Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs?
Hi Mark, *So, for the record, the actual issue is that GROMACS does not implement *some kind of tapering of non-bonded interactions that would apply between *particle pairs at short range? This is the extract from the supporting information of the paper I refer to that explains the issue (³http://irep.ntu.ac.uk/id/eprint/4652/7/4652__Perry_SI.pdf²): "Full inclusion of attractive nonbond interactions between 1,4 bonded atoms in most force fields (CHARMM, CVFF, PCFF, COMPASS) requires a 4% increase in r (the equilibrium bond length parameter of the forcefield) Partial inclusion of attractive nonbond interactions between 1,4 bonded atoms (50% of Lennard-Jones interactions and 5/6 of Coulomb interactions in AMBER) requires only a 2.5% increase, and 12-6 LJ parameters identical to these in CVFF and CHARMM can then be used." My particular issue was that I assumed gen-pairs=yes in the [ default ] section of the topology file meant that Gromacs automatically generated a list of atoms undergoing 1-4 interactions (as well as automatically generating the modified LJ and Coulomb parameters for those interactions). When I proceeded to manually specify the list of atoms undergoing 1-4 interactions within the [pairs] section of the topology, the silica system compressed during the simulation to produce the correct equilibrium bond length observed experimentally. Thanks, Amanda On 03/07/2017, 06:34, "gromacs.org_gmx-users-boun...@maillist.sys.kth.se on behalf of gromacs.org_gmx-users-requ...@maillist.sys.kth.se" wrote: >Send gromacs.org_gmx-users mailing list submissions to > gromacs.org_gmx-users@maillist.sys.kth.se > >To subscribe or unsubscribe via the World Wide Web, visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users >or, via email, send a message with subject or body 'help' to > gromacs.org_gmx-users-requ...@maillist.sys.kth.se > >You can reach the person managing the list at > gromacs.org_gmx-users-ow...@maillist.sys.kth.se > >When replying, please edit your Subject line so it is more specific >than "Re: Contents of gromacs.org_gmx-users digest..." > > >Today's Topics: > > 1. Re: Anybody using Silica InterfaceFF on Gromacs? (Mark Abraham) > 2. Re: Can not open file: run.trr (Mark Abraham) > 3. Re: Can not open file: run.trr (Mark Abraham) > 4. Re: Anybody using Silica InterfaceFF on Gromacs? (Alex) > 5. Multiple output log. xtc .edr files generated using mpirun > (Abhishek Acharya) > 6. Grid Spacing (Apramita Chand) > > >------ > >Message: 1 >Date: Mon, 03 Jul 2017 00:03:52 + >From: Mark Abraham >To: gmx-us...@gromacs.org >Subject: Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs? >Message-ID: > >Content-Type: text/plain; charset="UTF-8" > >Hi, > >On Fri, Jun 30, 2017 at 2:46 AM Alex wrote: > >> > >> > >> >> >> >> He he, childish :) >> > >> > David, no offense intended. I just think that when applied to solids, >>the >> entire concept of what works so well for biomolecular systems becomes a >>bit >> of a joke. And vice versa, to be fair. Spoken from experience, really >>-- we >> here used Gromacs to simulate things that I keep telling people not to >> simulate with Gromacs, and it got published!. :) >> > >Sure - one should use a tool capable of doing the job, and use its >optional >capabilities accordingly. Choosing non-bonded exclusions based on bonds, >in >a way that does not implement the intended model physics, is indeed wrong, >but that doesn't make the use of modified 1-4 pair interactions "childish" >in general ;-) > > >> In any case, I second what was said above re: # of exclusions. >>Solid-state >> potentials use smooth drop-offs to exclude long-range interactions >>between >> close neighbor sharing elements, so looking into David's suggestion may >>in >> fact fix the issues immediately. >> > >So, for the record, the actual issue is that GROMACS does not implement >some kind of tapering of non-bonded interactions that would apply between >particle pairs at short range? > >Mark > > >> Alex >> -- >> Gromacs Users mailing list >> >> * Please search the archive at >> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before >> posting! >> >> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >> >> * For (un)subscribe requests visit >> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or >> send a mail to gmx-users
Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs?
Sure - one should use a tool capable of doing the job, and use its optional capabilities accordingly. Choosing non-bonded exclusions based on bonds, in a way that does not implement the intended model physics, is indeed wrong, but that doesn't make the use of modified 1-4 pair interactions "childish" in general ;-) You are correct. My point only holds in a situation where people indeed are doing that: using an inappropriate tool. And you see how many people on this list ask questions about doing solid-state simulations with Gromacs. I have a dual background and could give many examples where the use of what works for proteins/lipids in water is, well, childish, for lack of a better word in solid-state systems. I hope noone takes offense. :) The "problem" is that, thanks to your hard work, Gromacs is delightful to use, much more so than LAMMPS, and my hope is that one day you guys will just implement things like REBO and ReaxFF and MEAM, and Weber-Stillinger and what-not. If you do that, I will fly to your location, give you a hug, and personally buy you a steak dinner. In other words, my hope is that one day Gromacs will be an appropriate tool for all of classical MD, solid or liquid state. So, for the record, the actual issue is that GROMACS does not implement some kind of tapering of non-bonded interactions that would apply between particle pairs at short range? Within the question initially posed (about silica), Gromacs is fairly capable of everything necessary, it is just a matter of carefully reading the documentation on the original solid-state potential and then setting up the adopted forcefield, including the nrexl, etc. The much bigger issue (with solid-state things in general) is the springs for bonds, angles, etc, and use permanent molecular topologies. Wrong phonon spectra, wrong elastic properties away from 0K, wrong rippling spectra of graphene and likely all atomically thin membranes, inability to do covalent bond breaking/formation, etc, etc. That is, things Gromacs was never intended for. But then again, see what I said above. Alex -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs?
Hi, On Fri, Jun 30, 2017 at 2:46 AM Alex wrote: > > > > > >> > >> He he, childish :) > > > > David, no offense intended. I just think that when applied to solids, the > entire concept of what works so well for biomolecular systems becomes a bit > of a joke. And vice versa, to be fair. Spoken from experience, really -- we > here used Gromacs to simulate things that I keep telling people not to > simulate with Gromacs, and it got published!. :) > Sure - one should use a tool capable of doing the job, and use its optional capabilities accordingly. Choosing non-bonded exclusions based on bonds, in a way that does not implement the intended model physics, is indeed wrong, but that doesn't make the use of modified 1-4 pair interactions "childish" in general ;-) > In any case, I second what was said above re: # of exclusions. Solid-state > potentials use smooth drop-offs to exclude long-range interactions between > close neighbor sharing elements, so looking into David's suggestion may in > fact fix the issues immediately. > So, for the record, the actual issue is that GROMACS does not implement some kind of tapering of non-bonded interactions that would apply between particle pairs at short range? Mark > Alex > -- > Gromacs Users mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before > posting! > > * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists > > * For (un)subscribe requests visit > https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or > send a mail to gmx-users-requ...@gromacs.org. > -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs?
is quite small, only about 8000 atoms. I have run this system >for 100 ns, which took roughly about 2 days. Hence, a run of 1 >microsecond would take about 20 days. I am trying to shorten it down to 2 >days by using more than 1 node. >Thanks, >Thanh Le >> On Jun 29, 2017, at 3:10 PM, >>gromacs.org_gmx-users-requ...@maillist.sys.kth.se wrote: >> >>> http://www.gromacs.org/Support/Mailing_Lists >>><http://www.gromacs.org/Support/Mailing_Lists> > > >-- > >Message: 4 >Date: Thu, 29 Jun 2017 22:37:54 + >From: Mark Abraham >To: gmx-us...@gromacs.org, gromacs.org_gmx-users@maillist.sys.kth.se >Subject: Re: [gmx-users] gromacs.org_gmx-users Digest, Vol 158, Issue > 186 >Message-ID: > >Content-Type: text/plain; charset="UTF-8" > >Hi, > >Don't even consider running on more than one node. You can see for >yourself >by comparing the performance of even just e.g. > >gmx mdrun -nt 1 -pin on >gmx mdrun -nt 2 -pin on >gmx mdrun -nt 14 -pin on >gmx mdrun -nt 28 -pin on > >... to run on different numbers of cores. Parallel efficiency drops off as >you approach 100 atoms per core. > >Further, the factor of seven in the core count is a surefire way to be >inefficient, because the domain decomposition will have to partition in >seven domains in one direction. I would consider running three simulations >per node, with 9,10,9 cores per simulation, using gmx mdrun -nt x -pin on >-pin_offset y for suitable x and y. But try the above experiment first. > >Mark > >On Fri, Jun 30, 2017 at 12:21 AM Thanh Le wrote: > >> Hi Mr. Abraham. >> My system is quite small, only about 8000 atoms. I have run this system >> for 100 ns, which took roughly about 2 days. Hence, a run of 1 >>microsecond >> would take about 20 days. I am trying to shorten it down to 2 days by >>using >> more than 1 node. >> Thanks, >> Thanh Le >> > On Jun 29, 2017, at 3:10 PM, >> gromacs.org_gmx-users-requ...@maillist.sys.kth.se wrote: >> > >> >> http://www.gromacs.org/Support/Mailing_Lists < >> http://www.gromacs.org/Support/Mailing_Lists> >> -- >> Gromacs Users mailing list >> >> * Please search the archive at >> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before >> posting! >> >> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >> >> * For (un)subscribe requests visit >> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or >> send a mail to gmx-users-requ...@gromacs.org. >> > > >-- > >Message: 5 >Date: Thu, 29 Jun 2017 22:37:54 + >From: Mark Abraham >To: gmx-us...@gromacs.org, gromacs.org_gmx-users@maillist.sys.kth.se >Subject: Re: [gmx-users] gromacs.org_gmx-users Digest, Vol 158, Issue > 186 >Message-ID: > >Content-Type: text/plain; charset="UTF-8" > >Hi, > >Don't even consider running on more than one node. You can see for >yourself >by comparing the performance of even just e.g. > >gmx mdrun -nt 1 -pin on >gmx mdrun -nt 2 -pin on >gmx mdrun -nt 14 -pin on >gmx mdrun -nt 28 -pin on > >... to run on different numbers of cores. Parallel efficiency drops off as >you approach 100 atoms per core. > >Further, the factor of seven in the core count is a surefire way to be >inefficient, because the domain decomposition will have to partition in >seven domains in one direction. I would consider running three simulations >per node, with 9,10,9 cores per simulation, using gmx mdrun -nt x -pin on >-pin_offset y for suitable x and y. But try the above experiment first. > >Mark > >On Fri, Jun 30, 2017 at 12:21 AM Thanh Le wrote: > >> Hi Mr. Abraham. >> My system is quite small, only about 8000 atoms. I have run this system >> for 100 ns, which took roughly about 2 days. Hence, a run of 1 >>microsecond >> would take about 20 days. I am trying to shorten it down to 2 days by >>using >> more than 1 node. >> Thanks, >> Thanh Le >> > On Jun 29, 2017, at 3:10 PM, >> gromacs.org_gmx-users-requ...@maillist.sys.kth.se wrote: >> > >> >> http://www.gromacs.org/Support/Mailing_Lists < >> http://www.gromacs.org/Support/Mailing_Lists> >> -- >> Gromacs Users mailing list >> >> * Please search the archive at >> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before >> posting! >> >> * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists >> >> * For (un)subscribe requests visit &
Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs?
> > >> >> He he, childish :) > > David, no offense intended. I just think that when applied to solids, the entire concept of what works so well for biomolecular systems becomes a bit of a joke. And vice versa, to be fair. Spoken from experience, really -- we here used Gromacs to simulate things that I keep telling people not to simulate with Gromacs, and it got published!. :) In any case, I second what was said above re: # of exclusions. Solid-state potentials use smooth drop-offs to exclude long-range interactions between close neighbor sharing elements, so looking into David's suggestion may in fact fix the issues immediately. Alex -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs?
On 28/06/17 20:22, Alex wrote: This community is mostly focused on other things. If you have solid silica under a non-native (to Gromacs) forcefield and all the bonded parameters have been copied correctly, there may be issues with your partial charges, LJ parameters, and mixing rules. Also make sure your original forcefield uses the same approach to neighbor searching as Gromacs does. Also to my knowledge, very few solid-state potentials use the childish 1-4 pair interaction. He he, childish :) Anyway you may in addition check the number of exclusions set in your topology. Alex On 6/28/2017 11:49 AM, Diez Fernandez, Amanda wrote: Hi, I converted the InterfaceFF silica parameters to use in Gromacs (and to be compatible with the AMBER forcefield) and have gotten some problems, namely I am getting a slightly bigger equilibrium bond length than I should. I was wondering if there were other Gromacs users out there that have used InterfaceFF silica parameters in Gromacs. Many thanks, Amanda -- David van der Spoel, Ph.D., Professor of Biology Dept. of Cell & Molec. Biol., Uppsala University. Box 596, 75124 Uppsala, Sweden. Phone: +46184714205. Fax: +4618511755. sp...@xray.bmc.uu.se http://folding.bmc.uu.se -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
Re: [gmx-users] Anybody using Silica InterfaceFF on Gromacs?
This community is mostly focused on other things. If you have solid silica under a non-native (to Gromacs) forcefield and all the bonded parameters have been copied correctly, there may be issues with your partial charges, LJ parameters, and mixing rules. Also make sure your original forcefield uses the same approach to neighbor searching as Gromacs does. Also to my knowledge, very few solid-state potentials use the childish 1-4 pair interaction. Alex On 6/28/2017 11:49 AM, Diez Fernandez, Amanda wrote: Hi, I converted the InterfaceFF silica parameters to use in Gromacs (and to be compatible with the AMBER forcefield) and have gotten some problems, namely I am getting a slightly bigger equilibrium bond length than I should. I was wondering if there were other Gromacs users out there that have used InterfaceFF silica parameters in Gromacs. Many thanks, Amanda -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
[gmx-users] Anybody using Silica InterfaceFF on Gromacs?
Hi, I converted the InterfaceFF silica parameters to use in Gromacs (and to be compatible with the AMBER forcefield) and have gotten some problems, namely I am getting a slightly bigger equilibrium bond length than I should. I was wondering if there were other Gromacs users out there that have used InterfaceFF silica parameters in Gromacs. Many thanks, Amanda -- Gromacs Users mailing list * Please search the archive at http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List before posting! * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists * For (un)subscribe requests visit https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users or send a mail to gmx-users-requ...@gromacs.org.
