[CP2K:5202] Re: Implementing new functionals

[Michael Banck]

Hi,

On Tue, Apr 22, 2014 at 03:26:55PM -0700, August Melcher wrote:
> I'm currently getting this error when I run with the aug-cc-pVQZ basis, but 
> not with aug-cc-pVDZ. 
> 
>  ****************************************************************************
>  *** 15:21:22 ERRORL2 in hfx_libint_wrapper:initialize_libderiv processor 
> ***
>  *** 0  :: err=-300 the angular momentum needed exceeds the value assumed 
> ***
>  *** when configuring libderiv                                           
>  ***
>  ****************************************************************************

I can reproduce this with stock Debian/Ubuntu packages and the attached reduced
input file, which required much less memory.

There are two things here:

1. the HFX module appears to initialize libderiv, which should only be
required for forces, but your job is a single-point calculation.  I was
able to reproduce a failure with PSI3 when running a geometry
optimization of Ar2, but not for a single-point run, so I think CP2K
should be able to do it in principle as well.

2. The default --with-libderiv-max-am1 configure option for the libint
library is 3, and the aug-cc-pVQZ basis for Ar has an angular momentum
of 4.  So you can recompile libint with higher --with-libderiv-max-am1
(and maybe --with-libderiv-max-am2, just to be sure) options and link
CP2K against it.


Michael
-------------- next part --------------
A non-text attachment was scrubbed...
Name: ar2.inp
Type: chemical/x-gamess-input
Size: 883 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140423/1424b415/attachment.inp>