[gmx-users] OpenMM

[PACIELLO GIULIA]

I dont'know if this is the right way to test that things 
converge to the same value...but I have made an energy 
minimization on CPU of the two output (parallel and 
serial). I have obtained this results:

FROM PARALLEL OUTPUT
Potential Energy  = -1.6960029e+04
Maximum force     =  1.7282695e+03 on atom 364
Norm of force     =  4.6171844e+02

FROM SERIAL OUTPUT
Potential Energy  = -1.7814945e+04
Maximum force     =  2.1551072e+03 on atom 496
Norm of force     =  2.9537823e+02

Thanks,
Giulia




On Wed, 14 Apr 2010 16:58:37 +1000
  Mark Abraham <Mark.Abraham at anu.edu.au> wrote:
> On 14/04/2010 4:52 PM, PACIELLO GIULIA wrote:
>> Ok...so how could I know the energies among atoms?
> 
> Since it's not reported, you can't get breakdowns of 
>energies. That's a limitation from the use of GPU. If you 
>want this information, perhaps do your simulation on the 
>GPU and re-rerun selected frames from your trajectory on 
>a different machine that does not use GPUs.
> 
>> And how can I test if
>> my parallel code is running in the correct manner?
> 
>>> MD is chaotic. Tiny differences at an early stage lead 
>>>to wildly
>>> different results later. That's inevitable. However 
>>>ensemble averages
>>> should converge to the same value.
> 
> Do things converge to the same values? There's a chance 
>that mdrun -reprod will lead to equivalent total energies 
>after only one step, if that's been implemented that way.
> 
> Mark
> -- 
> gmx-users mailing list    gmx-users at gromacs.org
> http://lists.gromacs.org/mailman/listinfo/gmx-users
> Please search the archive at 
>http://www.gromacs.org/search before posting!
> Please don't post (un)subscribe requests to the list. 
>Use the www interface or send it to 
>gmx-users-request at gromacs.org.
> Can't post? Read 
>http://www.gromacs.org/mailing_lists/users.php