AMBER Archive (2003)

Subject: Sander box error and thanks on previous mdin error

From: kxg2248_at_njit.edu
Date: Thu Jan 02 2003 - 21:41:49 CST

First, thanks to all for the quick response to my first posting. I think
the combination of removing a line from after the &end statement, and
adding the FIND/SEARCH commands to find my non-standard residues helped a lot.

Now I am getting through to an actual MD run. But, sander indicates that
my molecules are going outside of the box. It says to rerun from a
restart, and I set my restart write frequency to 1, and I tried to rerun
it. But I get a different error on restart.

Here is the error message I get after the restart:
| Flags: SGIFFT
HAS_FTN_ERFC
  getting new box info from bottom of inpcrd
| peek_ewald_inpcrd: Box info not found in inpcrd
  compute_nfft: failed to get good fft array size
  x = 4.4184937307205338E+25

Here is the restart file:
amAT1

    24 0.7400000E-01
  -26.7364864 -28.8505423 -30.2734019 -25.9030359 -28.4078195 -29.7462229
  -26.8225338 -29.8887011 -29.9884067 -26.5286425 -28.7431283 -31.7307417
  -26.4283118 -27.3320194 -32.1447242 -26.3174364 -27.2748172 -33.2185005
  -25.5734264 -26.8675380 -31.6758662 -27.6223584 -29.3911477 -32.4806968
  -28.5830542 -29.1179001 -32.0646159 -27.4868747 -30.7158029 -32.5147052
  -27.3856374 -31.3587726 -31.4804784 -27.4675706 -31.4117851 -33.8557336
  -27.2979586 -32.7974648 -33.9737356 -27.1044208 -33.6110651 -32.8348034
  -27.9592736 -33.5397011 -32.1770080 -26.2100571 -33.3094092 -32.3078515
  -27.3168695 -33.3733296 -35.1923931 -27.4943290 -32.6344038 -36.3057666
  -27.5271339 -33.2746293 -37.5660964 -26.7247181 -32.9120584 -38.1931787
  -28.4791298 -33.1074800 -38.0479315 -27.6439335 -31.2478614 -36.1947718
  -27.8618806 -30.2430190 -37.6120903 -27.6190177 -30.6736133 -34.9743541
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000
    0.0000000 0.0000000 0.0000000 0.0000000 0.0000000 0.0000000

And here is the original mdin file (when it ran, sander said I needed to
restart it to fix the error of going outside the box):
minimization run with belly
  &cntrl
   ntb=1, nstlim = 500,
   ntpr = 100,
   cut=15.0, igb=0,
   ntwr=1, ntwx=100,
   ntx=1, irest=0,
   ntc=1, ntf=1, tol=0.00001,
   dt=0.002,
   ibelly=1,
   tempi=313.2,temp0=313.2,
   ntt=1,

   imin=0, maxcyc=10,ncyc=100,
  &end
Group input
500.0
FIND
* * * N1
* * * N9
SEARCH
RES 1 999
END
END

The only changes to the mdin file to run from the restart are going
to: ntx=5 and irest=1. I am wondering if the igb=0 is getting me in
trouble, as I also saw messages regarding Ewald and not using belly with
standard (not particle mesh) Ewald calculations.

Thanks again,
K. Gilbert