Hello,
On Wed, Apr 13, 2011 at 12:40 AM, Sangita Kachhap <sangita.imtech.res.in>wrote:
>
> Hello all
>
> I am running AMBER ADVANCED TUTORIALS TUTORIAL 3 - SECTION 3.5 Nmode
> analysis.
> Here I am confused about number of frames, numberof processor and thread.
>
> Input file is:
>
> Input file for running entropy calculations using NMode
> &general
> endframe=50, keep_files=2,
> /
> &nmode
> nmstartframe=1, nmendframe=50,
> nminterval=5, nmode_igb=1, nmode_istrng=0.1,
> /
>
>
> progress.log file is:
>
> 50 frames were read in and processed by ptraj for use in calculation.
> Processing 10 frames with normal mode analysis.
>
This is to be expected from your input file, but I think you know that
already.
>
> Starting calculations
>
>
> Starting nmode calculations...
> master thread is calculating 2 frames
>
>
> I am running it in 5 processor.
>
This is why. You are analyzing 10 frames total, with 5 processors. Thus,
each processor is taking 2 frames. In MPI-lingo, what you're actually
executing, though, is called a "thread". So what you're running here are 5
threads, not 5 "processors". (Although if you have 5 processors available,
one would hope that the 5 threads run on 5 processors). There's nothing in
MPI *against* running more "threads" than you have physical processors (it's
just inefficient). The master thread is the thread that does all of the
printout (for all intents and purposes, it's an arbitrarily assigned
thread). It will also *always* operate on the largest number of frames.
For instance, if you had 11 frames and 5 processors, the master thread would
analyze 3 frames while the rest analyzed 2.
Hope this helps clarify things,
Jason
> Here total frames processed are 10 then why calculation only for 2 frames.
>
> What is mean by "thread" in progress.log?
>
> So can anybody sugesst me to clear it.
>
> Sangita Kachhap
> JRF
> BIC,IMTECH
> CHANDIGARH
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>
--
Jason M. Swails
Quantum Theory Project,
University of Florida
Ph.D. Candidate
352-392-4032
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Apr 13 2011 - 01:30:04 PDT
