AMBER Archive (2002)

Subject: minimization problem

From: Uwe Richter (uwe_at_fmp-berlin.de)
Date: Tue Mar 19 2002 - 06:22:06 CST


Hi all,

I'm having problems with some of my minimizations (solvated
protein with ntb=1, sander_classic). They
do not reach a gradient below 0.1 (somewhere in this
range). The energy is jumping up and down a little in
these cases.
I've already varied the stepsize and fiddled around
with the ntmin,ncyc keywords, but this didn't solve
the problem. The only idea I have is to further decrease
the stepsize.
I wonder whether I can do anything else to get the
energy converged.

Any hint are very much appreciated,
Uwe

Here is my current input

 &cntrl
    timlim=999999, imin=1, nmropt=0,

    ntx=1, irest=0, ntrx=1,
    ntxo=1, ntpr=100,
    ntwr=200, ntwx=200, ntwv=200, ntwe=200,
    ntwxm=0, ntwvm=0, ntwem=0,
    ntwprt=0,

    ntf=1, ntb=1,
    idiel=1,dielc=1.0,
    cut=8, ntnb=1, nsnb=25,
    ntid=0, scnb=2.0, scee=1.2,
    cut2nd=0.0, ichdna=0,
    isftrp=0, rwell=0, ipol=0,

    ibelly=0, ntr=0,

    maxcyc=10000, ncyc=10, ntmin=0,
    dx0=0.01, dxm=0.05, drms=0.01,

    ntc=1, tol=0.0005,

    pencut=0,ipnlty=1,nmrmax=0,
 &end
 &wt type='END' &end