Re: [gmx-users] Re: Gromacs-4.6 on two Titans GPUs
First, there is no value in ascribing problems to the hardware if the simulation setup is not yet balanced, or not large enough to provide enough atoms and long enough rlist to saturate the GPUs, etc. Look at the log files and see what complaints mdrun makes about things like PME load balance, and the times reported for different components of the simulation, because these must differ between the two runs you report. diff -y -W 160 *log |less is your friend. Some (non-GPU-specific) background information in part 5 here http://www.gromacs.org/Documentation/Tutorials/GROMACS_USA_Workshop_and_Conference_2013/Topology_preparation%2c_%22What's_in_a_log_file%22%2c_basic_performance_improvements%3a_Mark_Abraham%2c_Session_1A (though I recommend the PDF version) Mark On Thu, Nov 7, 2013 at 6:34 AM, James Starlight jmsstarli...@gmail.comwrote: I've gone to conclusion that simulation with 1 or 2 GPU simultaneously gave me the same performance mdrun -ntmpi 2 -ntomp 6 -gpu_id 01 -v -deffnm md_CaM_test, mdrun -ntmpi 2 -ntomp 6 -gpu_id 0 -v -deffnm md_CaM_test, Doest it be due to the small CPU cores or addition RAM ( this system has 32 gb) is needed ? OR may be some extra options are needed in the config? James 2013/11/6 Richard Broadbent richard.broadben...@imperial.ac.uk Hi Dwey, On 05/11/13 22:00, Dwey Kauffman wrote: Hi Szilard, Thanks for your suggestions. I am indeed aware of this page. In a 8-core AMD with 1GPU, I am very happy about its performance. See below. My intention is to obtain a even better one because we have multiple nodes. ### 8 core AMD with 1 GPU, Force evaluation time GPU/CPU: 4.006 ms/2.578 ms = 1.554 For optimal performance this ratio should be close to 1! NOTE: The GPU has 20% more load than the CPU. This imbalance causes performance loss, consider using a shorter cut-off and a finer PME grid. Core t (s) Wall t (s)(%) Time: 216205.51027036.812 799.7 7h30:36 (ns/day)(hour/ns) Performance: 31.9560.751 ### 8 core AMD with 2 GPUs Core t (s) Wall t (s)(%) Time: 178961.45022398.880 799.0 6h13:18 (ns/day)(hour/ns) Performance: 38.5730.622 Finished mdrun on node 0 Sat Jul 13 09:24:39 2013 I'm almost certain that Szilard meant the lines above this that give the breakdown of where the time is spent in the simulation. Richard However, in your case I suspect that the bottleneck is multi-threaded scaling on the AMD CPUs and you should probably decrease the number of threads per MPI rank and share GPUs between 2-4 ranks. OK but can you give a example of mdrun command ? given a 8 core AMD with 2 GPUs. I will try to run it again. Regarding scaling across nodes, you can't expect much from gigabit ethernet - especially not from the cheaper cards/switches, in my experience even reaction field runs don't scale across nodes with 10G ethernet if you have more than 4-6 ranks per node trying to communicate (let alone with PME). However, on infiniband clusters we have seen scaling to 100 atoms/core (at peak). From your comments, it sounds like a cluster of AMD cpus is difficult to scale across nodes in our current setup. Let's assume we install Infiniband (20 or 40GB/s) in the same system of 16 nodes of 8 core AMD with 1 GPU only. Considering the same AMD system, what is a good way to obtain better performance when we run a task across nodes ? in other words, what dose mudrun_mpi look like ? Thanks, Dwey -- View this message in context: http://gromacs.5086.x6.nabble. com/Gromacs-4-6-on-two-Titans-GPUs-tp5012186p5012279.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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 -- 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
Re: [gmx-users] nose-hoover vs v-rescale in implicit solvent
I think either is correct for practical purposes. Mark On Thu, Nov 7, 2013 at 8:41 AM, Gianluca Interlandi gianl...@u.washington.edu wrote: Does it make more sense to use nose-hoover or v-rescale when running in implicit solvent GBSA? I understand that this might be a matter of opinion. Thanks, Gianluca - Gianluca Interlandi, PhD gianl...@u.washington.edu +1 (206) 685 4435 http://artemide.bioeng.washington.edu/ Research Scientist at the Department of Bioengineering at the University of Washington, Seattle WA U.S.A. - -- 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 -- 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: single point calculation with gromacs
On Wed, Nov 6, 2013 at 4:07 PM, fantasticqhl fantastic...@gmail.com wrote: Dear Justin, I am sorry for the late reply. I still can't figure it out. It isn't rocket science - your two .mdp files describe totally different model physics. To compare things, change as few things as necessary to generate the comparison. So use the same input .mdp file for the MD vs EM single-point comparison, just changing the integrator line, and maybe unconstrained-start (I forget the details). And be aware of http://www.gromacs.org/Documentation/How-tos/Single-Point_Energy Mark Could you please send me the mdp file which was used for your single point calculations. I want to do some comparison and then solve the problem. Thanks very much! All the best, Qinghua -- View this message in context: http://gromacs.5086.x6.nabble.com/single-point-calculation-with-gromacs-tp5012084p5012295.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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
[gmx-users] Error in Umbrella sampling command
Dear Sir Presently I am working with the example file as given in the umbrella sampling tutorial. While running the following command grompp -f npt_umbrella.mdp -c conf0.gro -p topol.top -n index.ndx -o npt0.tpr I got the following error. How to debug this error. Ignoring obsolete mdp entry 'title' Back Off! I just backed up mdout.mdp to ./#mdout.mdp.5# NOTE 1 [file npt_umbrella.mdp]: nstcomm nstcalcenergy defeats the purpose of nstcalcenergy, setting nstcomm to nstcalcenergy NOTE 2 [file npt_umbrella.mdp]: leapfrog does not yet support Nose-Hoover chains, nhchainlength reset to 1 WARNING 1 [file npt_umbrella.mdp]: You are generating velocities so I am assuming you are equilibrating a system. You are using Parrinello-Rahman pressure coupling, but this can be unstable for equilibration. If your system crashes, try equilibrating first with Berendsen pressure coupling. If you are not equilibrating the system, you can probably ignore this warning. ERROR 1 [file npt_umbrella.mdp]: Generating velocities is inconsistent with attempting to continue a previous run. Choose only one of gen-vel = yes and continuation = yes. Generated 165 of the 1596 non-bonded parameter combinations Excluding 3 bonded neighbours molecule type 'Protein_chain_A' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_B' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_C' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_D' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_E' turning all bonds into constraints... Excluding 2 bonded neighbours molecule type 'SOL' turning all bonds into constraints... Excluding 1 bonded neighbours molecule type 'NA' turning all bonds into constraints... Excluding 1 bonded neighbours molecule type 'CL' turning all bonds into constraints... Velocities were taken from a Maxwell distribution at 300 K There were 2 notes There was 1 warning --- Program grompp_mpi_d, VERSION 4.6.3 Source code file: /opt/apps/GROMACS/GROMACS-SOURCE/gromacs-4.6.3/src/kernel/gromp p.c, line: 1593 Fatal error: There was 1 error in input file(s) For more information and tips for troubleshooting, please check the GROMACS website at http://www.gromacs.org/Documentation/Errors I request you to kindly help me to debug the error Regards Arunima -- 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] Error in Umbrella sampling command
On 11/7/13 6:27 AM, Arunima Shilpi wrote: Dear Sir Presently I am working with the example file as given in the umbrella sampling tutorial. While running the following command grompp -f npt_umbrella.mdp -c conf0.gro -p topol.top -n index.ndx -o npt0.tpr I got the following error. How to debug this error. Ignoring obsolete mdp entry 'title' Back Off! I just backed up mdout.mdp to ./#mdout.mdp.5# NOTE 1 [file npt_umbrella.mdp]: nstcomm nstcalcenergy defeats the purpose of nstcalcenergy, setting nstcomm to nstcalcenergy NOTE 2 [file npt_umbrella.mdp]: leapfrog does not yet support Nose-Hoover chains, nhchainlength reset to 1 WARNING 1 [file npt_umbrella.mdp]: You are generating velocities so I am assuming you are equilibrating a system. You are using Parrinello-Rahman pressure coupling, but this can be unstable for equilibration. If your system crashes, try equilibrating first with Berendsen pressure coupling. If you are not equilibrating the system, you can probably ignore this warning. ERROR 1 [file npt_umbrella.mdp]: Generating velocities is inconsistent with attempting to continue a previous run. Choose only one of gen-vel = yes and continuation = yes. Either the run is starting for the first time (gen_vel = yes and continuation = no) or it is a continuation (gen_vel = no and continuation = yes). In this case, set continuation = no since it is the first run. -Justin Generated 165 of the 1596 non-bonded parameter combinations Excluding 3 bonded neighbours molecule type 'Protein_chain_A' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_B' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_C' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_D' turning all bonds into constraints... Excluding 3 bonded neighbours molecule type 'Protein_chain_E' turning all bonds into constraints... Excluding 2 bonded neighbours molecule type 'SOL' turning all bonds into constraints... Excluding 1 bonded neighbours molecule type 'NA' turning all bonds into constraints... Excluding 1 bonded neighbours molecule type 'CL' turning all bonds into constraints... Velocities were taken from a Maxwell distribution at 300 K There were 2 notes There was 1 warning --- Program grompp_mpi_d, VERSION 4.6.3 Source code file: /opt/apps/GROMACS/GROMACS-SOURCE/gromacs-4.6.3/src/kernel/gromp p.c, line: 1593 Fatal error: There was 1 error in input file(s) For more information and tips for troubleshooting, please check the GROMACS website at http://www.gromacs.org/Documentation/Errors I request you to kindly help me to debug the error Regards Arunima -- == Justin A. Lemkul, Ph.D. Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 601 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 == -- 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
[gmx-users] Re: CHARMM .mdp settings for GPU
Dear All, Any suggestions? Thank you. -- View this message in context: http://gromacs.5086.x6.nabble.com/CHARMM-mdp-settings-for-GPU-tp5012267p5012316.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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
[gmx-users] DSSP output
Hi ALL, Is there any way to get the percentage of each secondary structural content of a protein using do_dssp if I supply a single PDB to it? And how to plot the data of the -sc output from do_dssp? Any suggestion is welcome. Regards, Anirban -- 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] DSSP output
On 11/7/13 8:24 AM, Anirban wrote: Hi ALL, Is there any way to get the percentage of each secondary structural content of a protein using do_dssp if I supply a single PDB to it? The output of scount.xvg has the percentages, but it's also trivial to do it for one snapshot. The contents of scount.xvg are the number of residues present in each type of secondary structure, and you know the total number of residues... And how to plot the data of the -sc output from do_dssp? Like any multiple data set. xmgrace -nxy -Justin -- == Justin A. Lemkul, Ph.D. Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 601 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 == -- 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
[gmx-users] Problem compiling Gromacs 4.6.3 with CUDA
Hi, I'm having trouble compiling v 4.6.3 with GPU support using CUDA 5.5.22. The configuration runs okay and I have made sure that I have set paths correctly. I'm getting errors: $ make [ 0%] Building NVCC (Device) object src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o icc: command line warning #10006: ignoring unknown option '-dumpspecs' /usr/lib/gcc/x86_64-redhat-linux/4.4.7/../../../../lib64/crt1.o: In function `_start': (.text+0x20): undefined reference to `main' CMake Error at cuda_tools_generated_pmalloc_cuda.cu.o.cmake:206 (message): Error generating /apps/src/gromacs/gromacs-4.6.3/src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o make[2]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/./cuda_tools_generated_pmalloc_cuda.cu.o] Error 1 make[1]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/all] Error 2 make: *** [all] Error 2 Any help would be appreciated. Regards, Ahmed. -- Scanned by iCritical. -- 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: CHARMM .mdp settings for GPU
Hi, It's not easy to be explicit. CHARMM wasn't parameterized with PME, so the original paper's coulomb settings can be taken with a grain of salt for use with PME - others' success in practice should be a guideline here. The good news is that the default GROMACS PME settings are pretty good for at least some problems (http://pubs.acs.org/doi/abs/10.1021/ct4005068), and the GPU auto-tuning of parameters in 4.6 is designed to preserve the right sorts of things. LJ is harder because it would make good sense to preserve the way CHARMM did it, but IIRC you can't use something equivalent to the CHARMM LJ shift with the Verlet kernels, either natively or with a table. We hope to fix that in 5.0, but code is not written yet. I would probably use vdwtype = cut-off, vdw-modifier = potential-shift-verlet and rcoulomb=rlist=rvdw=1.2, but I don't run CHARMM simulations for a living ;-) Mark On Thu, Nov 7, 2013 at 1:42 PM, Rajat Desikan rajatdesi...@gmail.comwrote: Dear All, Any suggestions? Thank you. -- View this message in context: http://gromacs.5086.x6.nabble.com/CHARMM-mdp-settings-for-GPU-tp5012267p5012316.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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
[gmx-users] installing gromacs 4.6.1 with openmpi
Dear gromacs users I have installed gromacs 4.6.1 with cmake 2.8.12, fftw3.3.3 and openmpi-1.6.4 on a single machine with 8 cores(Red Hat Enterprise linux 6.1) . During openmpi installation ( I used make -jN) and also in gromacs installation ( I used make -j N command), everything seemed ok but when I want to use mpirun -np N mdrun I face this error: mpiexec failed: gethostbyname_ex failed for Bioinf2 (I can run mdrun with just one cpu).Any suggestion would be highly appreciated. thanks in advance, Niloofar -- 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] Problem compiling Gromacs 4.6.3 with CUDA
icc and CUDA is pretty painful. I'd suggest getting latest gcc. Mark On Thu, Nov 7, 2013 at 2:42 PM, ahmed.sa...@stfc.ac.uk wrote: Hi, I'm having trouble compiling v 4.6.3 with GPU support using CUDA 5.5.22. The configuration runs okay and I have made sure that I have set paths correctly. I'm getting errors: $ make [ 0%] Building NVCC (Device) object src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o icc: command line warning #10006: ignoring unknown option '-dumpspecs' /usr/lib/gcc/x86_64-redhat-linux/4.4.7/../../../../lib64/crt1.o: In function `_start': (.text+0x20): undefined reference to `main' CMake Error at cuda_tools_generated_pmalloc_cuda.cu.o.cmake:206 (message): Error generating /apps/src/gromacs/gromacs-4.6.3/src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o make[2]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/./cuda_tools_generated_pmalloc_cuda.cu.o] Error 1 make[1]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/all] Error 2 make: *** [all] Error 2 Any help would be appreciated. Regards, Ahmed. -- Scanned by iCritical. -- 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 -- 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] installing gromacs 4.6.1 with openmpi
Sounds like a non-GROMACS problem. I think you should explore configuring OpenMPI correctly, and show you can run an MPI test program successfully. Mark On Thu, Nov 7, 2013 at 5:51 PM, niloofar niknam niloofae_nik...@yahoo.comwrote: Dear gromacs users I have installed gromacs 4.6.1 with cmake 2.8.12, fftw3.3.3 and openmpi-1.6.4 on a single machine with 8 cores(Red Hat Enterprise linux 6.1) . During openmpi installation ( I used make -jN) and also in gromacs installation ( I used make -j N command), everything seemed ok but when I want to use mpirun -np N mdrun I face this error: mpiexec failed: gethostbyname_ex failed for Bioinf2 (I can run mdrun with just one cpu).Any suggestion would be highly appreciated. thanks in advance, Niloofar -- 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 -- 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
[gmx-users] Re: choosing force field
My protein contains metal ions which are parameterized only in gromos force field. Since I am a newbie to MD simulations, it would be difficult for me to parameterize those myself. Can you please guide me as per my previous mail which out of the two simulations should I consider more reliable-43a1 or 53a7? Thanks in advance. -- View this message in context: http://gromacs.5086.x6.nabble.com/choosing-force-field-tp5012242p5012322.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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: choosing force field
On 11/7/13 12:14 PM, pratibha wrote: My protein contains metal ions which are parameterized only in gromos force field. Since I am a newbie to MD simulations, it would be difficult for me to parameterize those myself. Can you please guide me as per my previous mail which out of the two simulations should I consider more reliable-43a1 or 53a7? AFAIK, there is no such thing as 53A7, and your original message was full of similar typos, making it nearly impossible to figure out what you were actually doing. Can you indicate the actual force field(s) that you have been using in case someone has any ideas? The difference between 53A6 and 54A7 should be quite pronounced, in my experience, thus any guesses as to what 53A7 should be doing are not productive because I don't know what that is. -Justin -- == Justin A. Lemkul, Ph.D. Postdoctoral Fellow Department of Pharmaceutical Sciences School of Pharmacy Health Sciences Facility II, Room 601 University of Maryland, Baltimore 20 Penn St. Baltimore, MD 21201 jalem...@outerbanks.umaryland.edu | (410) 706-7441 == -- 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
[gmx-users] Re: LIE method with PME
Hello I performed MD simulations of several Protein-ligand complexes and solvated Ligands using PME for log range electrostatics. I want to calculate the binding free energy using the LIE method, but when using g_energy I only get Coul-SR. How can I deal with Ligand-environment long range electrostatic interaction using gromacs? I have seen other discussion lists but I couldn't arrive to a solution. Could you please help me? Thank you Williams -- 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: CHARMM .mdp settings for GPU
Thank you, Mark. I think that running it on CPUs is a safer choice at present. On Thu, Nov 7, 2013 at 9:41 PM, Mark Abraham mark.j.abra...@gmail.comwrote: Hi, It's not easy to be explicit. CHARMM wasn't parameterized with PME, so the original paper's coulomb settings can be taken with a grain of salt for use with PME - others' success in practice should be a guideline here. The good news is that the default GROMACS PME settings are pretty good for at least some problems (http://pubs.acs.org/doi/abs/10.1021/ct4005068), and the GPU auto-tuning of parameters in 4.6 is designed to preserve the right sorts of things. LJ is harder because it would make good sense to preserve the way CHARMM did it, but IIRC you can't use something equivalent to the CHARMM LJ shift with the Verlet kernels, either natively or with a table. We hope to fix that in 5.0, but code is not written yet. I would probably use vdwtype = cut-off, vdw-modifier = potential-shift-verlet and rcoulomb=rlist=rvdw=1.2, but I don't run CHARMM simulations for a living ;-) Mark On Thu, Nov 7, 2013 at 1:42 PM, Rajat Desikan rajatdesi...@gmail.com wrote: Dear All, Any suggestions? Thank you. -- View this message in context: http://gromacs.5086.x6.nabble.com/CHARMM-mdp-settings-for-GPU-tp5012267p5012316.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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 -- Rajat Desikan (Ph.D Scholar) Prof. K. Ganapathy Ayappa's Lab (no 13), Dept. of Chemical Engineering, Indian Institute of Science, Bangalore -- 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: CHARMM .mdp settings for GPU
Reasonable, but CPU-only is not 100% conforming either; IIRC the CHARMM switch differs from the GROMACS switch (Justin linked a paper here with the CHARMM switch description a month or so back, but I don't have that link to hand). Mark On Thu, Nov 7, 2013 at 8:45 PM, rajat desikan rajatdesi...@gmail.comwrote: Thank you, Mark. I think that running it on CPUs is a safer choice at present. On Thu, Nov 7, 2013 at 9:41 PM, Mark Abraham mark.j.abra...@gmail.com wrote: Hi, It's not easy to be explicit. CHARMM wasn't parameterized with PME, so the original paper's coulomb settings can be taken with a grain of salt for use with PME - others' success in practice should be a guideline here. The good news is that the default GROMACS PME settings are pretty good for at least some problems (http://pubs.acs.org/doi/abs/10.1021/ct4005068), and the GPU auto-tuning of parameters in 4.6 is designed to preserve the right sorts of things. LJ is harder because it would make good sense to preserve the way CHARMM did it, but IIRC you can't use something equivalent to the CHARMM LJ shift with the Verlet kernels, either natively or with a table. We hope to fix that in 5.0, but code is not written yet. I would probably use vdwtype = cut-off, vdw-modifier = potential-shift-verlet and rcoulomb=rlist=rvdw=1.2, but I don't run CHARMM simulations for a living ;-) Mark On Thu, Nov 7, 2013 at 1:42 PM, Rajat Desikan rajatdesi...@gmail.com wrote: Dear All, Any suggestions? Thank you. -- View this message in context: http://gromacs.5086.x6.nabble.com/CHARMM-mdp-settings-for-GPU-tp5012267p5012316.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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 -- Rajat Desikan (Ph.D Scholar) Prof. K. Ganapathy Ayappa's Lab (no 13), Dept. of Chemical Engineering, Indian Institute of Science, Bangalore -- 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 -- 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: LIE method with PME
If the long-range component of your electrostatics model is not decomposable by group (which it isn't), then you can't use that with LIE. See the hundreds of past threads on this topic :-) Mark On Thu, Nov 7, 2013 at 8:34 PM, Williams Ernesto Miranda Delgado wmira...@fbio.uh.cu wrote: Hello I performed MD simulations of several Protein-ligand complexes and solvated Ligands using PME for log range electrostatics. I want to calculate the binding free energy using the LIE method, but when using g_energy I only get Coul-SR. How can I deal with Ligand-environment long range electrostatic interaction using gromacs? I have seen other discussion lists but I couldn't arrive to a solution. Could you please help me? Thank you Williams -- 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 -- 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] Problem compiling Gromacs 4.6.3 with CUDA
Did it a few days ago. Not so much of a problem here. But I compiled everything, including fftw, with it. The only error I got was that I should turn off the separable compilation, and that the user must be in the group video. My settings are (yes, I know it should go better with openmp, but openmp goes horrobly in our cluster, I don't know why): setenv CC /opt/intel/bin/icc setenv CXX /opt/intel/bin/icpc setenv F77 /opt/intel/bin/ifort setenv CMAKE_PREFIX_PATH /storage/home/johannes/lib/fftw/vanilla/ mkdir build cd build cmake .. -DGMX_GPU=ON -DCUDA_SEPARABLE_COMPILATION=OFF -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda -DGMX_OPENMP=OFF -DGMX_MPI=ON -DGMX_THREAD_MPI=OFF -DMPIEXEC_MAX_NUMPROCS=1024 -DBUILD_SHARED_LIBS=OFF -DGMX_PREFER_STATIC_LIBS=ON -DCMAKE_INSTALL_PREFIX=/storage/home/johannes/bin/gromacs/vanilla/ make make install cd .. rm -rf build On Thu, Nov 7, 2013 at 3:02 PM, Mark Abraham mark.j.abra...@gmail.comwrote: icc and CUDA is pretty painful. I'd suggest getting latest gcc. Mark On Thu, Nov 7, 2013 at 2:42 PM, ahmed.sa...@stfc.ac.uk wrote: Hi, I'm having trouble compiling v 4.6.3 with GPU support using CUDA 5.5.22. The configuration runs okay and I have made sure that I have set paths correctly. I'm getting errors: $ make [ 0%] Building NVCC (Device) object src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o icc: command line warning #10006: ignoring unknown option '-dumpspecs' /usr/lib/gcc/x86_64-redhat-linux/4.4.7/../../../../lib64/crt1.o: In function `_start': (.text+0x20): undefined reference to `main' CMake Error at cuda_tools_generated_pmalloc_cuda.cu.o.cmake:206 (message): Error generating /apps/src/gromacs/gromacs-4.6.3/src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o make[2]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/./cuda_tools_generated_pmalloc_cuda.cu.o] Error 1 make[1]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/all] Error 2 make: *** [all] Error 2 Any help would be appreciated. Regards, Ahmed. -- Scanned by iCritical. -- 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 -- 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 -- 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: CHARMM .mdp settings for GPU
Hi Mark! I think that this is the paper that you are referring to: dx.doi.org/10.1021/ct900549r Also for your reference, these are the settings that Justin recommended using with CHARMM in gromacs: vdwtype = switch rlist = 1.2 rlistlong = 1.4 rvdw = 1.2 rvdw-switch = 1.0 rcoulomb = 1.2 As you mention the switch function in gromacs is different than in CHARMM but it appears that the difference is very small. Gianluca On Thu, 7 Nov 2013, Mark Abraham wrote: Reasonable, but CPU-only is not 100% conforming either; IIRC the CHARMM switch differs from the GROMACS switch (Justin linked a paper here with the CHARMM switch description a month or so back, but I don't have that link to hand). Mark On Thu, Nov 7, 2013 at 8:45 PM, rajat desikan rajatdesi...@gmail.comwrote: Thank you, Mark. I think that running it on CPUs is a safer choice at present. On Thu, Nov 7, 2013 at 9:41 PM, Mark Abraham mark.j.abra...@gmail.com wrote: Hi, It's not easy to be explicit. CHARMM wasn't parameterized with PME, so the original paper's coulomb settings can be taken with a grain of salt for use with PME - others' success in practice should be a guideline here. The good news is that the default GROMACS PME settings are pretty good for at least some problems (http://pubs.acs.org/doi/abs/10.1021/ct4005068), and the GPU auto-tuning of parameters in 4.6 is designed to preserve the right sorts of things. LJ is harder because it would make good sense to preserve the way CHARMM did it, but IIRC you can't use something equivalent to the CHARMM LJ shift with the Verlet kernels, either natively or with a table. We hope to fix that in 5.0, but code is not written yet. I would probably use vdwtype = cut-off, vdw-modifier = potential-shift-verlet and rcoulomb=rlist=rvdw=1.2, but I don't run CHARMM simulations for a living ;-) Mark On Thu, Nov 7, 2013 at 1:42 PM, Rajat Desikan rajatdesi...@gmail.com wrote: Dear All, Any suggestions? Thank you. -- View this message in context: http://gromacs.5086.x6.nabble.com/CHARMM-mdp-settings-for-GPU-tp5012267p5012316.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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 -- Rajat Desikan (Ph.D Scholar) Prof. K. Ganapathy Ayappa's Lab (no 13), Dept. of Chemical Engineering, Indian Institute of Science, Bangalore -- 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 -- 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 - Gianluca Interlandi, PhD gianl...@u.washington.edu +1 (206) 685 4435 http://artemide.bioeng.washington.edu/ Research Scientist at the Department of Bioengineering at the University of Washington, Seattle WA U.S.A. - -- 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] Problem compiling Gromacs 4.6.3 with CUDA
You will do much better with gcc+openmp than icc-openmp! Mark On Thu, Nov 7, 2013 at 9:17 PM, Jones de Andrade johanne...@gmail.comwrote: Did it a few days ago. Not so much of a problem here. But I compiled everything, including fftw, with it. The only error I got was that I should turn off the separable compilation, and that the user must be in the group video. My settings are (yes, I know it should go better with openmp, but openmp goes horrobly in our cluster, I don't know why): setenv CC /opt/intel/bin/icc setenv CXX /opt/intel/bin/icpc setenv F77 /opt/intel/bin/ifort setenv CMAKE_PREFIX_PATH /storage/home/johannes/lib/fftw/vanilla/ mkdir build cd build cmake .. -DGMX_GPU=ON -DCUDA_SEPARABLE_COMPILATION=OFF -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda -DGMX_OPENMP=OFF -DGMX_MPI=ON -DGMX_THREAD_MPI=OFF -DMPIEXEC_MAX_NUMPROCS=1024 -DBUILD_SHARED_LIBS=OFF -DGMX_PREFER_STATIC_LIBS=ON -DCMAKE_INSTALL_PREFIX=/storage/home/johannes/bin/gromacs/vanilla/ make make install cd .. rm -rf build On Thu, Nov 7, 2013 at 3:02 PM, Mark Abraham mark.j.abra...@gmail.com wrote: icc and CUDA is pretty painful. I'd suggest getting latest gcc. Mark On Thu, Nov 7, 2013 at 2:42 PM, ahmed.sa...@stfc.ac.uk wrote: Hi, I'm having trouble compiling v 4.6.3 with GPU support using CUDA 5.5.22. The configuration runs okay and I have made sure that I have set paths correctly. I'm getting errors: $ make [ 0%] Building NVCC (Device) object src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o icc: command line warning #10006: ignoring unknown option '-dumpspecs' /usr/lib/gcc/x86_64-redhat-linux/4.4.7/../../../../lib64/crt1.o: In function `_start': (.text+0x20): undefined reference to `main' CMake Error at cuda_tools_generated_pmalloc_cuda.cu.o.cmake:206 (message): Error generating /apps/src/gromacs/gromacs-4.6.3/src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o make[2]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/./cuda_tools_generated_pmalloc_cuda.cu.o] Error 1 make[1]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/all] Error 2 make: *** [all] Error 2 Any help would be appreciated. Regards, Ahmed. -- Scanned by iCritical. -- 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 -- 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 -- 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 -- 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
mdrun on 8-core AMD + GTX TITAN (was: Re: [gmx-users] Re: Gromacs-4.6 on two Titans GPUs)
Let's not hijack James' thread as your hardware is different from his. On Tue, Nov 5, 2013 at 11:00 PM, Dwey Kauffman mpi...@gmail.com wrote: Hi Szilard, Thanks for your suggestions. I am indeed aware of this page. In a 8-core AMD with 1GPU, I am very happy about its performance. See below. My Actually, I was jumping to conclusions too early, as you mentioned AMD cluster, I assumed you must have 12-16-core Opteron CPUs. If you have an 8-core (desktop?) AMD CPU, than you may not need to run more than one rank per GPU. intention is to obtain a even better one because we have multiple nodes. Btw, I'm not sure it's an economically viable solution to install Infiniband network - especially if you have desktop-class machines. Such a network will end up costing $500 per machine just for a single network card, let alone cabling and switches. ### 8 core AMD with 1 GPU, Force evaluation time GPU/CPU: 4.006 ms/2.578 ms = 1.554 For optimal performance this ratio should be close to 1! NOTE: The GPU has 20% more load than the CPU. This imbalance causes performance loss, consider using a shorter cut-off and a finer PME grid. Core t (s) Wall t (s)(%) Time: 216205.51027036.812 799.7 7h30:36 (ns/day)(hour/ns) Performance: 31.9560.751 ### 8 core AMD with 2 GPUs Core t (s) Wall t (s)(%) Time: 178961.45022398.880 799.0 6h13:18 (ns/day)(hour/ns) Performance: 38.5730.622 Finished mdrun on node 0 Sat Jul 13 09:24:39 2013 Indeed, as Richard pointed out, I was asking for *full* logs, these summaries can't tell much, the table above the summary entitled R E A L C Y C L E A N D T I M E A C C O U N T I N G as well as other reported information across the log file is what I need to make an assessment of your simulations' performance. However, in your case I suspect that the bottleneck is multi-threaded scaling on the AMD CPUs and you should probably decrease the number of threads per MPI rank and share GPUs between 2-4 ranks. OK but can you give a example of mdrun command ? given a 8 core AMD with 2 GPUs. I will try to run it again. You could try running mpirun -np 4 mdrun -ntomp 2 -gpu_id 0011 but I suspect this won't help because your scaling issue Regarding scaling across nodes, you can't expect much from gigabit ethernet - especially not from the cheaper cards/switches, in my experience even reaction field runs don't scale across nodes with 10G ethernet if you have more than 4-6 ranks per node trying to communicate (let alone with PME). However, on infiniband clusters we have seen scaling to 100 atoms/core (at peak). From your comments, it sounds like a cluster of AMD cpus is difficult to scale across nodes in our current setup. Let's assume we install Infiniband (20 or 40GB/s) in the same system of 16 nodes of 8 core AMD with 1 GPU only. Considering the same AMD system, what is a good way to obtain better performance when we run a task across nodes ? in other words, what dose mudrun_mpi look like ? Thanks, Dwey -- View this message in context: http://gromacs.5086.x6.nabble.com/Gromacs-4-6-on-two-Titans-GPUs-tp5012186p5012279.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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: Gromacs-4.6 on two Titans GPUs
On Thu, Nov 7, 2013 at 6:34 AM, James Starlight jmsstarli...@gmail.com wrote: I've gone to conclusion that simulation with 1 or 2 GPU simultaneously gave me the same performance mdrun -ntmpi 2 -ntomp 6 -gpu_id 01 -v -deffnm md_CaM_test, mdrun -ntmpi 2 -ntomp 6 -gpu_id 0 -v -deffnm md_CaM_test, Doest it be due to the small CPU cores or addition RAM ( this system has 32 gb) is needed ? OR may be some extra options are needed in the config? GROMACS does not really need much (or fast) ram and it's most probably not configuration settings that are causing the lack of scaling. Given your setup my guess is that your hardware for your system is simply imbalanced to be efficiently used in GROMACS runs. Please post *full* log files (FYI use e.g. http://pastebin.com), that will help explain what is going on. James 2013/11/6 Richard Broadbent richard.broadben...@imperial.ac.uk Hi Dwey, On 05/11/13 22:00, Dwey Kauffman wrote: Hi Szilard, Thanks for your suggestions. I am indeed aware of this page. In a 8-core AMD with 1GPU, I am very happy about its performance. See below. My intention is to obtain a even better one because we have multiple nodes. ### 8 core AMD with 1 GPU, Force evaluation time GPU/CPU: 4.006 ms/2.578 ms = 1.554 For optimal performance this ratio should be close to 1! NOTE: The GPU has 20% more load than the CPU. This imbalance causes performance loss, consider using a shorter cut-off and a finer PME grid. Core t (s) Wall t (s)(%) Time: 216205.51027036.812 799.7 7h30:36 (ns/day)(hour/ns) Performance: 31.9560.751 ### 8 core AMD with 2 GPUs Core t (s) Wall t (s)(%) Time: 178961.45022398.880 799.0 6h13:18 (ns/day)(hour/ns) Performance: 38.5730.622 Finished mdrun on node 0 Sat Jul 13 09:24:39 2013 I'm almost certain that Szilard meant the lines above this that give the breakdown of where the time is spent in the simulation. Richard However, in your case I suspect that the bottleneck is multi-threaded scaling on the AMD CPUs and you should probably decrease the number of threads per MPI rank and share GPUs between 2-4 ranks. OK but can you give a example of mdrun command ? given a 8 core AMD with 2 GPUs. I will try to run it again. Regarding scaling across nodes, you can't expect much from gigabit ethernet - especially not from the cheaper cards/switches, in my experience even reaction field runs don't scale across nodes with 10G ethernet if you have more than 4-6 ranks per node trying to communicate (let alone with PME). However, on infiniband clusters we have seen scaling to 100 atoms/core (at peak). From your comments, it sounds like a cluster of AMD cpus is difficult to scale across nodes in our current setup. Let's assume we install Infiniband (20 or 40GB/s) in the same system of 16 nodes of 8 core AMD with 1 GPU only. Considering the same AMD system, what is a good way to obtain better performance when we run a task across nodes ? in other words, what dose mudrun_mpi look like ? Thanks, Dwey -- View this message in context: http://gromacs.5086.x6.nabble. com/Gromacs-4-6-on-two-Titans-GPUs-tp5012186p5012279.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- 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 -- 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 -- 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
[gmx-users] Re: LIE method with PME
Thank you Mark What do you think about making a rerun on the trajectories generated previously with PME but this time using coulombtype: cut-off? Could you suggest a cut off value? Thanks again Williams -- 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
[gmx-users] Question about make_ndx and g_angle
Dear Users, I am using openSUSE 12.3 and try to use make_ndx and g_angle. When I try the following command, there is an error message. ./make.ndx -f data.pdb ./make_ndx: error while loading shared libraries: libcudart.so.4:cannot open shared object file: No such file or directory Do I need cuda library in order to use make_ndx and g_angle ? Thanks. Best regards, Changwoon Jang -- 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: LIE method with PME
I'd at least use RF! Use a cut-off consistent with the force field parameterization. And hope the LIE correlates with reality! Mark On Nov 7, 2013 10:39 PM, Williams Ernesto Miranda Delgado wmira...@fbio.uh.cu wrote: Thank you Mark What do you think about making a rerun on the trajectories generated previously with PME but this time using coulombtype: cut-off? Could you suggest a cut off value? Thanks again Williams -- 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 -- 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] Problem compiling Gromacs 4.6.3 with CUDA
Really? An what about gcc+mpi? should I expect any improvement? On Thu, Nov 7, 2013 at 6:51 PM, Mark Abraham mark.j.abra...@gmail.comwrote: You will do much better with gcc+openmp than icc-openmp! Mark On Thu, Nov 7, 2013 at 9:17 PM, Jones de Andrade johanne...@gmail.com wrote: Did it a few days ago. Not so much of a problem here. But I compiled everything, including fftw, with it. The only error I got was that I should turn off the separable compilation, and that the user must be in the group video. My settings are (yes, I know it should go better with openmp, but openmp goes horrobly in our cluster, I don't know why): setenv CC /opt/intel/bin/icc setenv CXX /opt/intel/bin/icpc setenv F77 /opt/intel/bin/ifort setenv CMAKE_PREFIX_PATH /storage/home/johannes/lib/fftw/vanilla/ mkdir build cd build cmake .. -DGMX_GPU=ON -DCUDA_SEPARABLE_COMPILATION=OFF -DCUDA_TOOLKIT_ROOT_DIR=/usr/local/cuda -DGMX_OPENMP=OFF -DGMX_MPI=ON -DGMX_THREAD_MPI=OFF -DMPIEXEC_MAX_NUMPROCS=1024 -DBUILD_SHARED_LIBS=OFF -DGMX_PREFER_STATIC_LIBS=ON -DCMAKE_INSTALL_PREFIX=/storage/home/johannes/bin/gromacs/vanilla/ make make install cd .. rm -rf build On Thu, Nov 7, 2013 at 3:02 PM, Mark Abraham mark.j.abra...@gmail.com wrote: icc and CUDA is pretty painful. I'd suggest getting latest gcc. Mark On Thu, Nov 7, 2013 at 2:42 PM, ahmed.sa...@stfc.ac.uk wrote: Hi, I'm having trouble compiling v 4.6.3 with GPU support using CUDA 5.5.22. The configuration runs okay and I have made sure that I have set paths correctly. I'm getting errors: $ make [ 0%] Building NVCC (Device) object src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o icc: command line warning #10006: ignoring unknown option '-dumpspecs' /usr/lib/gcc/x86_64-redhat-linux/4.4.7/../../../../lib64/crt1.o: In function `_start': (.text+0x20): undefined reference to `main' CMake Error at cuda_tools_generated_pmalloc_cuda.cu.o.cmake:206 (message): Error generating /apps/src/gromacs/gromacs-4.6.3/src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir//./cuda_tools_generated_pmalloc_cuda.cu.o make[2]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/./cuda_tools_generated_pmalloc_cuda.cu.o] Error 1 make[1]: *** [src/gmxlib/cuda_tools/CMakeFiles/cuda_tools.dir/all] Error 2 make: *** [all] Error 2 Any help would be appreciated. Regards, Ahmed. -- Scanned by iCritical. -- 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 -- 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 -- 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 -- 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 -- 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
[gmx-users] Re: CHARMM .mdp settings for GPU
Dear All, The setting that I mentioned above are from Klauda et al., for a POPE membrane system. They can be found in charmm_npt.mdp in lipidbook (link below) http://lipidbook.bioch.ox.ac.uk/package/show/id/48.html Is there any reason not to use their .mdp parameters for a membrane-protein system? Justin's recommendation is highly valued since I am using his forcefield. Justin, your comments please To summarize: Klauda et al., suggest rlist = 1.0 rlistlong= 1.4 rvdw_switch = 0.8 vdwtype= Switch coulombtype = pme DispCorr= EnerPres ;only usefull with reaction-field and pme or pppm rcoulomb = 1.0 rcoulomb_switch= 0.0 rvdw = 1.2 Justin's recommendation (per mail above) vdwtype = switch rlist = 1.2 rlistlong = 1.4 rvdw = 1.2 rvdw-switch = 1.0 rcoulomb = 1.2 On Fri, Nov 8, 2013 at 2:20 AM, Gianluca Interlandi [via GROMACS] ml-node+s5086n5012329...@n6.nabble.com wrote: Hi Mark! I think that this is the paper that you are referring to: dx.doi.org/10.1021/ct900549r Also for your reference, these are the settings that Justin recommended using with CHARMM in gromacs: vdwtype = switch rlist = 1.2 rlistlong = 1.4 rvdw = 1.2 rvdw-switch = 1.0 rcoulomb = 1.2 As you mention the switch function in gromacs is different than in CHARMM but it appears that the difference is very small. Gianluca On Thu, 7 Nov 2013, Mark Abraham wrote: Reasonable, but CPU-only is not 100% conforming either; IIRC the CHARMM switch differs from the GROMACS switch (Justin linked a paper here with the CHARMM switch description a month or so back, but I don't have that link to hand). Mark On Thu, Nov 7, 2013 at 8:45 PM, rajat desikan [hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=0wrote: Thank you, Mark. I think that running it on CPUs is a safer choice at present. On Thu, Nov 7, 2013 at 9:41 PM, Mark Abraham [hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=1 wrote: Hi, It's not easy to be explicit. CHARMM wasn't parameterized with PME, so the original paper's coulomb settings can be taken with a grain of salt for use with PME - others' success in practice should be a guideline here. The good news is that the default GROMACS PME settings are pretty good for at least some problems (http://pubs.acs.org/doi/abs/10.1021/ct4005068), and the GPU auto-tuning of parameters in 4.6 is designed to preserve the right sorts of things. LJ is harder because it would make good sense to preserve the way CHARMM did it, but IIRC you can't use something equivalent to the CHARMM LJ shift with the Verlet kernels, either natively or with a table. We hope to fix that in 5.0, but code is not written yet. I would probably use vdwtype = cut-off, vdw-modifier = potential-shift-verlet and rcoulomb=rlist=rvdw=1.2, but I don't run CHARMM simulations for a living ;-) Mark On Thu, Nov 7, 2013 at 1:42 PM, Rajat Desikan [hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=2 wrote: Dear All, Any suggestions? Thank you. -- View this message in context: http://gromacs.5086.x6.nabble.com/CHARMM-mdp-settings-for-GPU-tp5012267p5012316.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- gmx-users mailing list[hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=3 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 [hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=4. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists -- gmx-users mailing list[hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=5 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 [hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=6. * Can't post? Read http://www.gromacs.org/Support/Mailing_Lists -- Rajat Desikan (Ph.D Scholar) Prof. K. Ganapathy Ayappa's Lab (no 13), Dept. of Chemical Engineering, Indian Institute of Science, Bangalore -- gmx-users mailing list[hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=7 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 [hidden email]http://user/SendEmail.jtp?type=nodenode=5012329i=8. * Can't
[gmx-users] after using ACPYPE , GROMACS OPLS itp file generated an atom type like opls_x with mass 0.000
Hi, i was trying to generate topology for p-phenylene vinylene polymer for OPLS forcefield using acpype . The itp file i got has the atomtype opls_x with mass 0.00. Is there any way to rectify this? After reading through how acpype works i found out this was one of the possible errors but there was no solution to it. This is a part of the itp file generated: [ atoms ] ; nr type resi res atom cgnr charge mass ; qtot bond_type 1 opls_145 1 LIG C1-0.117500 12.01100 ; qtot -0.118 CA 2 opls_145 1 LIGC12-0.055800 12.01100 ; qtot -0.173 CA 3 opls_145 1 LIGC23-0.117500 12.01100 ; qtot -0.291 CA 4 opls_145 1 LIGC34-0.131000 12.01100 ; qtot -0.422 CA 5 opls_145 1 LIGC45-0.125000 12.01100 ; qtot -0.547 CA 6 opls_145 1 LIGC56-0.131000 12.01100 ; qtot -0.678 CA 7 opls_x 1 LIGC67-0.099200 0.0 ; qtot -0.777 x 8 opls_x 1 LIGC78-0.105200 0.0 ; qtot -0.882 x 9 opls_145 1 LIGC89-0.048800 12.01100 ; qtot -0.931 CA 10 opls_145 1 LIGC9 10-0.119500 12.01100 ; qtot -1.051 CA 11 opls_145 1 LIG C10 11-0.118500 12.01100 ; qtot -1.169 CA 12 opls_145 1 LIG C11 12-0.051800 12.01100 ; qtot -1.221 CA 13 opls_145 1 LIG C12 13-0.118500 12.01100 ; qtot -1.339 CA 14 opls_145 1 LIG C13 14-0.119500 12.01100 ; qtot -1.459 CA 15 opls_x 1 LIG C14 15-0.101200 0.0 ; qtot -1.560 x 16 opls_x 1 LIG C15 16-0.103200 0.0 ; qtot -1.663 x 17 opls_145 1 LIG C16 17-0.049800 12.01100 ; qtot -1.713 CA 18 opls_145 1 LIG C17 18-0.119500 12.01100 ; qtot -1.833 CA 19 opls_145 1 LIG C18 19-0.119000 12.01100 ; qtot -1.952 CA 20 opls_145 1 LIG C19 20-0.050800 12.01100 ; qtot -2.002 CA 21 opls_145 1 LIG C20 21-0.119000 12.01100 ; qtot -2.121 CA 22 opls_145 1 LIG C21 22-0.119500 12.01100 ; qtot -2.241 CA 23 opls_x 1 LIG C22 23-0.102200 0.0 ; qtot -2.343 x 24 opls_x 1 LIG C23 24-0.102200 0.0 ; qtot -2.445 x 25 opls_145 1 LIG C24 25-0.050800 12.01100 ; qtot -2.496 CA 26 opls_145 1 LIG C25 26-0.119000 12.01100 ; qtot -2.615 CA 27 opls_145 1 LIG C26 27-0.119000 12.01100 ; qtot -2.734 CA 28 opls_145 1 LIG C27 28-0.050800 12.01100 ; qtot -2.785 CA 29 opls_145 1 LIG C28 29-0.119000 12.01100 ; qtot -2.904 CA 30 opls_145 1 LIG C29 30-0.119000 12.01100 ; qtot -3.023 CA 31 opls_x 1 LIG C30 31-0.102200 0.0 ; qtot -3.125 x 32 opls_x 1 LIG C31 32-0.102200 0.0 ; qtot -3.227 x 33 opls_145 1 LIG C32 33-0.050800 12.01100 ; qtot -3.278 CA 34 opls_145 1 LIG C33 34-0.119000 12.01100 ; qtot -3.397 CA 35 opls_145 1 LIG C34 35-0.119000 12.01100 ; qtot -3.516 CA 36 opls_145 1 LIG C35 36-0.050800 12.01100 ; qtot -3.567 CA 37 opls_145 1 LIG C36 37-0.119000 12.01100 ; qtot -3.686 CA 38 opls_145 1 LIG C37 38-0.119000 12.01100 ; qtot -3.805 CA 39 opls_x 1 LIG C38 39-0.102200 0.0 ; qtot -3.907 x 40 opls_x 1 LIG C39 40-0.102200 0.0 ; qtot -4.009 x 41 opls_145 1 LIG C40 41-0.050800 12.01100 ; qtot -4.060 CA 42 opls_145 1 LIG C41 42-0.119000 12.01100 ; qtot -4.179 CA 43 opls_145 1 LIG C42 43-0.119500 12.01100 ; qtot -4.299 CA 44 opls_145 1 LIG C43 44-0.049800 12.01100 ; qtot -4.348 CA 45 opls_145 1 LIG C44 45-0.119500 12.01100 ; qtot -4.468 CA 46 opls_145 1 LIG C45 46-0.119000 12.01100 ; qtot -4.587 CA 47 opls_x 1 LIG C46 47-0.103200 0.0 ; qtot -4.690 x 48 opls_x 1 LIG C47 48-0.101200 0.0 ; qtot -4.791 x -- 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
[gmx-users] mpi segmentation error in continuation of REMD simulation with gromacs 4.5.5
Dear all, I'm trying to continue a REMD simulation using gromacs4.5.5 under NPT ensemble, and I got the following errors when I tried to use 2 cores per replica: [node-ib-4.local:mpi_rank_25][error_sighandler] Caught error: Segmentation fault (signal 11) [node-ib-13.local:mpi_rank_63][error_sighandler] Caught error: Segmentation fault (signal 11) ... Surprisingly, it worked fine when I tried to use only 1 core per replica.. I have no idea what caused the problem.. Could you give me some advice? ps. the command I used is srun .../gromacs-4.5.5-mpi-slurm/bin/mdrun_infiniband -s remd_.tpr -multi 48 -replex 1000 -deffnm remd_ -cpi remd_.cpt -append Best Qin -- 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
[gmx-users] Ligand simulation
Dear users, Although this topic has been extensively discussed in the list previously, I am unclear about the solution for the problem.. While running ligand in water simulation (EM) with RF-0 I get the following message: -- Number of degrees of freedom in T-Coupling group rest is 24531.00 Largest charge group radii for Van der Waals: 0.458, 0.356 nm Largest charge group radii for Coulomb: 0.458, 0.356 nm NOTE 1 [file ../em.mdp]: The sum of the two largest charge group radii (0.814337) is larger than rlist (1.60) - rvdw (1.00) NOTE 2 [file ../em.mdp]: The sum of the two largest charge group radii (0.814337) is larger than rlist (1.60) - rcoulomb (1.40) -- But I continued for nvt and npt where I got the same notes NVT - -- Largest charge group radii for Van der Waals: 0.509, 0.487 nm Largest charge group radii for Coulomb: 0.509, 0.487 nm NOTE 1 [file ../nvt.mdp]: The sum of the two largest charge group radii (0.996343) NOTE 2 [file ../nvt.mdp]: The sum of the two largest charge group radii (0.996343) . -- for NPT - Number of degrees of freedom in T-Coupling group System is 16357.00 Largest charge group radii for Van der Waals: 0.787, 0.684 nm Largest charge group radii for Coulomb: 0.787, 0.684 nm NOTE 1 [file ../npt.mdp]: The sum of the two largest charge group radii (1.470764) . NOTE 2 [file ../npt.mdp]: The sum of the two largest charge group radii (1.470764) . -- For MD - -- Largest charge group radii for Van der Waals: 0.671, 0.605 nm Largest charge group radii for Coulomb: 0.671, 0.605 nm NOTE 1 [file md.mdp]: The sum of the two largest charge group radii (1.276104) .. NOTE 2 [file md.mdp]: The sum of the two largest charge group radii (1.276104) . -- The ligand is not broken, whole of it is inside the water in the beginning of the simulation, topology is ok because protein-ligand simulation with PME ran fine. Any suggestions are welcome. Thank you Regards kavya -- 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