Re: [gmx-users] Assistance needed running gromacs 4.6.3 on Blue Gene/P
On Wed, Aug 7, 2013 at 5:04 PM, Prentice Bisbal wrote: > I couldn't find anything explicit in the documentation, The compute node > kernel (CNK) is proprietary, but is said to be very Linux-like, so a signal > 6 should be the same as on Linux, which would be an abort signal. There are several calls to abort() in the GROMACS source and probably some more in external libs, so it's impossible to know where it fails; it's also possible that the signal comes from outside GROMACS. It would be very helpful to run a debug build and obtain a backtrace, as suggested by Mark. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Assistance needed running gromacs 4.6.3 on Blue Gene/P
Hi Prentice, The subject and first phrase mention version 4.6.3, but all scripts refer to 4.6.2. Are you sure you are not mixing versions ? Do you have another 4.6.x version working already ? > -DCMAKE_C_FLAGS="-O3 -qarch=450d -qtune=450" \ Could you try with a lower optimization level ? > BE_MPI (ERROR): "killed with signal 6" What is signal 6 on this platform ? Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Gromacs 4.6 crushes in PBS queue system
On Tue, Feb 19, 2013 at 1:32 PM, Tomek Wlodarski wrote: > The problem is that this is only message I got... But that's an important error. GROMACS 4.6 depends a lot on the compiler generated code, as opposed to 4.5 and previous which used hand-written assembler code. The build procedure detects the architecture and instructs the compiler to optimize for it. If the computer on which you build GROMACS 4.6 has a different architecture from the one on which the executables are run, you get exactly this error. F.e. if your build computer is has Intel Sandy Bridge CPUs and the compute nodes have AMD Magny Cours CPUs, the executable won't run. You have to instruct the build procedure to use the proper type of acceleration for the compute nodes where the executable will be run. > I also get this Warning: > -- > WARNING: Open MPI will create a shared memory backing file in a > directory that appears to be mounted on a network filesystem. OpenMPI users have found that shared memory performance was lower when the shared memory backing file was created on a shared file system. This mechanism is used by default when you run several MPI ranks on the same computer. There's some documentation on the OpenMPI site on the implications and how to control this. > but this I got also with gromacs 4.5.5 which is running ok so I think this > is not a problem in my case. ... so performance with GROMACS 4.5.5 might have also suffered from this :) > Like Alexey notice the problem is that my nodes have different > architecture.. but this was not a problem with gromacs 4.5.5 Indeed, for the reason I mentioned above. Check for example the presence/absence of "avx" in the "flags" section. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] dummy atom - regd
On Mon, Feb 18, 2013 at 12:10 PM, Mark Abraham wrote: > b) a particle needs to be at that immobile reference point - you can > measure a distance to a point whether or not there is something there. I'm not sure that it's the same use-case... I recently talked to someone trying to define such a fixed particle to be used as a pull group. This pull group could then be set as reference in the middle of the PBC box, such that the pull distances are smaller than half the box size. The fact that the particle is really fixed when the PBC box is allowed to change dimensions is not a big problem, if needed the size of the box for each trajectory frame can then be used to correct distances. I don't know what was the outcome, but I'm curious to see such a setup. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Many energygrps to output
On Thu, Feb 7, 2013 at 8:13 PM, Yun Shi wrote: > So instead of making an index file with 399 groups of each residue in > A and typing in rerun.mdp file 400 group names as > "energygrps = A1 A2 A3 A4 ... A399 B", and issuing g_energy command 399 times, There can be maximum 256 groups defined at once (Chapter 3.3 in manual), so you can't do this anyway... Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Getting the virial for each individual atom.
On Mon, Jan 21, 2013 at 7:06 PM, Marcin, Zachary wrote > dvxx+= x[i][xx] *f[i][XX]; > Compare the indexing above with: > This is the same as the formula that I am trying to solve for (Vi*Pi + > .5*SUM(Rij * Fij)). > ... the ones you want to have. See the difference ? The existing code takes one atom at a time (i), what you want takes 2 atoms at a time (i & j). You cannot calculate what you want from there, as the forces are already summed per atom; Fij are the pairwise forces calculated in the kernels corresponding to the different potentials. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Re: Simulation of 2D lattice model
On Thu, Jan 17, 2013 at 1:51 PM, James Starlight wrote: > So if I have force field with the C6/C12 terms (instead of > sigma\epsilon) I need to express sigma (which correspond to the Rmin > in LJ equation) as the (C12/C6)^0.5. Than if I want to increase sigma > ( and consequently to increase vdw radius of my atom) I must to > increase c12 or decreace C6 terms. Does it correct in general ? I think that I have already answered the question - please re-read my previous messages in the thread. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Re: Simulation of 2D lattice model
On Thu, Jan 17, 2013 at 10:59 AM, James Starlight wrote: > thank you for so detailed explanation. You're welcome. Now it's up to you to use it :) > Now it's only intresting to me > if it possible to change vdw radius (assuming it as the distance from > center of atom to it outer electronic shell)? Yes, it's possible. > E.g I want to change > such value for each node in my lattice model (e.g for CH2 united atom > that value might be > 1.5 and I want dicrease it to the 1 A ( as for > simple carbon). Might I do it via some simplest modifications in the > nonbonded.itp or should I do that by means of changes in > sigma/epsillon terms (Which I also must express from A(c6) and B(c12) > based on gromac's combination rules) ? If nonbonded.itp for your chosen force field is expressed in sigma/epsilon (I guess that this is the case because you talk about an existing value for sigma), then it's only a matter of changing the sigma parameter of that atomic species. Keep in mind that this is only a description of the change - it doesn't guarantee that the change is valid. If I'd be a reviewer I'd ask for a reason for changing one of the parameters (sigma) and keeping the other (epsilon) constant. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Re: Turn on pdb output after every step
On Mon, Jan 14, 2013 at 9:23 PM, santhosh wrote: > I want to output pdb during the run. Before every integration step, I want > to output the current configuration of the system. Before an integration step is after the previous integration step, which is where the normal trajectory writing happens... so it's still not clear how your desired mode of operation differs from what GROMACS currently offers. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] DD cells - regd
On Sat, Jan 12, 2013 at 10:04 AM, ramesh cheerla wrote: > This probably means your > constraint lengths are too long compared to the domain decomposition cell > size. Decrease the number of domain decomposition grid cells This gives another possible solution. You decrease the number of DD grid cells by reducing the numbers of cores on which mdrun runs. Or, for the same number of cores, you can control better the decomposition using the '-dd' option in case the constraints are aligned to one of the 3 main directions. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Re: Simulation of 2D lattice model
On Sun, Jan 6, 2013 at 1:44 PM, James Starlight wrote: > I mean absence of exponential factor in the C6 term :) > > So to change the vdw radius of the specified atom I should to varry > both c6 and c12 shouldn't it ? Hmm, to me these look like very basic force field questions. Did you try to look the Lennard-Jones potential in a good MD book ? Or in the vast amount of resources available for free online ? Or, even better, in the GROMACS manual ? That would take surely less than an hour, while your questions on the subject have stretched over several days. The 6-12 LJ potential is composed of an attractive and repulsive term. The 6-12 combination is often used because of computational efficiency, as the 12-term can be obtained by multiplying the 6-term with itself. Each of these terms has a constant, typically called C6 and C12, which tells how much each term contributes to the total potential. For example, to have only the 6-term, C12 can be set to zero - this is useful when one knows how much attractive and how much repulsive the potential should be. Most often though, one thinks of the LJ potential in terms of equilibrium distance (obtained through a combination rule) and potential well, expressed through the sigma and epsilon constants. The two pairs (C6/C12 and sigma/epsilon) are interrelated. The relation is given in the GROMACS manual, on the Wikipedia page related to the LJ potential and in many other places. GROMACS also comes with a tool (g_sigeps) which allows an easy transformation between them. As you can see from the formulas, sigma depends on both C6 and C12 and epsilon depends on both C6 and C12. So, to partly answer your question, to change the equilibrium distance (sigma) you need to vary both C6 and C12. It's only partly answered because you need to read about combination rules in the GROMACS manual to see how to get from the atom radius to LJ potential sigma... Please note that a particular force field uses only one of the pairs (C6/C12 or sigma/epsilon) - you can't mix and match. If you want to use f.e. OPLS-AA, all LJ interactions are expressed using the sigma/epsilon pair. If you want to introduce in this force field a new type of interaction based on another force field which is defined using the C6/C12 pair, you have to perform the conversion to sigma/epsilon. If, on the other hand, you design your own force field, you are free to use either of the two pairs but, once chosen, you have to be consistent and use that pair for all LJ interactions. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Re: Simulation of 2D lattice model
On Wed, Jan 2, 2013 at 2:58 PM, James Starlight wrote: > By the way now I'd like to simulate the same 2D flexible ( in non > planar hybridization) lattice but consist of non-hydrogens atoms only > ( only atoms forming lattice scaffold ). Does it possible to change > Carbon atom-valence (from 4 to 3) preventing tetragonal-like geometry > ( >c- ) or should I use anothe atom type for such lattice ? I can't imagine how it can be both "non planar hybridization" and "preventing tetragonal-like geometry", but maybe I don't understand your intentions. Assuming you do want a change in geometry - from tetragonal to planar - then you need to use different parameters to describe the different geometry, it's not only about changing valence. If you use OPLS-AA, this means a change of atom type from something like C sp3 (f.e. opls_139) to sp2 (f.e. opls_147) - both these types chosen for having zero charge. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Re: Simulation of 2D lattice model
On Wed, Jan 2, 2013 at 1:15 PM, James Starlight wrote: > so why my simulation of the sp3 lattice have been crashed with the > links warnings ( which always tells about wrong geometry in topology)? I might have missed something from the beginning of this discussion... Why do you use LINCS with a lattice ? What do you expect to obtain ? LINCS is normally used to keep bonds at their equilibrium length. If you do this in a lattice, the degrees of freedom will be strongly reduced, due to the coupled/restricted motions. How does this influence what you want to observe ? If you are not interested in the flexibility of the lattice itself, have you considered freezing it ? Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Re: What is the purpose of the [ pairs ] section?
Hi Andrew! I've also been a bit confused by the names when I started looking at the force field definitions in GROMACS. Although Justin's replies contain all the information (as usual :)), let's try to shed some light with an example: consider a chain formed by atoms i-j-k-l-m-n; - means that there is a bond between atoms. nrexcl specifies how many neighbors should be excluded when building lists of nonbonded interactions. nrexcl = 3 means that the first, second and third order neighbors (j, k and l for atom i; k, l and m for atom j; l, m and n for atom k; m and n for atom l; n for atom m) are excluded, such that the following pairs will be put on the lists to be passed to the nonbonded kernels: i m i n j n For another force field with nrexcl = 2, the following pairs would be passed to the nonbonded kernels: i l i m i n j m j n k n For another force field with nrexcl = 5, no pairs would be passed to the nonbonded kernels. More generally, any pairs which are neighbors of higher order than nrexcl will be put on the nonbonded lists. Parameters for these nonbonded interactions will be automatically computed from the topology data, based on the combination rules. If you have empty [bonds], [angles], [dihedrals] and [pairs] sections in the topology file, only the above interactions are considered for potential energy calculations. These nonbonded lists are dynamical and are remade based on neighbor search frequency; when using cutoffs, a pair whose atoms are further apart than the cutoff is not added to the list. Bonded interactions are specified in the [bonds], [angles] and [dihedrals] sections. Parameters for the bonded interactions are taken directly from the topology data after matching atom types. Any of these sections can be empty, if the force field is so designed; for OPLS-AA all these sections are present. There can be overlaps, f.e. if nrexcl = 2 and the topology file contains: [dihedrals] i j k l j k l m k l m n then atoms i and l, j and m and k and n will interact both as part of the dihedral angle potential and as nonbonded pair. Some force fields (like OPLS-AA) specify that the 1-4 interactions ((i l), (j m) and (k n) in the above example) should indeed be treated both as part of a dihedral angle potential and as nonbonded pairs; when treating them as nonbonded pairs there might be a scaling factor (0.5 for OPLS-AA). When there is no scaling factor (in other words, scaling factor is 1), a definition like above with nrexcl = 2 is all that is needed. For a scaling factor different from 1, the GROMACS implementation uses an explicit pairs list, called [pairs]. All 1-4 interactions should be included in the [pairs] section, which for the above example would look like: [pairs] i l j m k n It's the user's responsibility to add all needed (1-4 in case of OPLS-AA) pairs to the [pairs] section; pdb2gmx normally does this, but if you build the topology through other means then you have to make sure these pairs are generated. The interactions for these explicit pairs are always computed - even though they are computed through the nonbonded kernels, they are not subject to cutoffs. These explicit pairs also need parameters. The parameters are automatically generated when gen-pairs = yes from the topology data, based on the combination rules and the fudge (=scaling) factors. If gen-pairs = no, then the [pairs] section should contain both the atom indeces and the parameters. I found the name gen-pairs a bit misleading, I would have called it gen-pairs-params or something similar as it is only taken into consideration when looking for parameters; the [pairs] list is not automatically updated when generating the .tpr file. The [pairs] section offers some flexibility, as not only the 1-4 interactions but any other interaction that should be treated the same way can be included; a possible limitation is that only one fudge factor can be set for the whole list of explicit pairs. Care should be taken to not calculate the same interaction twice. With nrexcl = 3 and: [pairs] i m the nonbonded interaction between atoms i and m will be computed twice: once as part of explicit pairs list and once as part of nonbonded lists. Finally, I would like to point out that most of these can be found by reading the manual and experimenting with a toy molecular system (like the example above) using gmxdump and a combination of mdrun -rerun and g_energy. Cheers, Bogdan On Mon, Dec 10, 2012 at 9:13 PM, Andrew DeYoung wrote: > Javier and Mark, > > Thank you SO much! That is so helpful to me. > > If you have time, I have two follow-up questions: > > (1) From your answer, it seems that in OPLS, 1-4 interactions will be taken > into account (scaled by fudge) ONLY if the 1-4 atomic indices are specified > in [ pairs ], even if gen-pairs = yes and sigma and epsilon values are > provided in [ atomtypes ] in ffnonbonded.itp. > > But what about 1-5, 1-6, 1-7, ..., 1-Infinity nonbonded interactions, in a > protein for example?
Re: [gmx-users] Scientific notation in .itp files
Hi Andrew! In addition to Justin's reply, I would like to point out that this is something you can easily answer yourself by building a few atoms system using your .itp file and running gmxdump to look at the actual values stored in the .tpr. Cheers, Bogdan On Sat, Dec 8, 2012 at 6:53 PM, Andrew DeYoung wrote: > Hi, > > I have been trying to rigorously check my .itp (topology) files for errors. > Is it true that .itp files allow scientific notation? > > For example, suppose that I want to specify sigma = 0.35 nm in > ffnonbonded.itp. If I enter either > > 3.5e-1 > > or > > 3.5E-1 > > will these entries indeed specify the value 0.35? That is, is it true that > one can use "e" or "E" to specify "times 10 raised to the power"? > > Thanks for your time! > > Andrew DeYoung > Carnegie Mellon University -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Freeze group atoms changing position
On Thu, Oct 25, 2012 at 3:59 PM, Alex Marshall wrote: > Thanks Justin. I identified the offending waters using vmd (adding 1 to > resID and atom number since vmd starts counting at 0) and checked > confout.gro to make sure the coordinates matched up. I only have one group > for all frozen atoms in the system, and these guys are definitely in it. Are you using some kind of constraints ? Are you using energy group exclusions to avoid interactions between frozen atoms ? If you search the manual for "frozen" you'll find some warnings and recommendations. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] question about the minimization
On Thu, Oct 18, 2012 at 10:53 AM, wrote: > Steepest Descents converged to machine precision in 20385 steps, > but did not reach the requested Fmax < 10. > Potential Energy = -8.8805600e+05 > Maximum force = 3.2341940e+02 on atom 3050 > Norm of force = 2.0930219e+00 > > I do not see why it can not reach a lower force. Probably you run mdrun in single precision. To reach such low force you normally need a double precision mdrun. Do you actually need such a low force ? There were many messages on this subject from Justin on this very list, please look them up. Good luck with your simulations! Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] domain decomposition trick for highly parallel non-equilibrium runs
On Sat, Oct 6, 2012 at 6:03 PM, Dr. Vitaly Chaban wrote: > Would anyone perhaps suggest some tricks to increase the stability of > the highly parallel jobs (NP > 100), which use domain decomposition > among nodes. It would be more helpful to know what you are talking about, i.e. a definition of stability (or maybe of instability, which you probably try to avoid)... Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] MD problem with nwall=2
On Wed, Sep 5, 2012 at 8:37 PM, Peter C. Lai wrote: > Could be a result of not setting x/y compressibility = 0 as the manual > suggests you should do... As one who has also tried to use walls recently, I've also found this statement in the manual, but no further explanation. Anyone cares to expand on it ? For testing, I've recently performed a run with compressibility=0 for z and non-zero for x/y (so, exactly opposite to what the manual suggests :)). The box deformed as expected in x/y and the pressure seemed to be maintained correctly. Is there some reason for which this would be a random (i.e. not easily reproducible) result ? Or are there some less obvious problems with it ? Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Elastic network model
On Fri, Aug 31, 2012 at 8:42 PM, mohan maruthi sena wrote: > Hi all, > I have built topology for elastic network as follows(the > atoms which fall with in a distance less than 6 angtroms are bonded). > [bonds] > ;aiaj funct bo kb > 2 1 1 3.80444752888 61.9 Less than 6 Angstrom ??? You need to read the topology chapter of the manual again. And then again. It will save you lots of time later. Hint: some units are wrong. Good luck! Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Umbrella Sampling Pull code Problem
On Fri, Aug 31, 2012 at 1:45 PM, Steinbrecher, Thomas (IPC) wrote: > However, when I make the following two changes in my input file: > > pull_start = no > pull_init1 = 1.95 > > which should (?) amount to an equivalent setup, a very different trajectory > results in which the COM distance quickly increases to 2.7 nm and then > appears to be restrained there. (Visualization confirms, in the first case, > the groups remain in their starting conformation, in the second one, they are > pushed appart) My guess is that you should use -1.95 as pull_init1 (mind the minus sign) and that the COM distance you measure is actually close to 2.9 nm (which is 1.95*2). g_dist measures a distance which is always non-negative, however pull_init1 is used as a reference position so its sign depends on the geometry of pulling. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! * Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Software inconsistency error: One or more interactions were multiple assigned in the domain decomposition
On Fri, Apr 20, 2012 at 14:49, Olivia Waring wrote: > My invocation of mdrun, using 8 processors and the following mdp file: > ... > There are: 1184 Atoms > ... > DD step 4 load imb.: force 181.0% > > Not all bonded interactions have been properly assigned to the domain > decomposition cells Hi Olivia! The translation to English would be: "you have too few atoms for a domain decomposition on 8 processors to succeed, given the parameters in your setup." :) With so few atoms, you won't get much speedup, if at all, when running on more than 1 processor. The rule of thumb posted on this very list quite often is 1000 atoms/processor for a reasonably good scaling. To convince GROMACS 4.5.x to use only one processor on a multi-core machine, run with the '-nt 1' command line option. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] g_angle calculation
On Wed, Dec 21, 2011 at 18:44, SebastianWaltz wrote: > Hi all, > > I try to reproduce the results from g_angle using my own c++ code. Using > the formula > > θ=acos(r_ij * r_kj/|r_ij||r_kj|) > > with the acos from the cmath library, I obtain an angle which is always > ~5% larger when the angel calculated by g_angle. > Does g_angle use the same formula or does it calculate θ in a different way? The angles are calculated through src/gmxlib/bondfree.c::bond_angle() which does (after taking care of PBC): *costh=cos_angle(r_ij,r_kj); th=acos(*costh); cos_angle() is defined in include/vec.h to do r_ij * r_kj/|r_ij||r_kj|; acos() comes from the math lib, it's not a GROMACS function. So it's essentially what you seem to do. I guess approximations and precision loss might play a role, f.e. cos_angle() uses GROMACS' own gmx_invsqrt() - do you use math lib sqrt() to calculate the vector norm ? You could try cut-and-pasting the few pieces of code making up the GROMACS implementation of these functions - they are well contained in include/vec.h. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] h-bonds constraints
On Fri, Aug 19, 2011 at 20:02, Dr. Vitaly V. Chaban wrote: > What criterion does the grompp utility use to identify H-atoms when it > is requested in the MDP file to apply constraints=h-bonds? Thanks. Any atom name starting with H or h (so case-insensitive). Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] how to simulate a line charge
On Sat, Aug 13, 2011 at 05:43, Amit Choubey wrote: > The largest charge group contains 267 atoms. The maximum is 32. Column 6 in the [atoms] section of the .top file defines the charge group. If you want each atom to be in its own charge group, you can set this to the same number as the atom number (column 1). Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] How to suppress the screen info & how to overwrite files?
On Fri, Aug 12, 2011 at 19:23, Justin A. Lemkul wrote: > Try setting the environment variable GMX_MAXBACKUP to zero. This environment variable is read only by the 4.5 series. In 4.0.7 there is no way to control it, there is a hard-coded limit of 128. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] gromacs 2.1
On Sun, Jun 26, 2011 at 14:09, wrote: > /bin/sh: line 0: cd: include: No such file or directory There's a mistake in the Makefile: include is one level higher, but is treated as if it would be in the current level and a "cd .." is done afterwards. You can move include into src, or just add a link to it: cd src ln -s ../include . > dlist.c:197:1: error: pasting "." and "H" does not give a valid > preprocessing token A similar error appears for me also for src/kernel/tpbcmp.c. In both cases, look up the corresponding lines and remove the pasting/concatenation operator "##" as gcc seems to be smart and do what it's supposed to do even without it. F.e. the offending line in src/tools/dlist.c should look like: #define BBB(x) (dl->atm.x != -1) This is supported by several posts on the 'net, f.e.: http://groups.google.com/group/comp.lang.c.moderated/browse_thread/thread/33e7f5f44f25ca81 My tests were done on Fedora 14 with gcc (GCC) 4.5.1 20100924 (Red Hat 4.5.1-4). Good luck! Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Ubuntu Installation
On Tue, Jan 11, 2011 at 14:19, Justin A. Lemkul wrote: > > > Sergio Manzetti wrote: >> >> Hi Justin, it was a blank ./configure This defaults to something else than /usr/lib... > I don't know why a standard location like /usr/lib isn't be detected as > containing the libraries, but you can set LDFLAGS and CPPFLAGS accordingly > so that the libraries and headers are found. See the installation > instructions for details. Because the header files (.h) are also needed - there's none shown in the list. Most likely the list contains files from a fftw package installed with the distribution without the corresponding -devel package. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] changing the velocity in trajectory file
On Mon, Nov 29, 2010 at 16:16, sreelakshmi ramesh wrote: > so can anybody tell me how to hack the trr file so that i can change the > velocity parameter in it You can use the xdrfile library to read in a certain frame of the trr file, change the velocity in the array that you get, then use the xdrfile library again to write the trr file. It's up to you if you recreate only the one frame in the trr file or you copy all frames except for the one that you want to change... That's not so much about "hacking the trr file" as it is about using the library created for this purpose. But this is useful only when you know what velocities you want to generate. Are you not satisfied with the quality of the random velocities generated by GROMACS ? Or do you want to get a particular set of velocities that can't be generated by GROMACS ? Good luck, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] Pairwise forces
On Fri, Nov 19, 2010 at 12:15, Reiner Ribarics wrote: > Our group has already done some rather long simulations and according to the > script files forces are written every 1 steps and can be accessed via > trjconv with -cf option that assigns each atom a temperature factor. I was > wondering if that is the way gromacs stores the forces in the trajectory > file. Or are the temperature factors calculated from other (maybe pairwise) > forces? It's not clear to me what you try to do. Obtain atomic forces ? Obtain pairwise forces ? Know the format in which forces are stored in the trajectory file ? Something else entirely ? > I also had a look at FDA gromacs that is able to write out pairwise forces, > but unfortunately it falls back on the rather slow gromacs routines and we > want to avoid rerunning our simulations if possible. Then please take a look again. The forces are indeed saved by hooking into the C kernels, but in a 'mdrun -rerun' phase which reads back the trajectory and writes the forces, which avoids having a compute intensive phase with the C kernels and the FDA modifications. Kind regards, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-developers] Re: [gmx-users] unexpexted stop of simulation
On Thu, Nov 4, 2010 at 1:08 AM, Roland Schulz wrote: > BTW: Is it somehow possible to print the kernel error messages that are > shown by dmesg to the user from within GROMACS? That would help the user to > directly see the reason of the error. Thus I'm looking for a function > similar to strerror but which returns the kernel message not just the > message of the error code (which in this case was just "Input/Output > errror". Hi Roland! In general, it's not possible to make a connection between a message logged by the Linux kernel (which is then shown by dmesg or the system logging) and a particular call to an I/O function. More specifically, dmesg just dumps the kernel log buffer which, to my knowledge, doesn't contain time information, so the last message in the buffer and the last I/O operation cannot be correlated this way; the system logging (syslogd and similar) attaches time information, but it's usually only readable by root for security reasons - and even if it would be readable by the GROMACS user, there is no way to uniquely associate an entry in syslog with a particular I/O operation on a multiuser/multitasking OS. It might be doable using a tracing infrastructure in the kernel... but that's no longer generic. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
Re: [gmx-users] LJ cut-off distance
2010/10/22 C. Batistakis > > Dear Justin > > Actually, what I want to do is to kill the attraction part of the LJ > substrate-polymer interaction. So, I was thinking to put the cut-off equal > the 1,12σ (point with the lowest energy). > > Do you have any other idea how can I have only the repulsion part? You could switch from using sigma+epsilon to using c6+c12 for LJ interactions - see sections 5.3.3 and 5.7.1 in the manual; this would give you individual control over the attractive and repulsive terms. You don't mention if your forcefield parameters are all custom - if apart from your LJ particles you use some atom types from a forcefield with sigma+epsilon (like OPLSaa) you need to convert those to c6+c12 as well. Cheers, Bogdan -- gmx-users mailing listgmx-users@gromacs.org http://lists.gromacs.org/mailman/listinfo/gmx-users Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/Search before posting! Please don't post (un)subscribe requests to the list. Use the www interface or send it to gmx-users-requ...@gromacs.org. Can't post? Read http://www.gromacs.org/Support/Mailing_Lists
