The ABINIT Users Mailing List ( CLOSED )

[deyulu@yahoo.com]

Dear Abinit users:
     This is a technical question regarding the program DP 
(http://lepes.grenoble.cnrs.fr/theorie/olevano/dp/index.html), and I was 
advised to post the qustion in this forum. Hope people do not mind.
As I understand, in order to compute \epsilon_m, one only needs to know the 
first column of \chi by solving a linear equation. In the subroutine 
"creation_epsrpa", however, \epsilon_m is obtained by inverting the full 
matrix, which slows down the calculation. Is there any particular concern in 
this subroutine?

     Thanks
     Deyu