FIST:: multiple potential

[Teodoro Laino]

Hi all,
this message focus on FIST (people not interested in that part of  
cp2k can skip this message).
There was a major rewriting (yet another one!) of the pair_potential  
type and consequently of the
many-body.. Since it is important that future modifications preserve  
the object oriented structure I massively cleaned the
many-body and the pair_potential type, forcing (maybe even up to a  
paranoiac level) a hierarchical structure
for that part of the code..

An example is EAM.. EAM potential contains a pair_potential term and  
a many-body term.. In the old code the two were mixed..
I went for splitting them and treating each term in the proper module..
Within this cleaning I decided to enable the possibility to define a  
non-bonded potential (even ONFO) as an overlap of
different functional forms.. for example you may define a Lennard- 
Jones potential as:

         &LENNARD-JONES
           atoms O O
           EPSILON [kcalmol]  0.152073
           SIGMA   [angstrom] 3.1507
           RCUT    [angstrom] 11.4
         &END LENNARD-JONES

Or as:

         &LENNARD-JONES
           atoms O O
           EPSILON [kcalmol]  .07603650000000000000
           SIGMA   [angstrom] 3.1507
           RCUT    [angstrom] 11.4
         &END LENNARD-JONES
         &LENNARD-JONES
           atoms O O
           EPSILON [kcalmol]  .07603650000000000000
           SIGMA   [angstrom] 3.1507
           RCUT    [angstrom] 11.4
         &END LENNARD-JONES

They will give the same result (in the last case the keyword  
MULTIPLE_POTENTIAL must be switched on..)
In case the keyword MULTIPLE_POTENTIAL is not switched on the  
behavior is the usual one.. the last defined
potential is the one used and all the others are deleted because  
overwritten by the next defined..

Instead of two Lennard-Jones you can define a LENNARD-JONES and a  
GENPOT potential for the same
atomic pair.. or instead of only two potentials you can define it as  
a sum of N different kind of potentials..

Since this was a quite relevant change I hope that you will not  
experience any trouble with "regular" FIST jobs.
In case let me know,

Ciao from the sunny (as never was) Zurich,
teo

p.s.: Piecewise potentials in the next commit.. ;-)