Is this a single processor (sander, not sander.MPI) run? With a single
processor run, you don't have to worry about stderr getting lost for the
nonmaster nodes (should not happen, but not all is perfect in the parallel
world), and also there are more obscure error conditions where strange
things can happen on the nonmaster node without helpful error messages. If
there are namelist errors, they can be very obscure and will show up on
stderr, not in the mdout file. Also, namelist errors is a sort of touchy
issue, dependent on the compiler you are using in some cases (some compilers
in the past had their own "standard" for namelists, since there was no real
standard). As Carlos says, there are bad things that can happen when
specifying the &ewald section too, and I am not sure you have any real need
here. I would try, in &cntrl, nstlim=50, ntpr=1 to further debug (so you
get output with each step, and only do a few steps).
Regards - Bob Duke
----- Original Message -----
From: "parul sharma" <sharmaparul7373.gmail.com>
To: "AMBER Mailing List" <amber.ambermd.org>
Sent: Monday, February 23, 2009 4:43 AM
Subject: [AMBER] PME doesn't work
> Dear All,
>
> I did MD simulation with NVT for 100 ps, and then i proceeded with NPT,
> but when i try to include PME (the input for which is given below)
>
> &cntrl
> imin = 0, irest = 1, ntx = 7, ntb = 2, ntp = 1, pres0 = 1.0, taup = 2,
> cut = 10, tempi = 300.0, temp0 = 300.0,
> ntc=2, ntf=2, tol = 0.0001,
> ntt = 3, gamma_ln = 1.0,
> scee= 1.2, nstlim =250000, dt = 0.002,
> ntpr = 500, ntwx = 500, ntwr = 500, ntwe = 500,
> &ewald
> ew_type = 0 nfft1 = 25, nfft2 = 25, nfft3 = 20,
> a = 58.8103873, b = 47.4014584, c = 33.2335295,
> alpha = 90.0, beta = 90.0, gamma = 90.0
> order = 4, dsum_tol = 0.00001,
> /
>
> It gives me the following error in the output:
>
> | Flags: MPI
> getting new box info from bottom of inpcrd
> | INFO: Old style inpcrd file read
> | peek_ewald_inpcrd: Box info found
>
> Can anyone please help me in sorting the problem out?
>
> --
> With Best Regards
>
> Parul Sharma
> PhD Student, Durban University of Technology,
> Durban,
> South Africa
> _______________________________________________
> 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 Wed Feb 25 2009 - 01:09:24 PST
