149. Emptying Root's Trash in debian

My / partition was filling up because of /var/cache/apt/archives
The files in that folder are write-protected. I tried doing sudo rm *.deb but it didn't work. Doing sudo nautilus and deleting like that worked. However, the space doesn't get released as everything goes into the trash of root. And you can't access that easily using nautilus.

So....
sudo su
cd /root/.local/share/Trash/files
rm *.deb

148. popt, readopt and g09?

Update:
NOTE that readopt WORKS in G09 rev. B and rev. D. It DOESN'T work in G09 rev. A.

readopt is a whole lot easier to use than popt.

Original post:
For various reasons I want to be able to freeze some atoms while including other in my optimisation using gaussion 09. Readopt (http://www.gaussian.com/g_tech/g_ur/k_opt.htm) sounded like a good idea.

The problem
I can't get this to work:

%nprocshared=6
%Chk=reduced_solvate.chk
#P ub3lyp/lanl2dz 6D 10F opt=ReadOptimize Punch=(MO) Pop=() 

reduced_solvate

3 2 ! charge and multiplicity
 Cr     10.0000     10.0000     10.1100
 O     10.0000     10.0000     12.1300
 H     10.0000     10.8000     12.7000
 H     10.0000     9.20000     12.7000
 O     10.0000     7.84000     10.0000
 H     10.0000     7.15000     10.6900
[..]
 H     6.32536     3.63314     12.3506
 H     8.48233     15.9732     5.80567

noatoms atoms=1-19

readopt and rdopt don't work either -- they all give 'syntax error' (C64 flashback!)

 QPErr --- A syntax error was detected in the input line.
 #P ub3lyp/lanl2dz 6D 10F ReadOptimize Pu
                                                  '
 Last state="GCL"

The solution: 

I didn't come up with this: http://ccl.net/cgi-bin/ccl/message-new?2012+04+24+002+raw

Anyway, you can use popt but it takes a bit more preparation of the coordinates.

You can convert an .xyz file quickly by using:

cat molecule.xyz | gawk '{print $1,"-1",$2,$3,$4}'>popt_molecule.xyz

Change 0 to -1 for the atoms you want to optimise.

%nprocshared=6
%Chk=popt_solvate.chk
#Popt ub3lyp/lanl2dz 6D 10F Punch=(MO) scf=(maxcycle=1024)

reduced_solvate popt

3 2 ! charge and multiplicity
Cr 0 10.0000 10.0000 10.1100
O 0 10.0000 10.0000 12.1300
H 0 10.0000 10.8000 12.7000
H 0 10.0000 9.20000 12.7000
O 0 10.0000 7.84000 10.0000
H 0 10.0000 7.15000 10.6900
[..]
H -1 3.62686 7.14993 12.8142
H -1 5.12696 15.5239 9.01700
H -1 6.32536 3.63314 12.3506
H -1 8.48233 15.9732 5.80567

147. Oniom in nwchem -- with a little bit of help from gromacs and openbabel

Example -- I want to do explicit solvent modelling of methanol in water. This is obviously an articifical approach, but generally applicable.

This is a rough approach to doing oniom calculations using nwchem 6.0 -- this is a technical description, not a how-to when it comes to the science.

1. Pre-optimisation
Draw methanol and set up a simple calc using e.g. ecce to pre-optimise the structure with e.g. an implicit solvent model. Here's nwch.nw:

scratch_dir /scratch
Title "pre-oniom"
Start  pre-oniom
echo
charge 0
geometry autosym units angstrom
 C     0.00000     0.00000     0.00000
 H     -0.675500     -0.675500     0.675500
 H     0.675500     -0.675500     -0.675500
 H     -0.675500     0.675500     -0.675500
 O     0.866025     0.866025     0.866025
 H     1.51843     0.213620     1.51843
end
ecce_print /home/me/jobs/jobs/testing/old/pre-oniom/ecce.out
basis "ao basis" spherical print
  H library "6-31+G*"
  O library "6-31+G*"
  C library "6-31+G*"
END
dft
  mult 1
  XC b3lyp
  mulliken
end
driver
  default
end
cosmo
end
task dft optimize
task dft freq

2. Solvation using gromacs

Take the output, nwch.nwout, and use babel to export the optimised structure

babel -inwo nwch.nwout -oxyz molecule.xyz

The next few steps require gromacs:
editconf -f molecule.xyz -o molecule.gro -box 2 2 2
genbox -cp molecule.gro -cs spc216.gro -o solvated.gro
babel -igro solvated.gro -oxyz solvated.xyz

tail -n +3 solvated.xyz > oniom.nw

3. Putting it all together

The only 'trick' is how to define what part of the input belongs to the high level section and what belongs to the low level section -- for a solvation case like this, where whole molecules are treated by one method or another, use model. The last atom to be part of the high level section is the last of the six atoms in methanol, so the keyword is model 6. Atoms 7 to infinity are thus part of the sto-3g part, and atoms 1-6 part of the 6-31g* part.


memory stack 600 mb heap 200 mb global 800 mb

scratch_dir /scratch
Title "oniom"
Start oniom
echo
charge 0

geometry units angstrom
symmetry c1
C         10.12000       10.35000       10.00000
H          9.60000       10.72000       10.89000
H          9.60000       10.72000        9.11000
H         11.14000       10.74000       10.00000

[.]

H          0.46000        2.39000        6.86000

H          1.16000        1.73000        8.18000

end

basis sto-3g spherical

 * library sto-3g

end

basis 6-31g spherical

 * library 6-31g

end

oniom

    model 6

    low  dft basis sto-3g input "dft\; xc\; end"

    high dft basis "6-31g" input "dft\; xc\; end"

end

basis "ao basis" spherical print

  H library "6-31+G*"

  O library "6-31+G*"

  C library "6-31+G*"

END

task oniom

This takes forever to run, but run it does. The memory statement is important -- if the global memory is too small it will crash. Also, be aware that the amount of memory specified is per instance -- if you launch with mpirun -n 3, multiply by 3 to get the amount of memory that needs to be present (here, 2.4 GB physical RAM).