Hi David,
I did following test study: I had produced a shorter version of 6000 frames .nc file by using trajout cpptraj subroutine to produce 10 frames .nc file which I used in the following test cases.
Study 1 >> Frame 1-10 > GB = -47.601 , PB = -6.2504
Study 2 >> Frame 1-3 > GB = -47.0283, PB = -5.4966 , Frame 3-10 > GB = -47.2521, PB = -6.1103
Study 3 >> Frame 1-5 > GB = -46.6376, PB = -5.0888 , Frame 5-10 > GB = -48.1015, PB = -6.686
In Study 2 & 3 I used the same 1-10.nc of Study 1 to split it into 1-3&3-10 and 1-5&5-10 frames .nc respectively.
They doesn’t add up to same value.
——
Jatin Kashyap
Ph.D. Student
Dr. Dibakar Datta Group
Department of Mechanical and Industrial Engineering
New Jersey Institute of Technology (NJIT)
University Heights
Newark, NJ 07102-1982
Phone- (201)889-5783
Email- jk435.njit.edu
> On Mar 31, 2021, at 8:23 AM, David A Case <david.case.rutgers.edu> wrote:
>
> On Wed, Mar 31, 2021, Jatin Kashyap wrote:
>>
>> I never did the solution number 3.
>>>
>>> I would stick with one core, and use parallelization to farm the snapshots
>>> out to a number of separate jobs, all running at the same time.
>
> MM-PB/GBSA is completely linear. If you analyze half of your snapshots in
> one job, and the other half in a second job, you can just add up the
> results. Same for a division of the frames into a smaller number of
> subframes.
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Fri Apr 02 2021 - 13:30:02 PDT
