[CPMD-list] VDW energy

[Axel Kohlmeyer]

On Mon, 9 Apr 2007, [GB2312] ÓÚ°×Èã wrote:

dear bairu,

BY> Dear all cpmd-lists,
BY> Dear Axel,
BY> Thanks for your kind help. I have run again according to your suggestion,
BY> but the problem is still there. Why there is no wavefunction velocities in
BY> the restart file? Maybe something goes wrong with my input file. Could you

they are not supposed to be there after a wavefunction optimization.

BY> please have a look at it and help me to find the problem? Thank you very
BY> much.
BY> Here is my input file:
BY> 
BY> &CPMD
BY>   OPTIMIZE GEOMETRY

this is the root of your problem. you claimed to have run
a wavefunction optimization, but instead ran a geometry
optimization. this is in principle equivalent, _but_ in a
geometry optimization the _forces_ on the ions (=atoms) are
stored in the same place as the velocities in an MD. so if
you RESTART with VELOCITIES, then your initial velocities will
be seeded from the forces and if those are not very, very low,
your dynamics will be completely screwed up.

so you have to either use a regular wavefunction optimization
or restart without velocities.

BY>   NOSE IONS MASSIVE
BY>   300.0 2500.0
BY>   NOSE ELECTRONS
BY>   0.007 15000.0

there is no need for this, so it is better not to put those
keywords there at all. it is a very bad habit of (too?) many 
CPMD users, to keep options that supposedly have no effect in 
the input file or disable them by setting the first letter to 
lowercase. this is not only confusing to new users but it may 
even backfire with some later versions of the code, since some 
functionality may have been added that would induce some unexpected
side effects or there could be new keywords that may match those 
'disabled' keywords.

BY>   ODIIS
BY>    5
BY>    MAXSTEP
BY>    500
BY>   STORE
BY>    100

geometry optimization writes a restart after each 
optimization step. no need to set STORE.

BY>   TIMESTEP
BY>    4.0
BY>   EMASS
BY>    700.0

both are not needed.

BY>   COMPRESS WRITE32

in the age of >100GB hard drives, compressed restart
files should only be used for test jobs or benchmarks.
this is a lossy compression and particularly if you
want to seed an MD, the additional noise will cancel
some of the effect of the wavefunction optimization.

BY> &END
BY> 
BY> &DFT
BY>   NEWCODE
BY>   FUNCTIONAL BLYP
BY>   GC-CUTOFF
BY>    1.0D-06
BY> &END
BY> 
BY> &VDW
BY>   VDW CORRECTION
BY>    1
BY>    C6  1 1 -19.489 4.535 3.0
BY> &END
BY> 
BY> &SYSTEM
BY>   SYMMETRY
BY>    0
BY>   CELL
BY>    15.000  1.0  1.0  0.0  0.0  0.0
BY>   CUTOFF
BY>   70.0
BY>   ANGSTROM
BY> &END
BY> 
BY> &ATOMS
BY> *N_MT_GIA_BLYP KLEINMAN-BYLANDER RAGGIO=1.0
BY>    LMAX=P LOC=P
BY>    4
BY> 7.500   5.500  6.953    1   1
BY> 7.500   5.500  8.047    2   1
BY> 6.953   9.500  7.500    3   1
BY> 8.047   9.500  7.500    4   1
BY> &END
BY> 


BY> My md input file is:
BY> 
BY> &CPMD
BY>   MOLECULAR DYNAMICS
BY>   QUENCH BO ELECTRONS
BY>   RESTART ACCUMULATORS WAVEFUNCTION COORDINATES VELOCITIES CELL
BY>   RESTART LATEST

you should only try to read values from the restart that you 
actually want to read. you want to _set_ initial velocities,
so why read them? your cell is fixed, why restart it? you have
not run any MD, why restart with accumulators? you only risk
that you run into some bug and you gain nothing.

BY>   TRAJECTORY XYZ
BY>   TEMPERATURE
BY>   50.0D0

note that the TEMPERATURE keyword has no effect when you
restart with velocities.

BY>   MAXSTEP
BY>    500
BY>   STORE
BY>    100
BY>   TIMESTEP
BY>    4.0
BY>   EMASS
BY>   700.0
BY>   COMPRESS WRITE16

a restart with write16 is completely useless
for 'production' MD. see above for the reasoning.

so please try again with a more careful input 
and then see if it works.

cheers,
  axel.

BY> &END
BY> 
BY> .....................................
BY> 

-- 
=======================================================================
Axel Kohlmeyer   akohlmey at cmm.chem.upenn.edu   http://www.cmm.upenn.edu
   Center for Molecular Modeling   --   University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582,  fax: 1-215-573-6233,  office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.