Re: [gmx-users] Performance on multiple GPUs per node
Hi, Without details of your benchmarks it's hard to comment on why do you not see performance improvement with multiple GPUs per node. Sharing some logs would be helpful. Are you comparing performance with N cores and varying number of GPUs? The balance of hardware resources is a key factor to scaling and my guess is that your runs are essentially CPU-bound, hence adding more GPUs does not help. Have a look at these papers: https://doi.org/10.1002/jcc.24030 https://doi.org/10.1007/978-3-319-15976-8_1 especially the former covers the topic quite well and both show scaling of <100k protein system to 32-64 nodes (dual socket/dual GPU). Cheers, -- Szilárd On Fri, Dec 11, 2015 at 11:54 AM, Jens Krüger < krue...@informatik.uni-tuebingen.de> wrote: > Dear all, > > we are currently planning a new cluster at our universities compute > centre. The big question on our side is, how many and which GPUs we should > put into the nodes. > > We have access to a test system with four Tesla K80s per Node. Using one > GPU node we can reach something like 23 ns/day for the ADH system (PME, > cubic) which is pretty much in line with e.g., > http://exxactcorp.com/index.php/solution/solu_list/84 > > When trying to use 2 or more GPUs on one node, the performance plunges to > below 10 ns/day no matter how we split the MPI/OMP threads. Has anybody of > you guys access to a comparable hardware setup? We would be interested in > benchmark data answering the question: Does GROMACS-5.1 scales on more than > one GPU per node? > > Thanks and best wishes, > > Jens > > > -- > 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] using OH ions in combination with CHARMM27 force field
Thanks Justin for your help. I will carry out the simulation with 0.1M NaOH which I had previously (the coordinate and the topology) and let you know if the desired interactions are taking place or not. Cheers, Soumadwip -- 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] Performance on multiple GPUs per node
Dear all, we are currently planning a new cluster at our universities compute centre. The big question on our side is, how many and which GPUs we should put into the nodes. We have access to a test system with four Tesla K80s per Node. Using one GPU node we can reach something like 23 ns/day for the ADH system (PME, cubic) which is pretty much in line with e.g., http://exxactcorp.com/index.php/solution/solu_list/84 When trying to use 2 or more GPUs on one node, the performance plunges to below 10 ns/day no matter how we split the MPI/OMP threads. Has anybody of you guys access to a comparable hardware setup? We would be interested in benchmark data answering the question: Does GROMACS-5.1 scales on more than one GPU per node? Thanks and best wishes, Jens -- 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] Regarding generating itp file.
Dear GROMACS Admins, I want to generate itp file of crystal structure ( calcium oxalate monohyrate). Can anyone suggest me how to generate itp file. Thanking you, Yours sincerely, Raju Lunkad -- 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] G53A6, non-symmetric selection metrix
Hi Mohsen, This table can indeed be quite confusing at first. I suggest you take a look at http://redmine.gromacs.org/issues/773#note-10 where there is a discussion of how to interpret table 8 (for a specific example, but it should highlight how it works). Hopefully that should answer your questions. Cheers Tom From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [gromacs.org_gmx-users-boun...@maillist.sys.kth.se] on behalf of Mohsen Ramezanpour [ramezanpour.moh...@gmail.com] Sent: 10 December 2015 23:42 To: Discussion list for GROMACS users Subject: [gmx-users] G53A6, non-symmetric selection metrix Dear All, Reading the parameters for Gromos 53A6 ff ( article by Oostenbrink *et al. Journal of computational chemistry* 25.13 (2004): 1656-1676.), I got confused about van der waals interactions between non-bonded atoms. for two atom types, we use combination rules to get (Cij 6) and (Cij 12), and these combination rules (eq. 15 in article) make use of selection matrix (table 8 in article). However this selection matrix is not symmetric. This means L-J interactions between two atom types depends to the order we put them in [pairs]. Is it correct? If this is true, then the order of atom numbers in [pair] matters. But it does not make sense, why (Cij 12) should be different from (Cji 12)? Thanks in advance for your comments Best, Mohsen -- *Rewards work better than punishment ...* -- 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] Regarding generating itp file.
On 12/11/15 7:01 AM, Raju Lunkad wrote: Dear GROMACS Admins, I want to generate itp file of crystal structure ( calcium oxalate monohyrate). Can anyone suggest me how to generate itp file. Ca2+ and water are likely already part of just about any force field. Oxalate is just two carboxylate groups, which is trivial to put together from acid building blocks. -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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] Gromacs 5.0.5 Parallelization GPU Tesla S1070
Hi. I'm getting another problem using Gromacs in GPU parallelization I have a Tesla S1070 (T10 Processor) and I'm trying to compile Gromacs 5.0.5 to recognize it. Firstly, let me specify my hardware/software information: - Ubuntu 14.04 server (no graphical interface) - Tesla S1070 GPU card (T10 Processor): --- PCI bridge: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 --- 3D controller: NVIDIA Corporation GT200GL [Tesla C1060 / M1060] I tried to follow the suggested tutorials (install nvidia driver, cuda and openMM) but mdrun says my card is incompatible, although it's recognized. I'm afraid if Gromacs 5.0.5 version is able to recognize my GPU card, because Tesla S1070 driver (270 version) uses cuda only in 2.2 version (and OpenMM needs higher cuda versions). So, there's a way to configure Tesla S1070 GPU card to get recognized using Gromacs 5.0.5? I mean, what drivers versions are needed (or what Gromacs versions supports it)? Best regards, Antoniel A. S. Gomes Bacharel em Ciências Biológicas, UEPB - Campus V Mestre em Biologia Celular e Molecular, UFPB - Campus I Doutorando em Biologia Geral e Aplicada, UNESP - Instituto de Biociências - Botucatu Lattes: http://lattes.cnpq.br/2330777925652141 -- 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] [MGMS-DS]: Molecular Modelling Workshop April 4-6, 2016 in Erlangen, Germany
Dear list subscribers, I am very happy to announce that the 30th Molecular Modelling Workshop in Erlangen takes place on April, Monday 4th to Wednesday 6th, 2016. Starting on Monday after lunch should allow to avoid travelling on weekend and keeping the expenses at a minimum. This time, the scientific program of the meeting is conducted by Prof. Tim Clark, Technical Director of the Computer-Chemistry-Center in Erlangen, Germany. As usual, the workshop is organised by the German Section of the Molecular Graphics and Modelling Society. ### Plenary Speakers ### Over the last few decades computational chemistry, molecular modelling, and chemoinformatics became more and more an integral part of research in various areas of the chemical industry and at universities. A lot of new tools and new approaches together with a huge increase in computer power facilitate this development. We are very pleased to announce Prof. Brian Shoichet University of California, San Francisco, USA and Dr. Marcus A. Neumann Avant-Garde Materials Simulation Deutschland GmbH giving plenary presentations as leading experts their respective fields. ### Poster and Lecture Awards ### As in the past years, there will be two poster awards of EUR 100 each and three lecture awards for the best oral presentations: 1st Winner: Travel bursary to attend the Young Modeller's Forum in London, UK, plus a speaker slot option at YMF 2016 (travel expenses are reimbursed up to EUR 500) 2nd Winner: EUR 200 travel expenses reimbursement 3rd Winner: EUR 100 travel expenses reimbursement Only undergraduate and graduate research students qualify for the poster and lecture awards. ### Registration ### As every year, an interesting and successful workshop depends on your contributions! Therefore let me invite you to submit talks and/or poster titles via the registration form accessible on the workshop website http://mmws2016.mgms-ds.de/index.php?m=register This website will provide all necessary information about the meeting! The deadline for all submissions is February 26th, 2016. +++ Please note that the week before the 30th Molecular Modelling Workshop, on Wednesday March 30th to Friday April 1st 2016, a Symposium Commemorating the Life and Work of Paul von Ragué Schleyer will take place at the same venue. If you are interested in this meeting, please follow the link http://schleyersymposium.mgms-ds.de for further information. +++ We are looking forward to meeting you in Erlangen soon! -Harald Lanig Chairman of the MGMS-DS e.V. http://www.mgms-ds.de -- PD Dr. Harald Lanig Universitaet Erlangen/Nuernberg Zentralinstitut fuer Scientific Computing (ZISC) Geschaeftsfuehrer Martensstrasse 5a, 91058 Erlangen Fon +49 9131-85 20781 harald.la...@fau.de Fax +49 9131-85 20785 http://www.zisc.uni-erlangen.de -- 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] gromacs.org_gmx-users Digest, Vol 140, Issue 43
Dear, In calcium oxalate monohydrate the charges are, Type charge Ca 2 C 0.992 O(oxal..) -0.996 O(water) -0.773 H 0.3665 calcium oxalate monohydrate having two type of bonding which are, Ox1 oxalate bounded with calcium only and Ox2, oxalate bounded with calcium and oxygen.So, Kindly help me for generating itp file for these type of interactions. On Fri, Dec 11, 2015 at 5:55 PM, < 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: G53A6, non-symmetric selection metrix (Piggot T.) >2. Performance on multiple GPUs per node (Jens Kr?ger) >3. Re: using OH ions in combination with CHARMM27 force field > (soumadwip ghosh) >4. Re: Performance on multiple GPUs per node (Szil?rd P?ll) >5. Re: Performance on multiple GPUs per node (Szil?rd P?ll) >6. Regarding generating itp file. (Raju Lunkad) >7. Re: Regarding generating itp file. (Justin Lemkul) > > > -- > > Message: 1 > Date: Fri, 11 Dec 2015 11:12:23 + > From: "Piggot T."> To: "gmx-us...@gromacs.org" > Subject: Re: [gmx-users] G53A6, non-symmetric selection metrix > Message-ID: > <6989ab580484164ba0a88714dbe8242057c15...@srv00046.soton.ac.uk> > Content-Type: text/plain; charset="us-ascii" > > Hi Mohsen, > > This table can indeed be quite confusing at first. I suggest you take a > look at http://redmine.gromacs.org/issues/773#note-10 where there is a > discussion of how to interpret table 8 (for a specific example, but it > should highlight how it works). Hopefully that should answer your questions. > > Cheers > > Tom > > From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [ > gromacs.org_gmx-users-boun...@maillist.sys.kth.se] on behalf of Mohsen > Ramezanpour [ramezanpour.moh...@gmail.com] > Sent: 10 December 2015 23:42 > To: Discussion list for GROMACS users > Subject: [gmx-users] G53A6, non-symmetric selection metrix > > Dear All, > > Reading the parameters for Gromos 53A6 ff ( article by Oostenbrink *et al. > Journal of computational chemistry* 25.13 (2004): 1656-1676.), I got > confused about van der waals interactions between non-bonded atoms. > > for two atom types, we use combination rules to get (Cij 6) and (Cij 12), > and these combination rules (eq. 15 in article) make use of selection > matrix (table 8 in article). However this selection matrix is not > symmetric. > > This means L-J interactions between two atom types depends to the order we > put them in [pairs]. Is it correct? > > If this is true, then the order of atom numbers in [pair] matters. But it > does not make sense, why (Cij 12) should be different from (Cji 12)? > > Thanks in advance for your comments > > Best, > Mohsen > > -- > *Rewards work better than punishment ...* > -- > 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: 2 > Date: Fri, 11 Dec 2015 11:54:22 +0100 > From: Jens Kr?ger > To: gromacs.org_gmx-users@maillist.sys.kth.se > Subject: [gmx-users] Performance on multiple GPUs per node > Message-ID: <566aab5e.4070...@informatik.uni-tuebingen.de> > Content-Type: text/plain; charset=utf-8; format=flowed > > Dear all, > > we are currently planning a new cluster at our universities compute > centre. The big question on our side is, how many and which GPUs we > should put into the nodes. > > We have access to a test system with four Tesla K80s per Node. Using one > GPU node we can reach something like 23 ns/day for the ADH system (PME, > cubic) which is pretty much in line with e.g., > http://exxactcorp.com/index.php/solution/solu_list/84 > > When trying to use 2 or more GPUs on one node, the performance plunges > to below 10 ns/day no matter how we split the MPI/OMP threads. Has > anybody of you guys access to a comparable hardware setup? We would be > interested in benchmark data answering the
Re: [gmx-users] Fwd: Help with Gromacs 5.1 mdp options for CHARMM27 force field
On 12/11/15 8:49 AM, Rakesh Ramachandran wrote: Hi Justin, Thanks for the reply. Are nstlist and rlist options important for GPU based gromacs and Verlet cutoff ? I am using nstlist value of 40 and rlist of 1.2 are these appropriate for CHARMM36. Verlet tunes both of these as needed. rlist is automatically adjusted to never miss interactions. The input value is a minimum value; mdrun takes care of the rest. Moreover why is force-switch a better option than other vdw-modifier options for CHARMM36 ? This is the method with which the force field was parametrized. It is absolutely required for membrane systems. Proteins and nucleic acids can be debated, but unless you have a rock-solid, verified, reason to change the vdW method, don't. -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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] [gmx-developers] Gromacs 5.0.5 Parallelization GPU Tesla S1070
Hi Antoniel, GROMACS 4.5 should work, just make sure to install a compatible OpenMM version. Cheers, -- Szilárd On Fri, Dec 11, 2015 at 2:30 PM, Antoniel A. S. Gomeswrote: > Hi Szilárd. > > Yeah, I saw that. But the problem is: in gromacs website ( > http://www.gromacs.org/Documentation/Installation_Instructions_4.5/GROMACS-OpenMM) > says my GPU card is compatible. Probably if I use an older version of > Gromacs it will works right? > > Thank you. > > Best regards, > > Antoniel A. S. Gomes > > Bacharel em Ciências Biológicas, UEPB - Campus V > Mestre em Biologia Celular e Molecular, UFPB - Campus I > Doutorando em Biologia Geral e Aplicada, UNESP - Instituto de Biociências > - Botucatu > Lattes: http://lattes.cnpq.br/2330777925652141 > > - Mensagem original - > De: "Szilárd Páll" > Para: "Discussion list for GROMACS users" > Enviadas: Sexta-feira, 11 de dezembro de 2015 11:25:33 > Assunto: Re: [gmx-users] [gmx-developers] Gromacs 5.0.5 Parallelization > GPU Tesla S1070 > > Hi, > > The hardware detection output should clearly state that your GPU is > incompatible. As the documentation points out [1], only compute capability > 2.0 or newer GPUs are supported with the CUDA acceleration in GROMACS. > > > http://manual.gromacs.org/documentation/5.1/install-guide/index.html#gpu-support > > Cheers, > > -- > Szilárd > > On Fri, Dec 11, 2015 at 2:01 PM, Antoniel A. S. Gomes > > wrote: > > > Hi. > > > > I'm getting another problem using Gromacs in GPU parallelization > > > > I have a Tesla S1070 (T10 Processor) and I'm trying to compile Gromacs > > 5.0.5 to recognize it. > > > > Firstly, let me specify my hardware/software information: > > > > - Ubuntu 14.04 server (no graphical interface) > > - Tesla S1070 GPU card (T10 Processor): > > --- PCI bridge: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro > > Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 > > --- 3D controller: NVIDIA Corporation GT200GL [Tesla C1060 / M1060] > > > > > > I tried to follow the suggested tutorials (install nvidia driver, cuda > and > > openMM) but mdrun says my card is incompatible, although it's recognized. > > > > I'm afraid if Gromacs 5.0.5 version is able to recognize my GPU card, > > because Tesla S1070 driver (270 version) uses cuda only in 2.2 version > (and > > OpenMM needs higher cuda versions). > > > > So, there's a way to configure Tesla S1070 GPU card to get recognized > > using Gromacs 5.0.5? I mean, what drivers versions are needed (or what > > Gromacs versions supports it)? > > > > Best regards, > > > > Antoniel A. S. Gomes > > > > Bacharel em Ciências Biológicas, UEPB - Campus V > > Mestre em Biologia Celular e Molecular, UFPB - Campus I > > Doutorando em Biologia Geral e Aplicada, UNESP - Instituto de Biociências > > - Botucatu > > Lattes: http://lattes.cnpq.br/2330777925652141 > > -- > > Gromacs Developers mailing list > > > > * Please search the archive at > > http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_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-developers > > or send a mail to gmx-developers-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. > -- > 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.
[gmx-users] g_membed: how to input position and orientation of a membrane protein into lipid bilayer?
Dear all, I would like to know who have already used g_membed and where we can define a specific position and orientation of the membrane protein into lipid bilayer. I have an issue due TM have a tilt angle of about 34°. Thank you! Rita. -- 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] Fwd: Help with Gromacs 5.1 mdp options for CHARMM27 force field
Hi Justin, Thanks for the reply. Are nstlist and rlist options important for GPU based gromacs and Verlet cutoff ? I am using nstlist value of 40 and rlist of 1.2 are these appropriate for CHARMM36. Moreover why is force-switch a better option than other vdw-modifier options for CHARMM36 ? Regards Rakesh On 9 December 2015 at 23:27, Justin Lemkulwrote: > > > On 12/9/15 12:42 PM, Rakesh Ramachandran wrote: > >> Hi Justin, >> >> Thanks for the reply. Sorry if this question was too long or I was >> not >> clear. >> >> 1) What are the optimal values of tau_t and tau_p for CHARMM36. I see >> in your tutorial that it is 0.1 0.1 for tau_t and 2.0 for tau_p. >> The CHARMM-GUI generated mdp settings had 1.0 1.0 for tau_t and 5.0 for >> tau_p. So are higher values more appropriate for CHARMM36 FF and what >> properties does it effect ? >> >> > Again, to be absolutely clear, values of tau_t and tau_p are in no way > connected with specific force fields. > > The values affect barostat response times and the distributions of > temperature and pressure values. Too short and the simulations can become > unstable, too long and the subsequent values may drift a bit before being > restored. Sometimes longer times are necessary for heterogeneous systems > like membranes. These are things you must decide based on whatever your > system is and whatever you find in the literature. I am not going to tell > you what values you should use. > > 3) I was asking Berendsen pressure coupling with respect to the >> equilibration stage, since I read in some earlier mailing post that >> Berendsen coupling is more appropriate for temperature (NVT) as well as >> pressure (NPT) equilibration. >> >> > Yes, for equilibration, Berendsen is often better because it relaxes > quickly. > > 4) Moreover once I do pbc correction I still see one or two ions sticking >> out of the rhombic dodecahedron box representation in pymol. Is it more of >> a visualization artifact ? >> >> > Yes, just a visualization artifact. > > > -Justin > > -- > == > > Justin A. Lemkul, Ph.D. > Ruth L. Kirschstein NRSA Postdoctoral Fellow > > Department of Pharmaceutical Sciences > School of Pharmacy > Health Sciences Facility II, Room 629 > University of Maryland, Baltimore > 20 Penn St. > Baltimore, MD 21201 > > jalem...@outerbanks.umaryland.edu | (410) 706-7441 > http://mackerell.umaryland.edu/~jalemkul > > == > -- > 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.
[gmx-users] water sorient
Dear all, I would like to calculate the water orientation near ubiqutin by gmx_sorient, gmx sorient -f md.trr -s md.tpr -o -no -ro -co -rc first, the sord.xvg shows the angle (i.e. theta 1) between the vector1 (nearest heavy protein atom to water oxygen) and vector 2 (water oxygen to the mid of of water hydrogen 1 and 2) is mostly 90 degrees, isn't that weird? I would expect water dipole more or less pointing to or away from the vector 1, depending on the charge of the heavy atom. 2nd, sori.xvg second column should be the distribution of cos (theta1), but it shows some value greater than 1. Any ideas? Thanks, Yao -- 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] [gmx-developers] Gromacs 5.0.5 Parallelization GPU Tesla S1070
Hi Szilárd. Yeah, I saw that. But the problem is: in gromacs website (http://www.gromacs.org/Documentation/Installation_Instructions_4.5/GROMACS-OpenMM) says my GPU card is compatible. Probably if I use an older version of Gromacs it will works right? Thank you. Best regards, Antoniel A. S. Gomes Bacharel em Ciências Biológicas, UEPB - Campus V Mestre em Biologia Celular e Molecular, UFPB - Campus I Doutorando em Biologia Geral e Aplicada, UNESP - Instituto de Biociências - Botucatu Lattes: http://lattes.cnpq.br/2330777925652141 - Mensagem original - De: "Szilárd Páll"Para: "Discussion list for GROMACS users" Enviadas: Sexta-feira, 11 de dezembro de 2015 11:25:33 Assunto: Re: [gmx-users] [gmx-developers] Gromacs 5.0.5 Parallelization GPU Tesla S1070 Hi, The hardware detection output should clearly state that your GPU is incompatible. As the documentation points out [1], only compute capability 2.0 or newer GPUs are supported with the CUDA acceleration in GROMACS. http://manual.gromacs.org/documentation/5.1/install-guide/index.html#gpu-support Cheers, -- Szilárd On Fri, Dec 11, 2015 at 2:01 PM, Antoniel A. S. Gomes wrote: > Hi. > > I'm getting another problem using Gromacs in GPU parallelization > > I have a Tesla S1070 (T10 Processor) and I'm trying to compile Gromacs > 5.0.5 to recognize it. > > Firstly, let me specify my hardware/software information: > > - Ubuntu 14.04 server (no graphical interface) > - Tesla S1070 GPU card (T10 Processor): > --- PCI bridge: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro > Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 > --- 3D controller: NVIDIA Corporation GT200GL [Tesla C1060 / M1060] > > > I tried to follow the suggested tutorials (install nvidia driver, cuda and > openMM) but mdrun says my card is incompatible, although it's recognized. > > I'm afraid if Gromacs 5.0.5 version is able to recognize my GPU card, > because Tesla S1070 driver (270 version) uses cuda only in 2.2 version (and > OpenMM needs higher cuda versions). > > So, there's a way to configure Tesla S1070 GPU card to get recognized > using Gromacs 5.0.5? I mean, what drivers versions are needed (or what > Gromacs versions supports it)? > > Best regards, > > Antoniel A. S. Gomes > > Bacharel em Ciências Biológicas, UEPB - Campus V > Mestre em Biologia Celular e Molecular, UFPB - Campus I > Doutorando em Biologia Geral e Aplicada, UNESP - Instituto de Biociências > - Botucatu > Lattes: http://lattes.cnpq.br/2330777925652141 > -- > Gromacs Developers mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_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-developers > or send a mail to gmx-developers-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. -- 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] Regarding generating itp file.
Please heed this: >> When replying, please edit your Subject line so it is more specific >> than "Re: Contents of gromacs.org_gmx-users digest..." ...and don't reply to the whole digest. On 12/11/15 8:02 AM, Raju Lunkad wrote: Dear, In calcium oxalate monohydrate the charges are, Type charge Ca 2 C 0.992 O(oxal..) -0.996 O(water) -0.773 H 0.3665 calcium oxalate monohydrate having two type of bonding which are, Ox1 oxalate bounded with calcium only and Ox2, oxalate bounded with calcium and oxygen.So, Kindly help me for generating itp file for these type of interactions. If you want bonds between those atom types (probably not necessary given the strong electrostatic interaction between the carboxylate and Ca2+) you'll have to do some sort of QM calculation to get the vibrational frequency and subsequently the force constant. -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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] [gmx-developers] Gromacs 5.0.5 Parallelization GPU Tesla S1070
Hi, The hardware detection output should clearly state that your GPU is incompatible. As the documentation points out [1], only compute capability 2.0 or newer GPUs are supported with the CUDA acceleration in GROMACS. http://manual.gromacs.org/documentation/5.1/install-guide/index.html#gpu-support Cheers, -- Szilárd On Fri, Dec 11, 2015 at 2:01 PM, Antoniel A. S. Gomeswrote: > Hi. > > I'm getting another problem using Gromacs in GPU parallelization > > I have a Tesla S1070 (T10 Processor) and I'm trying to compile Gromacs > 5.0.5 to recognize it. > > Firstly, let me specify my hardware/software information: > > - Ubuntu 14.04 server (no graphical interface) > - Tesla S1070 GPU card (T10 Processor): > --- PCI bridge: NVIDIA Corporation NF200 PCIe 2.0 switch for Quadro > Plex S4 / Tesla S870 / Tesla S1070 / Tesla S2050 > --- 3D controller: NVIDIA Corporation GT200GL [Tesla C1060 / M1060] > > > I tried to follow the suggested tutorials (install nvidia driver, cuda and > openMM) but mdrun says my card is incompatible, although it's recognized. > > I'm afraid if Gromacs 5.0.5 version is able to recognize my GPU card, > because Tesla S1070 driver (270 version) uses cuda only in 2.2 version (and > OpenMM needs higher cuda versions). > > So, there's a way to configure Tesla S1070 GPU card to get recognized > using Gromacs 5.0.5? I mean, what drivers versions are needed (or what > Gromacs versions supports it)? > > Best regards, > > Antoniel A. S. Gomes > > Bacharel em Ciências Biológicas, UEPB - Campus V > Mestre em Biologia Celular e Molecular, UFPB - Campus I > Doutorando em Biologia Geral e Aplicada, UNESP - Instituto de Biociências > - Botucatu > Lattes: http://lattes.cnpq.br/2330777925652141 > -- > Gromacs Developers mailing list > > * Please search the archive at > http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_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-developers > or send a mail to gmx-developers-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.
[gmx-users] parameter for repeating units with cgenff
Hi All, I am using cgenff to derive parameters for some polymer (A-...-C). The repeating block B has some Oxygen O atom which joins the next residue (-BO-). When I derive charges for B residue with cgenff then I get net -1 charge on B residue because of this unsatsified O atom. I want to keep B residue formal charge zero. I get formal charge zero if I derive charges on B-OH residue but removing H will again will give some non-intergral formal charge. I am not sure what should done in this regard. If someone has some experience with this or suggestion I will highly appreciate it. Thanks Jiomm -- 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] parameter for repeating units with cgenff
On 12/11/15 10:17 AM, gromacs query wrote: Hi All, I am using cgenff to derive parameters for some polymer (A-...-C). The repeating block B has some Oxygen O atom which joins the next residue (-BO-). When I derive charges for B residue with cgenff then I get net -1 charge on B residue because of this unsatsified O atom. I want to keep B residue formal charge zero. I get formal charge zero if I derive charges on B-OH residue but removing H will again will give some non-intergral formal charge. I am not sure what should done in this regard. If someone has some experience with this or suggestion I will highly appreciate it. To derive suitable monomer units, you have to include all the linking atoms; if it's an ether of some sort, you need B-O-CH3 or whatever is similar to then apply to the actual polymer itself. Just parametrizing B in the absence of flanking atoms is inappropriate. -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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] AMBER force fields, ff12SB and ff14SB for GROMACS
> Date: Thu, 10 Dec 2015 18:02:32 -0700 > From: Michael Shirts> To: Discussion list for GROMACS users > Cc: gromacs.org_gmx-users@maillist.sys.kth.se > Subject: Re: [gmx-users] AMBER force fields, ff12SB and ff14SB for > GROMACS > Message-ID: > > Content-Type: text/plain; charset=UTF-8 > > Great! It's important to have all data validated carefully. Have you > put together a set of tests, with coverage of all the atom types and > interaction types, showing that the molecules give the same energies > in both GROMACS and AMBER? Validating conversions is quite important > so that people can trust the results. I have considered every atoms, atom types, nonbonds and bonds interaction, atoms in every residues and nucleic acids to make sure they are the same as in AMBER. If some ones pick up any those things to compare with original one, I can make sure they will be the same. As you said the test is very important, but it takes a lot of time and I am working on this. It will be much better if other gromacs users like you participate the testing work. > > On Thu, Dec 10, 2015 at 4:01 PM, Man Hoang Viet > wrote: >> Dear GROMACS Admins, >> >> I have implemented AMBER force field ff12SB and ff14SB into GROMACS and >> want to share them with all gromacs users. I have tried to upload it via >> below link >> http://www.gromacs.org/Downloads/User_contributions/Force_fields >> However, I can not do register to upload. Could you please tell me how I >> can share the amber force fields to gromacs users. >> >> Thank you very much. >> >> Yours sincerely, >> -- >> Viet Man, >> >> Postdoctoral Research Associate, >> Physics Department, NCSU > Viet Man, Postdoctoral Research Associate, Physics Department, NCSU -- 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] Free energy of ice crystals (Thermodynamic Integration)
Hello, I would like to perform some free energy calculations of ice Ic (cubic ice) at various temperatures. However, I'm having difficulty finding information on how Gromacs can be used to do this. The goal is to gradually turn off VDW and Coulomb interactions, just as one would do in a liquid, but also turn on virtual spring-like potentials to create an Einstein crystal, which is where I'm having trouble. I know you can fix atom locations, and considering bond potentials are spring-like, my best idea is to create a virtual copy of each molecule (not virtual sites, but just atoms that have no charge and don't interact with other atoms/molecules), fix it in the crystal lattice, and create bonds between corresponding atoms with an equilibrium distance of 0. But this means I have to simulate twice as many atoms, and more importantly I don't know how to couple only some of a molecule's bonds with lambda. Am I even on the right track, or is there a more straightforward way to do this? Ar e there any tutorials for free energies of solids that I could take a look at? (I couldn't find any) Thanks for your help! N. H. -- 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] Free energy of ice crystals (Thermodynamic Integration)
IF it's NVT, then you should just be able to add harmonic potentials using the position restraints (not constraints) -- the harmonic restraints can be controlled by lambda. You should only need to alter the charge and VDW interactions -- no need to adjust the bonds. We have been doing some work on crystals of small organics-- email me off the list and we might be able to give you some example scripts. On Fri, Dec 11, 2015 at 8:46 PM, Nathan K Houtzwrote: > Hello, > > > I would like to perform some free energy calculations of ice Ic (cubic ice) > at various temperatures. However, I'm having difficulty finding information > on how Gromacs can be used to do this. The goal is to gradually turn off VDW > and Coulomb interactions, just as one would do in a liquid, but also turn on > virtual spring-like potentials to create an Einstein crystal, which is where > I'm having trouble. I know you can fix atom locations, and considering bond > potentials are spring-like, my best idea is to create a virtual copy of each > molecule (not virtual sites, but just atoms that have no charge and don't > interact with other atoms/molecules), fix it in the crystal lattice, and > create bonds between corresponding atoms with an equilibrium distance of 0. > But this means I have to simulate twice as many atoms, and more importantly I > don't know how to couple only some of a molecule's bonds with lambda. Am I > even on the right track, or is there a more straightforward way to do this? Ar > e there any tutorials for free energies of solids that I could take a look > at? (I couldn't find any) > > > Thanks for your help! > > N. H. > -- > 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] parameter for repeating units with cgenff
Hi Justin Thanks for the reply. Yes your are right and indeed, sorry I forgot to mention, I tried with some CH3 and other groups which may be considered as an appropriate 'cap' which is normally followed by many. But in this case I will have to remove the cap e.g. B-O-CH3 to B-O- before joining to the next residue. This causes a net non-integral charge on B-O- which is not correct as disturbs overall integral charge of the polymer. Ideally the cap once removed should leave B-O with integral charge. I have tried many caps but none of them after being removed leave B-O to be integral. best regards, Vishal On Fri, Dec 11, 2015 at 4:43 PM, Justin Lemkulwrote: > > > On 12/11/15 10:17 AM, gromacs query wrote: > >> Hi All, >> >> I am using cgenff to derive parameters for some polymer (A-...-C). The >> repeating block B has some Oxygen O atom which joins the next residue >> (-BO-). When I derive charges for B residue with cgenff then I get net -1 >> charge on B residue because of this unsatsified O atom. I want to keep B >> residue formal charge zero. I get formal charge zero if I derive charges >> on >> B-OH residue but removing H will again will give some non-intergral formal >> charge. I am not sure what should done in this regard. >> >> If someone has some experience with this or suggestion I will highly >> appreciate it. >> >> > To derive suitable monomer units, you have to include all the linking > atoms; if it's an ether of some sort, you need B-O-CH3 or whatever is > similar to then apply to the actual polymer itself. Just parametrizing B > in the absence of flanking atoms is inappropriate. > > -Justin > > -- > == > > Justin A. Lemkul, Ph.D. > Ruth L. Kirschstein NRSA Postdoctoral Fellow > > Department of Pharmaceutical Sciences > School of Pharmacy > Health Sciences Facility II, Room 629 > University of Maryland, Baltimore > 20 Penn St. > Baltimore, MD 21201 > > jalem...@outerbanks.umaryland.edu | (410) 706-7441 > http://mackerell.umaryland.edu/~jalemkul > > == > -- > 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] G53A6, non-symmetric selection metrix
Thanks Tom! Got it! now I know how to interpret this table. Great Cheers Mohsen On Fri, Dec 11, 2015 at 4:12 AM, Piggot T.wrote: > Hi Mohsen, > > This table can indeed be quite confusing at first. I suggest you take a > look at http://redmine.gromacs.org/issues/773#note-10 where there is a > discussion of how to interpret table 8 (for a specific example, but it > should highlight how it works). Hopefully that should answer your questions. > > Cheers > > Tom > > From: gromacs.org_gmx-users-boun...@maillist.sys.kth.se [ > gromacs.org_gmx-users-boun...@maillist.sys.kth.se] on behalf of Mohsen > Ramezanpour [ramezanpour.moh...@gmail.com] > Sent: 10 December 2015 23:42 > To: Discussion list for GROMACS users > Subject: [gmx-users] G53A6, non-symmetric selection metrix > > Dear All, > > Reading the parameters for Gromos 53A6 ff ( article by Oostenbrink *et al. > Journal of computational chemistry* 25.13 (2004): 1656-1676.), I got > confused about van der waals interactions between non-bonded atoms. > > for two atom types, we use combination rules to get (Cij 6) and (Cij 12), > and these combination rules (eq. 15 in article) make use of selection > matrix (table 8 in article). However this selection matrix is not > symmetric. > > This means L-J interactions between two atom types depends to the order we > put them in [pairs]. Is it correct? > > If this is true, then the order of atom numbers in [pair] matters. But it > does not make sense, why (Cij 12) should be different from (Cji 12)? > > Thanks in advance for your comments > > Best, > Mohsen > > -- > *Rewards work better than punishment ...* > -- > 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. > -- *Rewards work better than punishment ...* -- 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] parameter for repeating units with cgenff
On 12/11/15 12:17 PM, gromacs query wrote: Hi Justin Thanks for the reply. Yes your are right and indeed, sorry I forgot to mention, I tried with some CH3 and other groups which may be considered as an appropriate 'cap' which is normally followed by many. But in this case I will have to remove the cap e.g. B-O-CH3 to B-O- before joining to the next residue. This causes a net non-integral charge on B-O- which is not correct as disturbs overall integral charge of the polymer. Ideally the cap once removed should leave B-O with integral charge. I have tried many caps but none of them after being removed leave B-O to be integral. Capping groups can't simply be chopped off; that will naturally lead to a useless topology. They need to be designed such that their parameters (charges, etc) are analogous to the full-length species such that they can be merged with flanking groups. -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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] parameter for repeating units with cgenff
Hi Justin Thanks for the explanation. Just few things. 1) When you say " Initial charge fitting would be done on a side chain ..." you mean using some QM method? Which seems expensive and needs expertise which I dont have. 2) Also, as you said assigning charges and types with analogy does this mean I can hand pick types and charges looking (carefully) some residues available with CHARMM and can directly use them? At the same time, then in this case, I should satisfy the condition of overall integral charge. Am just wondering in future if reviewer raises question :) On Fri, Dec 11, 2015 at 6:33 PM, Justin Lemkulwrote: > > > On 12/11/15 1:28 PM, gromacs query wrote: > >> Hi Justin >> >> I am bit lost I think. For e.g. in the amino acid residues library we have >> -CO-CH(R)-NH- which we can combine these amino acids in any way and in the >> rtp file it defines only these atoms and connecting atoms are mentioned >> specifically. >> >> Now lets say I have some new amino acid -CO-CH(X)-NH- which I want to >> combine with some existing amino acids. If I want to derive charges from >> Cgenff what kind of residue I should prepare? >> Is it NH2-CO-CH(X)-NH-COOH or may be CH3-NH-CO-CH(X)-NH-COCH3. But >> ultimately I will be requiring only -CO-CH(X)-NH- part. So not considering >> charges for the caps in topology will give non-integral charge or in other >> words this part -CO-CH(X)-NH- will always be non-integral. Sorry am >> confused. >> >> > The backbone group always has a charge of zero, so you only deal with the > side chain when deriving charges. > > Initial charge fitting would be done on a side chain analog that > terminates in a -CH3 that is analogous to CB. Once that work is done, the > model compound is merged with the backbone and CB (which is -CH2, so > generally the third H charge is just lumped into the C charge) and > torsional fitting is done for chi1/chi2/etc using a dipeptide model of the > amino acid. But at that point the charges are done. > > I also wouldn't use CGenFF for a modified amino acid; derive parameters by > analogy from the parent CHARMM force field. CGenFF is generalized, at the > expense of some accuracy. Mixing CGenFF into a polypeptide chain is not > the most robust approach. CHARMM has excellent coverage for most chemical > moieties. Simply assigning charges and types by analogy should be quite > straightforward. > > > -Justin > > -- > == > > Justin A. Lemkul, Ph.D. > Ruth L. Kirschstein NRSA Postdoctoral Fellow > > Department of Pharmaceutical Sciences > School of Pharmacy > Health Sciences Facility II, Room 629 > University of Maryland, Baltimore > 20 Penn St. > Baltimore, MD 21201 > > jalem...@outerbanks.umaryland.edu | (410) 706-7441 > http://mackerell.umaryland.edu/~jalemkul > > == > -- > 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] parameter for repeating units with cgenff
On 12/11/15 2:20 PM, gromacs query wrote: Hi Justin Thanks for the explanation. Just few things. 1) When you say " Initial charge fitting would be done on a side chain ..." you mean using some QM method? Which seems expensive and needs expertise which I dont have. Yes, these are generally the initial target data (gas-phase dipole moment, water interactions, etc). Most of these calculations are done by optimizing the structure at MP2/6-31G*, which is very straightforward in, e.g. Gaussian. The CGenFF paper has all the details; this is, in essence, what is done for the biomolecular CHARMM force field, as well. 2) Also, as you said assigning charges and types with analogy does this mean I can hand pick types and charges looking (carefully) some residues available with CHARMM and can directly use them? At the same time, then in this case, I should satisfy the condition of overall integral charge. Am just wondering in future if reviewer raises question :) This is what is generally done. In fact, this is largely what CGenFF is doing, too. Initial selection of parameters based on rules that reference the library of existing compounds. Refinements are then done based on the initial topology. -Justin On Fri, Dec 11, 2015 at 6:33 PM, Justin Lemkulwrote: On 12/11/15 1:28 PM, gromacs query wrote: Hi Justin I am bit lost I think. For e.g. in the amino acid residues library we have -CO-CH(R)-NH- which we can combine these amino acids in any way and in the rtp file it defines only these atoms and connecting atoms are mentioned specifically. Now lets say I have some new amino acid -CO-CH(X)-NH- which I want to combine with some existing amino acids. If I want to derive charges from Cgenff what kind of residue I should prepare? Is it NH2-CO-CH(X)-NH-COOH or may be CH3-NH-CO-CH(X)-NH-COCH3. But ultimately I will be requiring only -CO-CH(X)-NH- part. So not considering charges for the caps in topology will give non-integral charge or in other words this part -CO-CH(X)-NH- will always be non-integral. Sorry am confused. The backbone group always has a charge of zero, so you only deal with the side chain when deriving charges. Initial charge fitting would be done on a side chain analog that terminates in a -CH3 that is analogous to CB. Once that work is done, the model compound is merged with the backbone and CB (which is -CH2, so generally the third H charge is just lumped into the C charge) and torsional fitting is done for chi1/chi2/etc using a dipeptide model of the amino acid. But at that point the charges are done. I also wouldn't use CGenFF for a modified amino acid; derive parameters by analogy from the parent CHARMM force field. CGenFF is generalized, at the expense of some accuracy. Mixing CGenFF into a polypeptide chain is not the most robust approach. CHARMM has excellent coverage for most chemical moieties. Simply assigning charges and types by analogy should be quite straightforward. -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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. -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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] parameter for repeating units with cgenff
Hi Justin I am bit lost I think. For e.g. in the amino acid residues library we have -CO-CH(R)-NH- which we can combine these amino acids in any way and in the rtp file it defines only these atoms and connecting atoms are mentioned specifically. Now lets say I have some new amino acid -CO-CH(X)-NH- which I want to combine with some existing amino acids. If I want to derive charges from Cgenff what kind of residue I should prepare? Is it NH2-CO-CH(X)-NH-COOH or may be CH3-NH-CO-CH(X)-NH-COCH3. But ultimately I will be requiring only -CO-CH(X)-NH- part. So not considering charges for the caps in topology will give non-integral charge or in other words this part -CO-CH(X)-NH- will always be non-integral. Sorry am confused. Thanks On Fri, Dec 11, 2015 at 5:50 PM, Justin Lemkulwrote: > > > On 12/11/15 12:17 PM, gromacs query wrote: > >> Hi Justin >> >> Thanks for the reply. Yes your are right and indeed, sorry I forgot to >> mention, I tried with some CH3 and other groups which may be considered as >> an appropriate 'cap' which is normally followed by many. But in this case >> I >> will have to remove the cap e.g. B-O-CH3 to B-O- before joining to the >> next >> residue. This causes a net non-integral charge on B-O- which is not >> correct >> as disturbs overall integral charge of the polymer. Ideally the cap once >> removed should leave B-O with integral charge. I have tried many caps but >> none of them after being removed leave B-O to be integral. >> >> > Capping groups can't simply be chopped off; that will naturally lead to a > useless topology. They need to be designed such that their parameters > (charges, etc) are analogous to the full-length species such that they can > be merged with flanking groups. > > > -Justin > > -- > == > > Justin A. Lemkul, Ph.D. > Ruth L. Kirschstein NRSA Postdoctoral Fellow > > Department of Pharmaceutical Sciences > School of Pharmacy > Health Sciences Facility II, Room 629 > University of Maryland, Baltimore > 20 Penn St. > Baltimore, MD 21201 > > jalem...@outerbanks.umaryland.edu | (410) 706-7441 > http://mackerell.umaryland.edu/~jalemkul > > == > -- > 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] parameter for repeating units with cgenff
On 12/11/15 1:28 PM, gromacs query wrote: Hi Justin I am bit lost I think. For e.g. in the amino acid residues library we have -CO-CH(R)-NH- which we can combine these amino acids in any way and in the rtp file it defines only these atoms and connecting atoms are mentioned specifically. Now lets say I have some new amino acid -CO-CH(X)-NH- which I want to combine with some existing amino acids. If I want to derive charges from Cgenff what kind of residue I should prepare? Is it NH2-CO-CH(X)-NH-COOH or may be CH3-NH-CO-CH(X)-NH-COCH3. But ultimately I will be requiring only -CO-CH(X)-NH- part. So not considering charges for the caps in topology will give non-integral charge or in other words this part -CO-CH(X)-NH- will always be non-integral. Sorry am confused. The backbone group always has a charge of zero, so you only deal with the side chain when deriving charges. Initial charge fitting would be done on a side chain analog that terminates in a -CH3 that is analogous to CB. Once that work is done, the model compound is merged with the backbone and CB (which is -CH2, so generally the third H charge is just lumped into the C charge) and torsional fitting is done for chi1/chi2/etc using a dipeptide model of the amino acid. But at that point the charges are done. I also wouldn't use CGenFF for a modified amino acid; derive parameters by analogy from the parent CHARMM force field. CGenFF is generalized, at the expense of some accuracy. Mixing CGenFF into a polypeptide chain is not the most robust approach. CHARMM has excellent coverage for most chemical moieties. Simply assigning charges and types by analogy should be quite straightforward. -Justin -- == Justin A. Lemkul, Ph.D. Ruth L. Kirschstein NRSA Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 629 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 http://mackerell.umaryland.edu/~jalemkul == -- 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.