[gmx-users] still change after equilibriated

[Erik Lindahl]

Hi,

That is NOT correct.

Pressure is a macroscopic property well-defined over 6*10^23 atoms or  
so. The second you extrapolate this down to a couple of thousand  
atoms you will observe very large fluctuations, which means you have  
to run long simulations to get a good statistical average. This has  
nothing to do with whether you are coupling the pressure or not.

The Berendsen algorithm does not result in a pure NPT ensemble, but  
it is the only algorithm to _guarantee_ exponential relaxation to the  
correct average pressure.

In this case the fluctuations are an order of magnitude higher than  
the pressure difference, which probably means they are insignificant.

Cheers,

Erik

On Dec 22, 2005, at 8:46 AM, Yang Ye wrote:

> Despite possible simulation artifacts (protocol, solute, etc.), it is
> known that Gromacs' pressure coupling (at least Berendsen) could not
> give stable pressure. Check archive for similar topics.
>
> Yang Ye
>
> Rongliang Wu wrote:
>
>> Hello gmx-users,
>>
>> i have equilibriated a system after 7 ns, and out of curiosity i  
>> continued run until 9ns. but i found the following phenomenon:
>>          Statistics over 1000001 steps [ 5000.0000 thru 7000.0005  
>> ps ], 9 data sets
>>
>> Energy                      Average       RMSD     Fluct.       
>> Drift  Tot-Drift
>> --------------------------------------------------------------------- 
>> ----------
>> Potential                  -8174.09          0          0  
>> -0.0167593   -33.5187
>> Temperature                 299.735          0          0  
>> -2.3866e-06 -0.0047732
>> Pressure (bar)             0.993253    417.974    417.974  
>> -0.000471878  -0.943758
>> Box-X                       3.27009          0          0  
>> -7.43729e-07 -0.00148746
>> Density (SI)                967.094          0          0  
>> 0.000662033    1.32407
>> Pres-XX (bar)               3.73019    721.395    719.606  
>> -0.0879504   -175.901
>> Pres-YY (bar)               2.69262    624.848    623.924  
>> -0.0588169   -117.634
>> Pres-ZZ (bar)              -3.44305    728.557    723.708    
>> 0.145352    290.704
>> #Surf*SurfTen              -21.9101    2811.25    2780.78    
>> 0.714987    1429.98
>> Heat Capacity Cv:      12.4718 J/mol K (factor = 0)
>>
>> Statistics over 1000001 steps [ 7000.0005 thru 9000.0000 ps ], 9  
>> data sets
>>
>> Energy                      Average       RMSD     Fluct.       
>> Drift  Tot-Drift
>> --------------------------------------------------------------------- 
>> ----------
>> Potential                  -8229.97          0          0   
>> 0.0348891    69.7783
>> Temperature                 299.735          0          0  
>> 1.12632e-06 0.00225264
>> Pressure (bar)              1.10322    419.616    419.616  
>> -0.000406983  -0.813967
>> Box-X                       3.26972          0          0           
>> 0          0
>> Density (SI)                967.428          0          0  
>> 2.26927e-05  0.0453855
>> Pres-XX (bar)              -57.2399    737.021     736.43   
>> 0.0511137    102.227
>> Pres-YY (bar)                16.767    613.566    613.507  
>> -0.0148347   -29.6693
>> Pres-ZZ (bar)               43.7826    732.279    731.959     
>> -0.0375        -75
>> #Surf*SurfTen               209.332    2795.88    2793.91   
>> -0.181749   -363.498
>> Heat Capacity Cv:      12.4718 J/mol K (factor = 0)
>> 	
>> why the pressure in different orientations becomes much larger. is  
>> it that the only meaningful pressure value is the the scalar  
>> pressure?
>>
>>
>> Thanks!
>>
>> Regards!
>>
>> Sincerely,
>>
>> Rongliang Wu
>> State Key Laboratory
>> of Polymer Physics and Chemistry
>> Center of Molecular Science
>> Institute of Chemistry
>> Chinese Academy of Sciences
>> 				
>>
>> 　　　　　　　　　　wurl04 at iccas.ac.cn
>> 　　　　　　　　　　2005-12-20
>>
>>
>> --------------------------------------------------------------------- 
>> ---
>>
>> _______________________________________________
>> 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.
>>
>
> -- 
> /Regards,/
> Yang Ye
> /Computational Biology Lab
> School of Biological Sciences
> Nanyang Technological University
> Singapore
> Tel: 6316-2884
> /
>
> _______________________________________________
> 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.

-----------------------------------------------------------
Erik Lindahl  <lindahl at sbc.su.se>     Backup address:  
<erik.lindahl at gmail.com>
Assistant Professor, Computational Structural Biology
Stockholm Bioinformatics Center, Stockholm University, SE 106 91  
Stockholm
Phone: +46 8 5537 8564     Fax: +46 8 5537 8214