[Octopus-devel] cvs commit: marques/octopus/src derivatives.F90 derivatives_inc.F90 nl_operator.F90 poisson.F90 poisson3D.F90 poisson_cg.F90

[cvsusers at tddft.org]

acastro     2005/01/22 17:40:45 WET

  Modified files:
    octopus/src          derivatives.F90 derivatives_inc.F90 
                         nl_operator.F90 poisson.F90 poisson3D.F90 
                         poisson_cg.F90 
  Log:
  The CG_CORRECTED (PoissonSolver = 5) method is now able to work with
  curvilinear coordinates. The code will stop and complain if any other
  Poisson solver is to be used with curvilinear coordinates.
  
  However, it does not work with Gygy scheme, only
  with Modine's and Brigg's. I do not know why, but I thought it was about
  time to commit. In any case, I think there is a bug in the Gygy's
  implementation (the grid generated for one single atom in the center
  looks weird, with higher density of points in the border of the mesh).
  
  Revision  Changes    Path
  1.8       +11 -2     marques/octopus/src/derivatives.F90
  1.4       +30 -0     marques/octopus/src/derivatives_inc.F90
  1.4       +54 -25    marques/octopus/src/nl_operator.F90
  1.20      +5 -0      marques/octopus/src/poisson.F90
  1.23      +1 -1      marques/octopus/src/poisson3D.F90
  1.6       +36 -11    marques/octopus/src/poisson_cg.F90