8.0 In-Reply-To: References: <8b269b70-05d0-426a-a32a-8bf778e7792b@chem.uzh.ch> Message-ID: <52541643-c635-4ab1-b68d-ac3b73daf415@chem.uzh.ch> Hi Huiqun It seems you forgot to set NVFLAGS in your arch file which is why pw_cuda_z.o in libcp2kpwcuda.a is basically empty (specifically: __PW_CUDA is defined for gfortran, but undefined for nvcc). Usually you want to pass along all the preprocessor flags to nvcc plus optimization and the -arch flag. Take a look at the CRAY-XC30-gfortran-cuda.psmp for an example . Since you are using CUDA 8 you may also have to apply https://github.com/cp2k/cp2k/commit/e5a324e424bd79e0b2f6b87a00adf3d68eeba746 since that was fixed after the 4.1 release. Best regards, Tiziano On 27.07.2017 17:06, hqzhou wrote: > Hi, Tiziano > > Thank you for your quick response! Here are attached the compile log > and output of > nm libcp2kpwcuda.a. It seems that no such references exist in > libcp2kpwcuda.a. > > Thanks for resolving the problem! > > Huiqun > ------------------ Original ------------------ > *From: * "Tiziano M?ller"; > *Date: * Thu, Jul 27, 2017 05:10 PM > *To: * "hqzhou"; > *Cc: * "cp2k"; > *Subject: * Re: [CP2K:9263] Unddefined reference when compiling cp2k 4.1 > withCUDA 8.0 > Hi Huiqun > > Can you please attach the complete build log? > It might also be a good idea to do a `make distclean` and rebuild > everything from scratch for that. > > The mentioned symbols should be defined in > > lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a > > You can check for their presence using > `nm lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a` > > which should give something like this: > > [...] > pw_cuda_z.o: > [...] > 00000000000003a0 T pw_cuda_cfffg_z_ > 0000000000000f60 T pw_cuda_cff_z_ > 0000000000001680 T pw_cuda_cf_z_ > 00000000000019e0 T pw_cuda_fc_z_ > 00000000000012f0 T pw_cuda_ffc_z_ > 0000000000001f00 T pw_cuda_fg_z_ > 0000000000001d40 T pw_cuda_f_z_ > 0000000000000960 T pw_cuda_sfffc_z_ > 0000000000002310 T pw_cuda_sf_z_ > [...] > > If they are not there (after the rebuild), please also attach the > complete output of this command. > > Best regards, > Tiziano > > On 27.07.2017 08:22, hqzhou wrote: > > Hi, > > > > I'm compiling cp2k 4.1 with CUDA 8.0 using gnu compiler gfortran 4.8.5 > > on a CentOS 7.3 machine. > > For the compilation going on smoothly, I have made some modifications to > > the relevant source > > code: > > 1. change syncthreads() to __syncthreads() > > 2. added condition to support native atomicAdd() in CUDA 8.0 > > 3. change FFT_ALIGNMENT to CUFFT_COMPATIBILITY_FFTW_PADDING > > > > The compilation went almost OK, except the last step of creating > > cp2k.ssmp. I got following error. > > Actually, all undefined references are defined in pw/pw_cuda.F, am I > > wrong somewhere in my > > configuration (refer to the attached)? > > > > Thanks in advance! > > > > PS: I can run dbcsr_unittest.ssmp, libcusmm_unittest.ssmp without any > > problem. > > > > Huiqun Zhou > > @Earth Sciences, Nanjing University, China > > > > ======================= snippet of screen output > > ========================================== > > gfortran -I/usr/local/include -I/usr/local/cuda/include > > -I/usr/local/libint/include -I/usr/local/libxc/include -D__GRID_CORE=2 > > -D__PILAENV_BLOCKSIZE=1024 -D__ACC -D__DBCSR_ACC -D__PW_CUDA -D__LIBXC2 > > -D__LIBINT -D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5 -D__MAX_CONTR=4 > > -O2 -ffast-math -ffree-form -ffree-line-length-none -fopenmp > > -ftree-vectorize -funroll-loops -mtune=native -I/usr/local/include > > -I/usr/local/cuda/include -I/usr/local/libint/include > > -I/usr/local/libxc/include -D__COMPILE_ARCH="\"unclecuda-gnu\"" > > -D__COMPILE_DATE="\"Thu Jul 27 14:08:24 CST 2017\"" > > -D__COMPILE_HOST="\"unclecuda\"" -D__COMPILE_REVISION="\"svn:17462\"" > > -D__DATA_DIR="\"/home/zhou_huiqun/Cookplate/cp2k-4.1/data\"" > > -L/home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp -o > > /home/zhou_huiqun/Cookplate/cp2k-4.1/exe/unclecuda-gnu/cp2k.ssmp cp2k.o > > -lcp2kstart -lcp2kmc -lcp2kswarm -lcp2kmotion -lcp2kthermostat -lcp2kemd > > -lcp2ktmc -lcp2kmain -lcp2keri_mme -lcp2kminimax -lcp2khfxbase > > -lcp2ksubsys -lcp2kxc -lcp2kao -lcp2kinput -lcp2kpw -lcp2kpwcuda > > -lcp2kfft -lcp2kdbcsrwrap -ldbcsr -ldbcsrarnoldi -ldbcsrvec -ldbcsrmm > > -ldbcsrops -ldbcsrbblock -ldbcsrbdist -ldbcsrbase -ldbcsrdata -ldbcsrerr > > -lmicsmm -lclsmm -lcusmm -lcp2kacc -lcp2kaccmic -lcp2kaccopencl > > -lcp2kacccuda -lcp2kfm -lcp2kcommon -lcp2kmpiwrap -lcp2kgrid -lcp2kbase > > -L/usr/local/cuda/lib64 -lcudart -lcublas -lcufft -lrt > > /usr/lib64/libopenblas.a /usr/local/libxc/lib/libxcf90.a > > /usr/local/libxc/lib/libxc.a /usr/local/libint/lib/libderiv.a > > /usr/local/libint/lib/libint.a > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_f': > > pw_cuda.F:(.text+0x6c): undefined reference to `pw_cuda_f_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_sf': > > pw_cuda.F:(.text+0x1ef): undefined reference to `pw_cuda_sf_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_fg': > > pw_cuda.F:(.text+0x36e): undefined reference to `pw_cuda_fg_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d_ps': > > pw_cuda.F:(.text+0x3673): undefined reference to `pw_cuda_fc_z_' > > pw_cuda.F:(.text+0x3b6e): undefined reference to `pw_cuda_ffc_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d_ps': > > pw_cuda.F:(.text+0x6ca5): undefined reference to `pw_cuda_cf_z_' > > pw_cuda.F:(.text+0x7336): undefined reference to `pw_cuda_cff_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d': > > pw_cuda.F:(.text+0x779c): undefined reference to `pw_cuda_sfffc_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d': > > pw_cuda.F:(.text+0x79d0): undefined reference to `pw_cuda_cfffg_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_init': > > pw_cuda.F:(.text+0x7a55): undefined reference to `pw_cuda_init' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_finalize': > > pw_cuda.F:(.text+0x7a41): undefined reference to `pw_cuda_finalize' > > collect2: error: ld returned 1 exit status > > > > -- > > You received this message because you are subscribed to the Google > > Groups "cp2k" group. > > To unsubscribe from this group and stop receiving emails from it, send > > an email to cp2k+uns... at googlegroups.com > > . > > To post to this group, send email to cp... at googlegroups.com > > . > > Visit this group at https://groups.google.com/group/cp2k. > > For more options, visit https://groups.google.com/d/optout. > > -- > Tiziano M?ller > University of Zurich > Department of Chemistry > Winterthurerstrasse 190 > CH-8057 Z?rich > > Tel: +41 44 63 54234 > www.chem.uzh.ch > tiziano... at chem.uzh.ch > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com. > To post to this group, send email to cp... at googlegroups.com. > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com > . > To post to this group, send email to cp... at googlegroups.com > . > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. -- Tiziano M?ller University of Zurich Department of Chemistry Winterthurerstrasse 190 CH-8057 Z?rich Tel: +41 44 63 54234 www.chem.uzh.ch tiziano... at chem.uzh.ch From hqz... at nju.edu.cn Sat Jul 29 10:51:56 2017 From: hqz... at nju.edu.cn (=?utf-8?B?aHF6aG91?=) Date: Sat, 29 Jul 2017 18:51:56 +0800 Subject: [CP2K:9267] Unddefined reference when compiling cp2k 4.1 withCUDA 8.0 In-Reply-To: <52541643-c635-4ab1-b68d-ac3b73daf415@chem.uzh.ch> References: <8b269b70-05d0-426a-a32a-8bf778e7792b@chem.uzh.ch> <52541643-c635-4ab1-b68d-ac3b73daf415@chem.uzh.ch> Message-ID: ------------------ Original ------------------ From: "Tiziano M?ller"; Date: Fri, Jul 28, 2017 04:16 PM To: "hqzhou"; Cc: "cp2k"; Subject: Re: [CP2K:9267] Unddefined reference when compiling cp2k 4.1 withCUDA 8.0 Hi Huiqun It seems you forgot to set NVFLAGS in your arch file which is why pw_cuda_z.o in libcp2kpwcuda.a is basically empty (specifically: __PW_CUDA is defined for gfortran, but undefined for nvcc). Usually you want to pass along all the preprocessor flags to nvcc plus optimization and the -arch flag. Take a look at the CRAY-XC30-gfortran-cuda.psmp for an example . Since you are using CUDA 8 you may also have to apply https://github.com/cp2k/cp2k/commit/e5a324e424bd79e0b2f6b87a00adf3d68eeba746 since that was fixed after the 4.1 release. Best regards, Tiziano On 27.07.2017 17:06, hqzhou wrote: > Hi, Tiziano > > Thank you for your quick response! Here are attached the compile log > and output of > nm libcp2kpwcuda.a. It seems that no such references exist in > libcp2kpwcuda.a. > > Thanks for resolving the problem! > > Huiqun > ------------------ Original ------------------ > *From: * "Tiziano M?ller"; > *Date: * Thu, Jul 27, 2017 05:10 PM > *To: * "hqzhou"; > *Cc: * "cp2k"; > *Subject: * Re: [CP2K:9263] Unddefined reference when compiling cp2k 4.1 > withCUDA 8.0 > Hi Huiqun > > Can you please attach the complete build log? > It might also be a good idea to do a `make distclean` and rebuild > everything from scratch for that. > > The mentioned symbols should be defined in > > lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a > > You can check for their presence using > `nm lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a` > > which should give something like this: > > [...] > pw_cuda_z.o: > [...] > 00000000000003a0 T pw_cuda_cfffg_z_ > 0000000000000f60 T pw_cuda_cff_z_ > 0000000000001680 T pw_cuda_cf_z_ > 00000000000019e0 T pw_cuda_fc_z_ > 00000000000012f0 T pw_cuda_ffc_z_ > 0000000000001f00 T pw_cuda_fg_z_ > 0000000000001d40 T pw_cuda_f_z_ > 0000000000000960 T pw_cuda_sfffc_z_ > 0000000000002310 T pw_cuda_sf_z_ > [...] > > If they are not there (after the rebuild), please also attach the > complete output of this command. > > Best regards, > Tiziano > > On 27.07.2017 08:22, hqzhou wrote: > > Hi, > > > > I'm compiling cp2k 4.1 with CUDA 8.0 using gnu compiler gfortran 4.8.5 > > on a CentOS 7.3 machine. > > For the compilation going on smoothly, I have made some modifications to > > the relevant source > > code: > > 1. change syncthreads() to __syncthreads() > > 2. added condition to support native atomicAdd() in CUDA 8.0 > > 3. change FFT_ALIGNMENT to CUFFT_COMPATIBILITY_FFTW_PADDING > > > > The compilation went almost OK, except the last step of creating > > cp2k.ssmp. I got following error. > > Actually, all undefined references are defined in pw/pw_cuda.F, am I > > wrong somewhere in my > > configuration (refer to the attached)? > > > > Thanks in advance! > > > > PS: I can run dbcsr_unittest.ssmp, libcusmm_unittest.ssmp without any > > problem. > > > > Huiqun Zhou > > @Earth Sciences, Nanjing University, China > > > > ======================= snippet of screen output > > ========================================== > > gfortran -I/usr/local/include -I/usr/local/cuda/include > > -I/usr/local/libint/include -I/usr/local/libxc/include -D__GRID_CORE=2 > > -D__PILAENV_BLOCKSIZE=1024 -D__ACC -D__DBCSR_ACC -D__PW_CUDA -D__LIBXC2 > > -D__LIBINT -D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5 -D__MAX_CONTR=4 > > -O2 -ffast-math -ffree-form -ffree-line-length-none -fopenmp > > -ftree-vectorize -funroll-loops -mtune=native -I/usr/local/include > > -I/usr/local/cuda/include -I/usr/local/libint/include > > -I/usr/local/libxc/include -D__COMPILE_ARCH="\"unclecuda-gnu\"" > > -D__COMPILE_DATE="\"Thu Jul 27 14:08:24 CST 2017\"" > > -D__COMPILE_HOST="\"unclecuda\"" -D__COMPILE_REVISION="\"svn:17462\"" > > -D__DATA_DIR="\"/home/zhou_huiqun/Cookplate/cp2k-4.1/data\"" > > -L/home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp -o > > /home/zhou_huiqun/Cookplate/cp2k-4.1/exe/unclecuda-gnu/cp2k.ssmp cp2k.o > > -lcp2kstart -lcp2kmc -lcp2kswarm -lcp2kmotion -lcp2kthermostat -lcp2kemd > > -lcp2ktmc -lcp2kmain -lcp2keri_mme -lcp2kminimax -lcp2khfxbase > > -lcp2ksubsys -lcp2kxc -lcp2kao -lcp2kinput -lcp2kpw -lcp2kpwcuda > > -lcp2kfft -lcp2kdbcsrwrap -ldbcsr -ldbcsrarnoldi -ldbcsrvec -ldbcsrmm > > -ldbcsrops -ldbcsrbblock -ldbcsrbdist -ldbcsrbase -ldbcsrdata -ldbcsrerr > > -lmicsmm -lclsmm -lcusmm -lcp2kacc -lcp2kaccmic -lcp2kaccopencl > > -lcp2kacccuda -lcp2kfm -lcp2kcommon -lcp2kmpiwrap -lcp2kgrid -lcp2kbase > > -L/usr/local/cuda/lib64 -lcudart -lcublas -lcufft -lrt > > /usr/lib64/libopenblas.a /usr/local/libxc/lib/libxcf90.a > > /usr/local/libxc/lib/libxc.a /usr/local/libint/lib/libderiv.a > > /usr/local/libint/lib/libint.a > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_f': > > pw_cuda.F:(.text+0x6c): undefined reference to `pw_cuda_f_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_sf': > > pw_cuda.F:(.text+0x1ef): undefined reference to `pw_cuda_sf_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_fg': > > pw_cuda.F:(.text+0x36e): undefined reference to `pw_cuda_fg_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d_ps': > > pw_cuda.F:(.text+0x3673): undefined reference to `pw_cuda_fc_z_' > > pw_cuda.F:(.text+0x3b6e): undefined reference to `pw_cuda_ffc_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d_ps': > > pw_cuda.F:(.text+0x6ca5): undefined reference to `pw_cuda_cf_z_' > > pw_cuda.F:(.text+0x7336): undefined reference to `pw_cuda_cff_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d': > > pw_cuda.F:(.text+0x779c): undefined reference to `pw_cuda_sfffc_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d': > > pw_cuda.F:(.text+0x79d0): undefined reference to `pw_cuda_cfffg_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_init': > > pw_cuda.F:(.text+0x7a55): undefined reference to `pw_cuda_init' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_finalize': > > pw_cuda.F:(.text+0x7a41): undefined reference to `pw_cuda_finalize' > > collect2: error: ld returned 1 exit status > > > > -- > > You received this message because you are subscribed to the Google > > Groups "cp2k" group. > > To unsubscribe from this group and stop receiving emails from it, send > > an email to cp2k+uns... at googlegroups.com > > . > > To post to this group, send email to cp... at googlegroups.com > > . > > Visit this group at https://groups.google.com/group/cp2k. > > For more options, visit https://groups.google.com/d/optout. > > -- > Tiziano M?ller > University of Zurich > Department of Chemistry > Winterthurerstrasse 190 > CH-8057 Z?rich > > Tel: +41 44 63 54234 > www.chem.uzh.ch > tiziano... at chem.uzh.ch > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com. > To post to this group, send email to cp... at googlegroups.com. > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com > . > To post to this group, send email to cp... at googlegroups.com > . > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. -- Tiziano M?ller University of Zurich Department of Chemistry Winterthurerstrasse 190 CH-8057 Z?rich Tel: +41 44 63 54234 www.chem.uzh.ch tiziano... at chem.uzh.ch -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. -------------- next part -------------- An HTML attachment was scrubbed... URL: From hqz... at nju.edu.cn Sat Jul 29 10:53:51 2017 From: hqz... at nju.edu.cn (=?utf-8?B?aHF6aG91?=) Date: Sat, 29 Jul 2017 18:53:51 +0800 Subject: [CP2K:9267] Unddefined reference when compiling cp2k 4.1 withCUDA 8.0 In-Reply-To: <52541643-c635-4ab1-b68d-ac3b73daf415@chem.uzh.ch> References: <8b269b70-05d0-426a-a32a-8bf778e7792b@chem.uzh.ch> <52541643-c635-4ab1-b68d-ac3b73daf415@chem.uzh.ch> Message-ID: Tiziano, It works now, thank you very much! Huiqun ------------------ Original ------------------ From: "Tiziano M?ller"; Date: Fri, Jul 28, 2017 04:16 PM To: "hqzhou"; Cc: "cp2k"; Subject: Re: [CP2K:9267] Unddefined reference when compiling cp2k 4.1 withCUDA 8.0 Hi Huiqun It seems you forgot to set NVFLAGS in your arch file which is why pw_cuda_z.o in libcp2kpwcuda.a is basically empty (specifically: __PW_CUDA is defined for gfortran, but undefined for nvcc). Usually you want to pass along all the preprocessor flags to nvcc plus optimization and the -arch flag. Take a look at the CRAY-XC30-gfortran-cuda.psmp for an example . Since you are using CUDA 8 you may also have to apply https://github.com/cp2k/cp2k/commit/e5a324e424bd79e0b2f6b87a00adf3d68eeba746 since that was fixed after the 4.1 release. Best regards, Tiziano On 27.07.2017 17:06, hqzhou wrote: > Hi, Tiziano > > Thank you for your quick response! Here are attached the compile log > and output of > nm libcp2kpwcuda.a. It seems that no such references exist in > libcp2kpwcuda.a. > > Thanks for resolving the problem! > > Huiqun > ------------------ Original ------------------ > *From: * "Tiziano M?ller"; > *Date: * Thu, Jul 27, 2017 05:10 PM > *To: * "hqzhou"; > *Cc: * "cp2k"; > *Subject: * Re: [CP2K:9263] Unddefined reference when compiling cp2k 4.1 > withCUDA 8.0 > Hi Huiqun > > Can you please attach the complete build log? > It might also be a good idea to do a `make distclean` and rebuild > everything from scratch for that. > > The mentioned symbols should be defined in > > lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a > > You can check for their presence using > `nm lib/unclecuda-gnu/ssmp/libcp2kpwcuda.a` > > which should give something like this: > > [...] > pw_cuda_z.o: > [...] > 00000000000003a0 T pw_cuda_cfffg_z_ > 0000000000000f60 T pw_cuda_cff_z_ > 0000000000001680 T pw_cuda_cf_z_ > 00000000000019e0 T pw_cuda_fc_z_ > 00000000000012f0 T pw_cuda_ffc_z_ > 0000000000001f00 T pw_cuda_fg_z_ > 0000000000001d40 T pw_cuda_f_z_ > 0000000000000960 T pw_cuda_sfffc_z_ > 0000000000002310 T pw_cuda_sf_z_ > [...] > > If they are not there (after the rebuild), please also attach the > complete output of this command. > > Best regards, > Tiziano > > On 27.07.2017 08:22, hqzhou wrote: > > Hi, > > > > I'm compiling cp2k 4.1 with CUDA 8.0 using gnu compiler gfortran 4.8.5 > > on a CentOS 7.3 machine. > > For the compilation going on smoothly, I have made some modifications to > > the relevant source > > code: > > 1. change syncthreads() to __syncthreads() > > 2. added condition to support native atomicAdd() in CUDA 8.0 > > 3. change FFT_ALIGNMENT to CUFFT_COMPATIBILITY_FFTW_PADDING > > > > The compilation went almost OK, except the last step of creating > > cp2k.ssmp. I got following error. > > Actually, all undefined references are defined in pw/pw_cuda.F, am I > > wrong somewhere in my > > configuration (refer to the attached)? > > > > Thanks in advance! > > > > PS: I can run dbcsr_unittest.ssmp, libcusmm_unittest.ssmp without any > > problem. > > > > Huiqun Zhou > > @Earth Sciences, Nanjing University, China > > > > ======================= snippet of screen output > > ========================================== > > gfortran -I/usr/local/include -I/usr/local/cuda/include > > -I/usr/local/libint/include -I/usr/local/libxc/include -D__GRID_CORE=2 > > -D__PILAENV_BLOCKSIZE=1024 -D__ACC -D__DBCSR_ACC -D__PW_CUDA -D__LIBXC2 > > -D__LIBINT -D__LIBINT_MAX_AM=6 -D__LIBDERIV_MAX_AM1=5 -D__MAX_CONTR=4 > > -O2 -ffast-math -ffree-form -ffree-line-length-none -fopenmp > > -ftree-vectorize -funroll-loops -mtune=native -I/usr/local/include > > -I/usr/local/cuda/include -I/usr/local/libint/include > > -I/usr/local/libxc/include -D__COMPILE_ARCH="\"unclecuda-gnu\"" > > -D__COMPILE_DATE="\"Thu Jul 27 14:08:24 CST 2017\"" > > -D__COMPILE_HOST="\"unclecuda\"" -D__COMPILE_REVISION="\"svn:17462\"" > > -D__DATA_DIR="\"/home/zhou_huiqun/Cookplate/cp2k-4.1/data\"" > > -L/home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp -o > > /home/zhou_huiqun/Cookplate/cp2k-4.1/exe/unclecuda-gnu/cp2k.ssmp cp2k.o > > -lcp2kstart -lcp2kmc -lcp2kswarm -lcp2kmotion -lcp2kthermostat -lcp2kemd > > -lcp2ktmc -lcp2kmain -lcp2keri_mme -lcp2kminimax -lcp2khfxbase > > -lcp2ksubsys -lcp2kxc -lcp2kao -lcp2kinput -lcp2kpw -lcp2kpwcuda > > -lcp2kfft -lcp2kdbcsrwrap -ldbcsr -ldbcsrarnoldi -ldbcsrvec -ldbcsrmm > > -ldbcsrops -ldbcsrbblock -ldbcsrbdist -ldbcsrbase -ldbcsrdata -ldbcsrerr > > -lmicsmm -lclsmm -lcusmm -lcp2kacc -lcp2kaccmic -lcp2kaccopencl > > -lcp2kacccuda -lcp2kfm -lcp2kcommon -lcp2kmpiwrap -lcp2kgrid -lcp2kbase > > -L/usr/local/cuda/lib64 -lcudart -lcublas -lcufft -lrt > > /usr/lib64/libopenblas.a /usr/local/libxc/lib/libxcf90.a > > /usr/local/libxc/lib/libxc.a /usr/local/libint/lib/libderiv.a > > /usr/local/libint/lib/libint.a > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_f': > > pw_cuda.F:(.text+0x6c): undefined reference to `pw_cuda_f_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_sf': > > pw_cuda.F:(.text+0x1ef): undefined reference to `pw_cuda_sf_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_fg': > > pw_cuda.F:(.text+0x36e): undefined reference to `pw_cuda_fg_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d_ps': > > pw_cuda.F:(.text+0x3673): undefined reference to `pw_cuda_fc_z_' > > pw_cuda.F:(.text+0x3b6e): undefined reference to `pw_cuda_ffc_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d_ps': > > pw_cuda.F:(.text+0x6ca5): undefined reference to `pw_cuda_cf_z_' > > pw_cuda.F:(.text+0x7336): undefined reference to `pw_cuda_cff_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_c1dr3d_3d': > > pw_cuda.F:(.text+0x779c): undefined reference to `pw_cuda_sfffc_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_r3dc1d_3d': > > pw_cuda.F:(.text+0x79d0): undefined reference to `pw_cuda_cfffg_z_' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_init': > > pw_cuda.F:(.text+0x7a55): undefined reference to `pw_cuda_init' > > > /home/zhou_huiqun/Cookplate/cp2k-4.1/lib/unclecuda-gnu/ssmp/libcp2kpw.a(pw_cuda.o): > > > In function `__pw_cuda_MOD_pw_cuda_finalize': > > pw_cuda.F:(.text+0x7a41): undefined reference to `pw_cuda_finalize' > > collect2: error: ld returned 1 exit status > > > > -- > > You received this message because you are subscribed to the Google > > Groups "cp2k" group. > > To unsubscribe from this group and stop receiving emails from it, send > > an email to cp2k+uns... at googlegroups.com > > . > > To post to this group, send email to cp... at googlegroups.com > > . > > Visit this group at https://groups.google.com/group/cp2k. > > For more options, visit https://groups.google.com/d/optout. > > -- > Tiziano M?ller > University of Zurich > Department of Chemistry > Winterthurerstrasse 190 > CH-8057 Z?rich > > Tel: +41 44 63 54234 > www.chem.uzh.ch > tiziano... at chem.uzh.ch > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com. > To post to this group, send email to cp... at googlegroups.com. > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. > > -- > You received this message because you are subscribed to the Google > Groups "cp2k" group. > To unsubscribe from this group and stop receiving emails from it, send > an email to cp2k+uns... at googlegroups.com > . > To post to this group, send email to cp... at googlegroups.com > . > Visit this group at https://groups.google.com/group/cp2k. > For more options, visit https://groups.google.com/d/optout. -- Tiziano M?ller University of Zurich Department of Chemistry Winterthurerstrasse 190 CH-8057 Z?rich Tel: +41 44 63 54234 www.chem.uzh.ch tiziano... at chem.uzh.ch -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. -------------- next part -------------- An HTML attachment was scrubbed... URL: From hqz... at nju.edu.cn Sat Jul 29 11:12:28 2017 From: hqz... at nju.edu.cn (=?utf-8?B?aHF6aG91?=) Date: Sat, 29 Jul 2017 19:12:28 +0800 Subject: libcusmm, libsmm, libxsmm: mutual exclusive or not? Message-ID: Hi, libcusmm, libsmm and libxsmm are all used for small matrix multiplication, I wonder whether it is sufficient to include only the GPU version, libcusmm, on my desktop computer with a Tesla K40 equipped. I have tried to include libsmm_dnn and it seems OK, but I don't know if it is meaningful. Huiqun Zhou @Earth Sciences, Nanjing University -------------- next part -------------- An HTML attachment was scrubbed... URL: From alfio.... at gmail.com Sun Jul 30 06:55:16 2017 From: alfio.... at gmail.com (Alfio Lazzaro) Date: Sat, 29 Jul 2017 23:55:16 -0700 (PDT) Subject: libcusmm, libsmm, libxsmm: mutual exclusive or not? In-Reply-To: References: Message-ID: <8745a653-9ec5-4483-a362-f3c41df25aa7@googlegroups.com> Hi Huiqun Zhou, libcusmm + libsmm_dnn is OK. Alfio Il giorno sabato 29 luglio 2017 13:12:48 UTC+2, Huiqun Zhou ha scritto: > > Hi, > > libcusmm, libsmm and libxsmm are all used for small matrix multiplication, > I wonder whether it is sufficient to include only the GPU version, > libcusmm, on my > desktop computer with a Tesla K40 equipped. I have tried to include > libsmm_dnn > and it seems OK, but I don't know if it is meaningful. > > Huiqun Zhou > @Earth Sciences, Nanjing University > -------------- next part -------------- An HTML attachment was scrubbed... URL: From natalie... at gmail.com Mon Jul 31 00:51:47 2017 From: natalie... at gmail.com (Natalie Austin) Date: Sun, 30 Jul 2017 17:51:47 -0700 (PDT) Subject: Solvation model SCCS convergence issues In-Reply-To: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> References: <884f0531-de9a-4ec4-8613-fe132f95d0ab@googlegroups.com> Message-ID: <2330f888-8efc-4211-bd9e-2837bca3b9ea@googlegroups.com> My system is non-periodic which is why I used that solver. But I will test the solvation calculations with periodicity turned on. Thanks, Natalie -------------- next part -------------- An HTML attachment was scrubbed... URL: From hqz... at nju.edu.cn Mon Jul 31 06:45:31 2017 From: hqz... at nju.edu.cn (=?utf-8?B?aHF6aG91?=) Date: Mon, 31 Jul 2017 14:45:31 +0800 Subject: [CP2K:9272] Re: libcusmm, libsmm, libxsmm: mutual exclusive or not? In-Reply-To: <8745a653-9ec5-4483-a362-f3c41df25aa7@googlegroups.com> References: <8745a653-9ec5-4483-a362-f3c41df25aa7@googlegroups.com> Message-ID: Hi, Alfio Thank you very much! I also made clear the internal mechanism, they have fallback relationship. Huiqun ------------------ Original ------------------ From: "Alfio Lazzaro"; Date: Sun, Jul 30, 2017 02:55 PM To: "cp2k"; Subject: [CP2K:9272] Re: libcusmm, libsmm, libxsmm: mutual exclusive or not? Hi Huiqun Zhou, libcusmm + libsmm_dnn is OK. Alfio Il giorno sabato 29 luglio 2017 13:12:48 UTC+2, Huiqun Zhou ha scritto:Hi, libcusmm, libsmm and libxsmm are all used for small matrix multiplication, I wonder whether it is sufficient to include only the GPU version, libcusmm, on my desktop computer with a Tesla K40 equipped. I have tried to include libsmm_dnn and it seems OK, but I don't know if it is meaningful. Huiqun Zhou @Earth Sciences, Nanjing University -- You received this message because you are subscribed to the Google Groups "cp2k" group. To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com. To post to this group, send email to cp... at googlegroups.com. Visit this group at https://groups.google.com/group/cp2k. For more options, visit https://groups.google.com/d/optout. -------------- next part -------------- An HTML attachment was scrubbed... URL: From rybk... at gmail.com Mon Jul 31 11:39:43 2017 From: rybk... at gmail.com (Vladimir Rybkin) Date: Mon, 31 Jul 2017 04:39:43 -0700 (PDT) Subject: Basic tutorial for all-electron calculations (HF, MP2, ...) In-Reply-To: <9ef46dff-8f71-49b0-8bdd-1ad2470d6f30@googlegroups.com> References: <9ef46dff-8f71-49b0-8bdd-1ad2470d6f30@googlegroups.com> Message-ID: Dear Christian, I have added my MP2/RPA tutorial, which includes input files on the page of the passed User Tutorial 2017. https://www.cp2k.org/events:2017_user_tutorial:about#th_july_9am_-_5pm You can find more examples - HF and MP2/RPA - in the exercises: https://www.cp2k.org/exercises:2017_uzh_cp2k-tutorial:index Yours, Vladimir ???????, 22 ???? 2017 ?., 1:42:33 UTC+2 ???????????? Christian Leitold ???????: > > Hi, > > I am new in the field of electronic calculations, and I am looking for > some basic tutorials. In particular, right now I am interested in doing HF > and MP2 calculations for simple molecules in the gas phase, no PBCs, and no > DFT. The tutorials I could find so far look really good, but I could only > find them for DFT calculations. Could you maybe point me to some examples > of Hartree-Fock-only calculations? > > Thanks, > Christian > -------------- next part -------------- An HTML attachment was scrubbed... URL: From christia... at gmail.com Mon Jul 31 17:45:33 2017 From: christia... at gmail.com (Christian Leitold) Date: Mon, 31 Jul 2017 10:45:33 -0700 (PDT) Subject: Basic tutorial for all-electron calculations (HF, MP2, ...) In-Reply-To: References: <9ef46dff-8f71-49b0-8bdd-1ad2470d6f30@googlegroups.com> Message-ID: <96007229-1f4d-4715-a62d-79196efe72e5@googlegroups.com> Hi Vladimir, Thanks a lot, I will definitely have a look at these documents! Best, Christian On Monday, July 31, 2017 at 4:39:43 AM UTC-7, Vladimir Rybkin wrote: > > Dear Christian, > > I have added my MP2/RPA tutorial, which includes input files on the page > of the passed User Tutorial 2017. > https://www.cp2k.org/events:2017_user_tutorial:about#th_july_9am_-_5pm > > You can find more examples - HF and MP2/RPA - in the exercises: > https://www.cp2k.org/exercises:2017_uzh_cp2k-tutorial:index > > Yours, > > Vladimir > > > > ???????, 22 ???? 2017 ?., 1:42:33 UTC+2 ???????????? Christian Leitold > ???????: >> >> Hi, >> >> I am new in the field of electronic calculations, and I am looking for >> some basic tutorials. In particular, right now I am interested in doing HF >> and MP2 calculations for simple molecules in the gas phase, no PBCs, and no >> DFT. The tutorials I could find so far look really good, but I could only >> find them for DFT calculations. Could you maybe point me to some examples >> of Hartree-Fock-only calculations? >> >> Thanks, >> Christian >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: