On Wed, Mar 12, 2008 at 8:46 PM, Nichols, = Patrick J=20 <patrick.nichols@pnl.gov>=20 wrote:

Chona,

    Two things I would suggest=20 adding.

nwpw

   ...

   cutoff 40.0

end

set=20 nwpw:lcao_skip .true.

Sorry, I always set my own cutoffs so I did not put = this into=20 the file. 40-50 is a good starting=20 point.

The lcao=20 skip will prevent another errquit from occurring = ( it=20 will try to generate the LCAO guess and find no=20 information).

Take=20 Care,

   Patrick

From: asheroth@gmail.com [mailto:asheroth@gmail.com] On=20 Behalf Of Chona S. Guiang
Sent: Wednesday, March 12, = 2008 3:57=20 PM
To: Nichols, Patrick J
Cc:=20 nwchem-users
Subject: Re: [NWCHEM] CPI=20 pseudopotentials

Hi Patrick,

   Thanks for the help on = this. =20 I downloaded the CPI file you enclosed and saved it to my permanent=20 directory.   My job made more progress this time, although = now I get=20 the following error/warning:

failed keyword recognition=20 forcutoff

I terminated the job for the time = being.

Please do=20 send me the format for a library file. That will be very useful = for=20 later.

Thanks,
Chona

On Wed, Mar 12, 2008 at 6:02 PM, Nichols, = Patrick J=20 <patrick.nichols@pnl.gov> wrote:

Chona,

   The user manual is a little "confusing" on = this topic.=20 There are two ways to read a cpi file into NWChem. From the library = (a file=20 in NWCHEM_NWPW_LIBRARY/CPI/ called just Ti e.g.)  in which = the=20 input should look like

nwpw

=20 ..

=20 pseudopotentials

      Ti library = CPI

      ...

  =20 end

=20 ...

end

The other way is to read it from a file in your = permanent=20 directory ( a file called ti.cpi) or if no permanent directory is = assigned=20 the same directory as your *.nw file. The input is as=20 so.

nwpw

  =20 ...

=20 pseudopotentials

       Ti CPI=20 ti.cpi

=20 end

end

I have attached a cpi file I = got from=20 the Opium homepage ( GGA ) for Ti and modified to = work=20 with NWChem. You can check if the format fits your own=20 file.

If=20 you wish to place to read this from a library the format = changes.=20 If=20 you place this cpi in your "permanent directory" and use the same = *.nw this=20 should work.

If you=20 want an example of the format for a library file I can send that = too.=20

The=20 format is a clunky too. In the cpi file, there should not be a blank line at the = end of the=20 file or bad things will happen ( infinite semicore corrections=20 usually).

Hope=20 this helps...

     Patrick = Nichols

From: owner-nwchem-users@emsl.pnl.gov [mailto:owner-nwchem-users@emsl.pnl.gov] On Behalf Of = Chona=20 S. Guiang
Sent: Wednesday, March 12, 2008 8:58 = AM
To:=20 nwchem-users
Subject: [NWCHEM] CPI=20 pseudopotentials

Hi,

   I'm trying to read in CPI=20 pseudopotentials that I generated from Opium, but I'm getting the = following=20 error:

     >>> JOB=20 STARTED       AT Wed Mar 12 11:45:44=20 2008 =20 = <<<
         =20 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D input data = =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D<= BR>Error: <CPI>=20 section not found

This is how I specified the use of the CPI=20 pseudopotentials in my input::

nwpw
  =20 ...
   = pseudopotentials
     =20 Ti CPI Ti.cpi
      O =20 CPI   O.cpi
   end
  =20 ...
end

as per the instructions in the manual and a = previous=20 thread on CPI in the support archives. I placed the files in = the=20 $NWCHEM_NWPW_LIBRARY/pspw_default directory I'm pretty sure this = variable is=20 defined because the output file does indicate the correct = path.

In=20 case I misread the instructions the first time, I also tried writing = "CPI"=20 on the first line of the CPI files but this produced the same error=20 message.

Is there something else I'm=20 = missing?

Thanks,
Chona
=