Hi I tried to use the H2O-32_SPME.inp in the tests/Fist directory by changing only the length of the simulation (1 ns) and I got also a significant energy drift (see attachment). Then I tried to reduce the integration time step to 1.0 fs. The drift become lower but it do not vanish. I did also 1ns of NVE dynamics (starting from the last restart of the previous simulations) and here the drift become even more clear. The inputs that I used are below. Best regards Paolo NVT input: &FORCE_EVAL METHOD FIST &MM &FORCEFIELD &BEND ATOMS H O H K 0. THETA0 1.8 &END BEND &BOND ATOMS O H K 0. R0 1.8 &END BOND &CHARGE ATOM O CHARGE -0.8476 &END CHARGE &CHARGE ATOM H CHARGE 0.4238 &END CHARGE &NONBONDED &LENNARD-JONES atoms O O EPSILON 78.198 SIGMA 3.166 RCUT 11.4 &END LENNARD-JONES &LENNARD-JONES atoms O H EPSILON 0.0 SIGMA 3.6705 RCUT 11.4 &END LENNARD-JONES &LENNARD-JONES atoms H H EPSILON 0.0 SIGMA 3.30523 RCUT 11.4 &END LENNARD-JONES &END NONBONDED &END FORCEFIELD &POISSON &EWALD EWALD_TYPE spme ALPHA .5 GMAX 21 O_SPLINE 6 &END EWALD &END POISSON &END MM &SUBSYS &CELL ABC 9.865 9.865 9.865 &END CELL &COORD O -4.583 5.333 1.560 H2O H -3.777 5.331 0.943 H2O H -5.081 4.589 1.176 H2O ... &END COORD &END SUBSYS &END FORCE_EVAL &GLOBAL PROJECT H2O RUN_TYPE MD &END GLOBAL &MOTION &CONSTRAINT &G3X3 DISTANCES 1.8897268 1.8897268 3.0859239 MOLECULE 1 ATOMS 1 2 3 &END G3X3 &END CONSTRAINT &MD ENSEMBLE NVT STEPS 1000000 TIMESTEP 1.0 TEMPERATURE 300.0 &THERMOSTAT REGION MOLECULE &NOSE LENGTH 3 YOSHIDA 3 TIMECON 1000 MTS 2 &END NOSE &END &END MD &END MOTION NVE input: &FORCE_EVAL METHOD FIST &MM &FORCEFIELD &BEND ATOMS H O H K 0. THETA0 1.8 &END BEND &BOND ATOMS O H K 0. R0 1.8 &END BOND &CHARGE ATOM O CHARGE -0.8476 &END CHARGE &CHARGE ATOM H CHARGE 0.4238 &END CHARGE &NONBONDED &LENNARD-JONES atoms O O EPSILON 78.198 SIGMA 3.166 RCUT 11.4 &END LENNARD-JONES &LENNARD-JONES atoms O H EPSILON 0.0 SIGMA 3.6705 RCUT 11.4 &END LENNARD-JONES &LENNARD-JONES atoms H H EPSILON 0.0 SIGMA 3.30523 RCUT 11.4 &END LENNARD-JONES &END NONBONDED &END FORCEFIELD &POISSON &EWALD EWALD_TYPE spme ALPHA .5 GMAX 21 O_SPLINE 6 &END EWALD &END POISSON &END MM &END FORCE_EVAL &GLOBAL PROJECT H2O RUN_TYPE MD &END GLOBAL &MOTION &CONSTRAINT &G3X3 DISTANCES 1.8897268 1.8897268 3.0859239 MOLECULE 1 ATOMS 1 2 3 &END G3X3 &END CONSTRAINT &MD ENSEMBLE NVE STEPS 1000000 TIMESTEP 1.0 TEMPERATURE 300.0 &END MD &END MOTION &EXT_RESTART RESTART_FILE_NAME H2O-1_1000000.restart &END EXT_RESTART On Monday, September 17, 2012 1:09:54 PM UTC+2, jgh wrote:<blockquote class="gmail_quote" style="margin: 0;margin-left: 0.8ex;border-left: 1px #ccc solid;padding-left: 1ex;">Hi I did some tests with the inputs from ../tests/Fist and get much better energy conservation that you show. Probably your input use a not optimal setting for the cutoffs and Ewald parameters (I was using SPME). regards Juerg -------------------------------------------------------------- Juerg Hutter Phone : ++41 44 635 4491 Physical Chemistry Institute FAX : ++41 44 635 6838 University of Zurich E-mail: <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BckEGd5oO7QJ">hut...@pci.uzh.ch</a> Winterthurerstrasse 190 CH-8057 Zurich, Switzerland --------------------------------------------------------------- -----<a href="javascript:" target="_blank" gdf-obfuscated-mailto="BckEGd5oO7QJ">cp...@googlegroups.com</a> wrote: ----- To: <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BckEGd5oO7QJ">cp...@googlegroups.com</a> From: marco masia Sent by: <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BckEGd5oO7QJ">cp...@googlegroups.com</a> Date: 09/12/2012 03:39PM Subject: [CP2K:4029] Re: Energy conservation NVT/NVE simulation Dear all, recently I have noticed the same problem with FIST simulations. I took the H2O-32_ewald.inp input, equilibrated it for 1 ns and run a 1ns NVE simulation from the thermalized configuration. I have tried to refine many simulation parameters (e.g. time step, RCUT, Ewald pars) and also (thinking that there was a bad implementation of constraints) I run the simulations for a flexible model. I always get a non-negligible drift (see attachment); the graph shows block averaged energies of 1 ns simulations performed with cp2k and obeli.x (a program of mine). The values are shifted for a better comparison. I have been looking into this, thinking that the problem was in the input file...but since other people are experiencing it with QS, I wonder if there is some bug in the velocity verlet routines. The results are platform independent (I tried both on mac and linux). best marco On Wednesday, September 5, 2012 3:32:26 AM UTC-4, <a>albe...@virgilio.it</a> wrote: Dear all, I'm carrying out some tests to reproduce the VDOS of water and I used as starting points the input files 'H2O-32.inp' and 'H2O-64.inp' in tests/QS/benchmark directory. In both cases I ran a NVT simulation at 300K for 15 ps followed by an NVE simulation on the equilibrated system. I'm puzzled about the results in the energies: in both NVT and NVE the plot of "Cons Qty" show a constant and linear drift while temperature, kinetic and potential energy are equilibrated and stable. Since "Cons Qty" should be conserved and this is not the case I think that there are some problems in the setup of the simulation. I did not modify the parameters of the benchmark inputs but I just added the thermostat section in NVT and I modified only the printing option. Thank you very much for your help and suggestions, Alberto -- You received this message because you are subscribed to the Google Groups "cp2k" group. To view this discussion on the web visit <a href="https://groups.google.com/d/msg/cp2k/-/sAiVZ2dob1sJ" target="_blank">https://groups.google.com/d/msg/cp2k/-/sAiVZ2dob1sJ</a>. To post to this group, send email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BckEGd5oO7QJ">cp...@googlegroups.com</a>. To unsubscribe from this group, send email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="BckEGd5oO7QJ">cp2k+uns...@googlegroups.com</a>. For more options, visit this group at <a href="http://groups.google.com/group/cp2k?hl=en" target="_blank">http://groups.google.com/group/cp2k?hl=en</a>. [attachment "energy-conservation.pdf" removed by Jürg Hutter/at/UZH]</blockquote>