On Wed, Sep 29, 2010, Vijay Manickam Achari wrote:
> According to Case's advice, I have tried to run small number of steps
> using fewer nodes. It runs in 16 nodes but more than this number it
> terminates indicating the same problem about as shake.
>
> Why this happens and where the mistake could be?
Did you run a 10-step simulation and look at the output on every step? Are
there differences at step 1? You need to look as closely as you can at pmemd
below and above 16 nodes, and versus sander. Saying there is a shake error is
of no new help--it just means something is wrong. Finding the exact
differences in short outputs may help in locating the problem. I don't recall
hearing of a symptom like this before, so additional (and detailed)
information is likely to be required.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Sep 29 2010 - 05:00:05 PDT
