[Octopus-users] XC Library in parallel

Tue Oct 6 09:50:26 WEST 2009

Hi David,

Here is an example of an inp file that sets the negative coordinates to zero
when run in parallel. Also if you have any suggestions as to why the code is
not able to find the XC library in parallel I would be grateful - I have
used openmpi to build.

*CalculationMode = go        # go=geometry optimisation | gs = ground state
| unocc = unnocupied states.............
units = eV_Angstrom
FromScratch = yes

#Geometry Optimisation Values
GOMaxIter   = 100        #Max no of interations
GOMethod    = cg_bfgs2         #Optimisation Method: |steep = steep descent|
cg_fr = fletcher-Reeves cg | cg_pr = Polak-Ribiere cg | cg_bfgs =
VectorBroyden-Flecther-Goldfarb-Shanno quasi-newton | cg_bfgs = More
efficieint version
GOTolerence = 0.05        # Force stopping criteria
GOStep      = 0.2

#Simulation Box
BoxShape = sphere
Spacing  = 0.13
Radius   = 10

%Species
  'N' | 14 | spec_ps_psf | 7 | 1 | 0
  'C' | 12 | spec_ps_psf | 6 | 1 | 0
  'H' | 1  | spec_ps_psf | 1 | 1 | 0
%

#XYZCoordinates = "IND.xyz"
%Coordinates
"N"  |   -2.009895 | -0.375276  |   0.219476| yes
"C"  |   -0.773073 |  1.493705  |  -0.492799| yes
"C"  |    0.067581 |  0.542777  |   0.299037| yes
"C"  |   -2.044650 |  0.684557  |  -0.772146| yes
"C"  |   -0.725321 | -0.537337  |   0.684986| yes
"C"  |    1.378760 |  0.630671  |   0.707703| yes
"C"  |   -0.199670 | -1.551800  |   1.471828| yes
"C"  |    1.919562 | -0.379852  |   1.500712| yes
"C"  |    1.129735 | -1.455652  |   1.873242| yes
"H"  |   -0.284686 |  1.839696  |  -1.419337| yes
"H"  |   -0.995268 |  2.402804  |   0.098843| yes
"H"  |   -2.023651 |  0.276702  |  -1.806465| yes
"H"  |   -2.962904 |  1.290181  |  -0.681757| yes
"H"  |   -2.617172 | -1.181502  |   0.107667| yes
"H"  |    1.990779 |  1.493972  |   0.416522| yes
"H"  |   -0.815082 | -2.403512  |   1.786756| yes
"H"  |    2.961415 | -0.322358  |   1.827095| yes
"H"  |    1.556359 | -2.244437  |   2.500590| yes
%

#XCFunctional = gga_x_pw91 +  gga_c_pw91         *

Thanks again for your help.

Conn

On Tue, Oct 6, 2009 at 12:06 AM, David Strubbe <dstrubbe at berkeley.edu>wrote:

> Can you send an input file that suffers from this problem about coordinates
> being set to zero?
>
> Thanks,
> David
>
>
> On Mon, Oct 5, 2009 at 2:31 PM, Conn O'Rourke <conn.orourke at googlemail.com
> > wrote:
>
>> Hi David,
>>
>> Sorry about about being vague.
>>
>> If I input the coordinates in a block in the inp file and run in parallel
>> the coordinates get set to zero, a problem which doesn't occur if octopus
>> read them in from an .xyz file.
>>
>> If I run the code with the same inp file in serial the code accepts the
>> coordinates as negative.
>>
>> Thanks,
>> Conn
>>
>>
>> On Mon, Oct 5, 2009 at 8:09 PM, David Strubbe <dstrubbe at berkeley.edu>wrote:
>>
>>> Conn,
>>>
>>> Can you be more specific about what "seems to dislike having negative
>>> coordinate values" means?
>>>
>>> David Strubbe
>>>
>>> On Mon, Oct 5, 2009 at 10:31 AM, Conn O'Rourke <
>>> conn.orourke at googlemail.com> wrote:
>>>
>>>> Hi All,
>>>>
>>>> I have built the mpi version of octopus-3.1.0 and it runs fine with the
>>>> default LDA  XC functional, but when I try to use a different functional the
>>>> code can't seem to find it.
>>>> I have added the libxc.a path to $LD_LIBRARY_PATH but it still cannot
>>>> find it and a blank is left under the theory level section of the log file:
>>>>
>>>> **************************** Theory Level ****************************
>>>> Input: [TheoryLevel = dft]
>>>>
>>>> Exchange and correlation:
>>>>
>>>>
>>>> Any suggestions as to what I might have done incorrectly?
>>>>
>>>> Also the code in parallel seems to dislike having negative coordinate
>>>> values if input in the inp file, but accepts them if input as a .xyz file.
>>>> is this a known bug or something I have done incorrectly in my build.
>>>>
>>>> Thanks in advance.
>>>>
>>>> Conn
>>>>
>>>> _______________________________________________
>>>> Octopus-users mailing list
>>>> Octopus-users at tddft.org
>>>> http://www.tddft.org/mailman/listinfo/octopus-users
>>>>
>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.tddft.org/pipermail/octopus-users/attachments/20091006/5ec4b0ca/attachment-0001.html 