Dear Jason
Thanks for your answer.
Is pmemd related to GPU, only?
If I install ambertools on cpu, can I install pmemd? or not?
On Mon, May 4, 2015 at 4:56 AM, Jason Swails <jason.swails.gmail.com> wrote:
> On Sun, May 3, 2015 at 12:10 PM, David A Case <case.biomaps.rutgers.edu>
> wrote:
>
> > On Sun, May 03, 2015, Atila Petrosian wrote:
> > >
> > > Should I do both of parallel installation and pmemd installation?
> >
> > This depends on whether you have computers available to you with multiple
> > cores. Most computers do these days, so it is generally worth making a
> > parallel installation.
> >
> > >
> > > Two general question : which of pmemd and sander.MPI is faster?
> >
> > For a given number of cores, pmemd.MPI will generally be faster than
> > sander.MPI.
> >
> > > which of pmemd and sander.MPI is more accurate?
> >
> > Both give the same results.
> >
> > >
> > > I want to do MM-PB(GB)SA in future. Which of pmemd and sander.MPI is
> more
> > > appropriate for me?
> >
> > The "expensive" part of MM-PBSA is creating the solvated trajectory in
> the
> > first place. So what is worthwhile optimizing is the speed of "ordinary"
> > simulations.
> >
> > My advice: if you are just beginning, don't worry about pmemd for now.
> > Use your AmberTools14 installation, and gain experience in running and
> > analyzing simulations. Later, you can return to the question of how best
> > to install pmemd. Basic instructions are here:
> >
> > http://jswails.wikidot.com/installing-ambertools-14-and-older-amber
> >
> > but they won't work unchanged for you, since for some reason you can't
> get
> > the
> > old AmberTools1.5 to build. I think I can persuade some deveoloper to
> work
> > up details of how to compile just pmemd from an old distribution (without
> > requiring a complete build of AmberTools), and use it alongside
> > AmberTools14.
> >
>
> I can add it to my Wiki if it's of interest, but it's an awful lot of
> headache to go through for just doubling the speed of sander (compared to
> the more-than-an-order-of-magnitude boost you get from GPUs, which will not
> work with Amber 11).
>
> All the best,
> Jason
>
> --
> Jason M. Swails
> BioMaPS,
> Rutgers University
> Postdoctoral Researcher
> _______________________________________________
> 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 Mon May 04 2015 - 11:00:09 PDT
