[gmx-users] "constraints:none" segmentation fault

David van der Spoel spoel at xray.bmc.uu.se
Mon Apr 21 08:25:01 CEST 2003

On Sun, 2003-04-20 at 22:47, Giorgos Karvounis wrote:
> Dear all,
> I want to simulate a b-hairpin in spc water in a NVE ensemble. I
> basically want to "reproduce" the simulation of beta-hairpin of
> Thermistat from the extended conformation at 360 K ( A. Bonvin et al.
> J.M.B (2000) 296,255-268).  I am using PME without any coupling
> (neither  T nor P). I run first a position restraints run , with T
> coupling on , then i turn it off for a md run. The problem is that the
> temperature is gradually cooling down... looking at the user's list, it
> seems that others faced the same problem and appartently a fisrt
> approach to this problem is to use FLEX_SPC and "constraints: none". My
> problem is that, when i use "constraints :none" in the .mdp file, after
> 50 steps , i get a segmentation fault and the run crashes.
> So basically, after having read many emails about PME and energy
> conservation, and still being quite confused about it, can you please
> summarise wich are the best parameters for this type of run and if you
> can tell me a possible reason for the segmentation fault?
> I am grateful for any advice.
It is rather tricky to get good energy conservation, since PME is still
approximate (due to grid and cut-off in short range energy function),
besides that you still have a Vanderwaals cut-off and probably a number
of nstlist > 1 which affects energy conservation. If you use a twin
range cut-off it will be worse still. I don't think constraints are very
important though also are approximate. If you are lucky these errors
cancel out roughly, but this is not a real solution to the energy
conservation problem of course. 

If you do use T coupling with PME however, you can monitor the
temperature coupling constants lambda afterwards (they are stored in the
energy file) and check how much your system is disturbed by it.
Then redo the same simulation without PME and compare...

> Thank you in advance
> Giorgos
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the 
> www interface or send it to gmx-users-request at gromacs.org.
Groeten, David.
Dr. David van der Spoel, 	Dept. of Cell & Mol. Biology
Husargatan 3, Box 596,  	75124 Uppsala, Sweden
phone:	46 18 471 4205		fax: 46 18 511 755
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://xray.bmc.uu.se/~spoel

More information about the gromacs.org_gmx-users mailing list