Implementing new functionals

[August Melcher]

I got -128.868055451878774 using that input file. However I was looking at 
the xc_b97.F file with Narbe today and he noticed what is likely the cause 
of the problem we've been noticing. Currently, the exchange energy density 
is being evaluated without attenuation. The exchange energy density in 
range-separated functionals uses the erfc(w*r)/r operator instead of 1/r, 
so there is an additional attenuation function (F) that appears in the 
integrand. The spin-polarized equation for F can be found in the wB97X-V 
paper (Equation 9). The only modification that needs to be made to the 
existing maple code is the inclusion of the F functional between e_lsda_x_a 
and gx_a (i.e. for the energy evaluation only, it would look like:

exc = scale_x * (e_lsda_x_a * *Fx_a* * gx_a + e_lsda_x_b * *Fx_b* *gx_b) + 
scale_c * (e_lsda_c_ab * gc_ab + e_lsda_c_a * gc_a + e_lsda_c_b * gc_b)

Would you mind modifying the source code or else directing me to the b97 
maple file so that I can do so?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20140604/2c612c7c/attachment.htm>