Aw: [gmx-users] GPU-based workstation

There's also these, but 1 chip runs 6K US, they can get performance up to 2.3 teraflops per chip though double percission...but have no clue about integration with GPU's...Intell also sells their chips on PCIe cards...but get only about 350 Gflops, and run 1K US$.   http://en.wikipedia.org/wiki/

Re: [gmx-users] grommp problem in gromacs 4.6.1

Looks like it. Please open an issue at gromacs.redmine.org and attach enough files and instructions to reproduce the problem. Thanks Mark On May 24, 2013 9:48 PM, "cyberjhon" wrote: > Hi guys > > I am trying to generate a .tpr file using grommp in gromacs 4.6.1. > Unfortunately I am obtaining a

Aw: Re: [gmx-users] the "-dist" flag of g_hbond tool

  The hydrogen bonding energy would have/is usefull to myself.  An example, I use the .ndx as you did below for protein-protein interactions only.  I get around 25 and 28 for two different states.  The interesting part is the 25 is about 7 times the delG, however the hydrogen bonds move much less

Aw: [gmx-users] GPU-based workstation

Dear Dr. Starlight,   Dont know the answere to all, but funny I was looking at performance on varied others web sites.  I use a core i7 970, but it seems their newest chip is almost the same as the i7 in performance (thier newer chips dont scale past 12 cores for some internal chip based design,

Re: [gmx-users] On pressure controlling

On 5/24/13 3:49 PM, Bao Kai wrote: Dear all, I am wondering if I can make the box only scale in one single direction, such as the z direction when controlling the pressure so that I can fix the area of the box in xy plane. I just want to fix the area of the interface during the NPT simulation

[gmx-users] On pressure controlling

Dear all, I am wondering if I can make the box only scale in one single direction, such as the z direction when controlling the pressure so that I can fix the area of the box in xy plane. I just want to fix the area of the interface during the NPT simulation. Thanks a lot. Best, Kai -- gmx-use

[gmx-users] grommp problem in gromacs 4.6.1

Hi guys I am trying to generate a .tpr file using grommp in gromacs 4.6.1. Unfortunately I am obtaining a very strange error: --- Program grompp, VERSION 4.6.1 Source code file: /N/u/ortgrp/Quarry/src/gromacs-4.6.1/src/kernel/toppush.c, line: 133

[gmx-users] Re: Have your ever got a real NVE simulation (good energy conservation) in gromacs?

Hi Guys Thanks for your reply 1. Ok guys I will use the following Double precision gromacs dt=1fs pbc = xyz nstlist = 5 rlist= 1.7 rlist_long = 2.0 coulombtype = PME-Switch rcoulomb_switch = 1.2 rcoulo

Re: [gmx-users] the "-dist" flag of g_hbond tool

It used to be. I didn't realise it was still in the code. We experimented a bit with having a continuous bond criterion instead of a binary measure. It didn't do for us what we hoped it would so it was abandoned. I know that some people are using Espiniozas empirical formula for bond energy, how

Re: [gmx-users] the "-dist" flag of g_hbond tool

Yes. I have looked at it already. I may need to spend time to understand it. By the way, in the source code, it seems some part are written for calculating hydrogen bonding energy, but I haven't see any "flag command" could give a output of "hydrogen bonding energy" file. Is it still under develop

Re: [gmx-users] the "-dist" flag of g_hbond tool

Hm. That is peculiar. The source code has the answer of course. I can have a look next week to see why that is. Erik On 24 May 2013, at 14:11, CHEN Pan wrote: > Hi, > > I have 512 donors and 1024 acceptors. > > I have just tested "g_hbond" with my standard crystal structure, which I > should

[gmx-users] Re: GPU-based workstation

In particular I'd like to more carefully choose gpu card (and number of cards) as well as CPU 1) In case of GPU I have to possible options 2 x GeForce 680 (256 bit, 4gb ram ) or 1 x Geforce 690 (320 bit, 6 gb ram). >From what option the best performance should be expected ? What GPU's parameters

Re: [gmx-users] the "-dist" flag of g_hbond tool

Hi, I have 512 donors and 1024 acceptors. I have just tested "g_hbond" with my standard crystal structure, which I should get 512 hydrogen bonds. And the output "hbnum.xvg" does show 512 hydrogen bonds, which is correct. But the "hbdist.xvg" file still shows that the summation of population is 20

Re: [gmx-users] Re: REMD analysis

Dear Sir, Thank you for the advice. I have not understood the things properly, especially the convergence of REMD. I got two relevant papers : 1. Convergence of replica exchange molecular dynamics 2. Convergence and sampling efficiency in replica exchange simulations of peptide folding in explici

Re: [gmx-users] the "-dist" flag of g_hbond tool

Hi, See below On 24 May 2013, at 11:45, CHEN Pan wrote: > Dear Gromacs users, > > I am confused about the g_hbond tools. > > 1) When I use "-dist" to get the distribution of hydrogen bonding distance, > I found that the summation of the population is always 200 (the y-column > below). I am no

[gmx-users] GPU-based workstation

Dear Gromacs Users! I'd like to build new workstation for performing simulation on GPU with Gromacs 4.6 native cuda support. Recently I've used such setup with Core i5 cpu and nvidia 670 GTX video and obtain good performance ( ~ 20 ns\day for typical 60.000 atom system with SD integrator) Now

Re: [gmx-users] REMD:subsystems are not compatible.

Depends what's in the _prev files. The original problem might be http://redmine.gromacs.org/issues/801, in which case you should try 4.5.6 On Fri, May 24, 2013 at 1:01 PM, suhani nagpal wrote: > Hello Mark > > Thanks for the suggestions. > > I used gmxcheck on my _prev.cpt files, they all show

Re: [gmx-users] REMD:subsystems are not compatible.

Hello Mark Thanks for the suggestions. I used gmxcheck on my _prev.cpt files, they all show last time step - 7.3 ns but when I resume remd with _prev.cpt files , it's starting from 0 ps. Did i miss something ? On Fri, May 24, 2013 at 3:37 PM, Mark Abraham wrote: > Probably your set of state_

Re: [gmx-users] Re: REMD analysis

Hi to everybody, Bharat, maybe i didn't follow exactly the wole tale, but is it possible you are running xmgrace without the -nxy option? You are probably visualizing the data related the 1st replica several times! Francesco 2013/5/24 Mark Abraham > On Fri, May 24, 2013 at 10:44 AM, bharat gup

Re: [gmx-users] REMD:subsystems are not compatible.

Probably your set of state_.cpt files is mutually inconsistent after some crash, because some files were written and some were not. Use gmxcheck to find out the times. The good news is that the _prev.cpt files will allow you to construct a self-consistent set. Mark On Fri, May 24, 2013 at 11:36

Re: [gmx-users] Re: REMD analysis

On Fri, May 24, 2013 at 10:44 AM, bharat gupta wrote: > Dear Sir, > > Thank you for your detailed response to my query. I understood the concept > of ordered arrangement of ensembles in replica_index.xvg. But I have a > doubt, you said that " *At time 4, replicas in ensemble 1 and 2 have > exchang

[gmx-users] the "-dist" flag of g_hbond tool

Dear Gromacs users, I am confused about the g_hbond tools. 1) When I use "-dist" to get the distribution of hydrogen bonding distance, I found that the summation of the population is always 200 (the y-column below). I am not sure if it's was done with normalization or not, if yes, the summation s

[gmx-users] REMD:subsystems are not compatible.

Greetings I'm running REMD with 96 replicas. It has run till 7.3 ns after extending it once using mdrun_mpi -s md_.tpr -multi 96 -replex 2000 -cpi state_.cpt -noappend And if i put to restart this again it shows the error: --- Program mdrun_m

Re: [gmx-users] Re: REMD analysis

Dear Sir, Thank you for your detailed response to my query. I understood the concept of ordered arrangement of ensembles in replica_index.xvg. But I have a doubt, you said that " *At time 4, replicas in ensemble 1 and 2 have exchanged. So replica 0 is now in ensemble 2, which is expressed by 0 in

Re: [gmx-users] Re: REMD analysis

At time 0 we have an set of replicas and an (ordered) set of ensembles. We could label these however we liked, but for (in)convenience we use 0-(n-1) for both. The rows of the matrices in the *.xvg files change with time. At time 2, replicas in ensemble 0 and 1 have exchanged, so replica 0 is now i

Re: [gmx-users] g_hbond

Hi, Just to clarify: The third column are the number of acceptor-donor pairs that fulfil the distance criterion but not the angle criterion. Erik On 24 May 2013, at 09:06, CHEN Pan wrote: > Hi Maggin, > The middle column is the total number of hydrogen bonds in your system > under the definit

Re: [gmx-users] Have your ever got a real NVE simulation (good energy conservation) in gromacs?

Hi, I have achieved good energy conservation in NVE simulations. This was single proteins in the gas phase, using infinite cut-offs, constraints applied to hydrogens, 0.5 fs or 1.0 fs time steps, double precision and virtual sites if I'm not mistaken. We had problems with energy conservation fo

[gmx-users] Re: residue labeling in g_rmsf

Hi, you can use Matlab software to do it! maggin -- View this message in context: http://gromacs.5086.x6.nabble.com/residue-labeling-in-g-rmsf-tp5008474p5008514.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- gmx-users mailing listgmx-users@gromacs.org http:/

[gmx-users] Re: g_hbond

Hi, Pan Thank you very much for your kind help! Thank you! maggin -- View this message in context: http://gromacs.5086.x6.nabble.com/g-hbond-tp5008507p5008513.html Sent from the GROMACS Users Forum mailing list archive at Nabble.com. -- gmx-users mailing listgmx-users@gromacs.org http://

Re: [gmx-users] residue labeling in g_rmsf

Hi Arunima, In bash you can also combine the residue names and such with the RMSF graph: paste <(grep "^\(ATOM\|HETATM\)" structure.pdb | cut -b 17-26 | uniq) <(grep "^[^#@]" rmsf.xvg) To make it a bit more clear, the redirection <( ... ) treats the output of the command between parentheses as a

[gmx-users] Re: Have your ever got a real NVE simulation (good energy conservation) in gromacs?

Hi John, I was dealing with this issue recently. Have you considered these points? 1. Run gromacs in double-precision (mdrun_d) 2. use PME-Switch for the treatment of the Coulombic interaction 3. follow the guidelines from here: http://www.gromacs.org/Documentation/Terminology/NVE

Re: [gmx-users] g_hbond

Hi Maggin, The middle column is the total number of hydrogen bonds in your system under the definition of the hydrogen bonds criteria you have defined (here you used the default value). The third column is the number of acceptor-donor paris in the system. The first column is the simulation time, yo