ntc in sander6

From: Xavier Deupi <xavi_at_yogi.uab.es>
Date: Sat 25 Aug 2001 05:08:12 -0700

Hi all,

Recently I've been compiling and testing sander6 in different machines.
First I tried with an SGI Origin 3800, and I could start to send MD
simulations without much problem... But when I tried an IBM SP3 and a Cray
T3E, I was warned that

> this parallel version only works for ntc < 3

Browsing the amber web page I've found that

        "The bonds, angles, dihedrals, SHAKE (only on bonds involving
        hydrogen when NTC=2), nonbonded energies and forces, pairlist
        creation, and integration steps are parallelized."

So I set ntc=2, and changed dt from 0.002 to 0.001. With these changes,
both machines could start to run the jobs without problems...

My question is... why the Origin accepts ntc=3, while the IBM and the Cray
do not? (I think this also happened running sander6 under Linux, but I'm
not completely sure...). And, what's more important for me, how is the
speed of the calculations affected by this decrease in dt? (my first tests
seem to show that it is not affected, but I have to be sure before assure
that to my advisor :-) )

Thanks

xavi

________________________________________________________

  Xavier Deupi
  Computational Medicine Lab
  Biostatistics Unit. School of Medicine
  Universitat Autonoma de Barcelona. CATALUNYA (Spain)
  Phone : (3493)-581.23.48
  E-mail: xavi_at_yogi.uab.es
________________________________________________________
Received on Sat Aug 25 2001 - 05:08:12 PDT
Custom Search