Detailed Description

Note: If parallel mode is distributed certain combination of "in_use" and "in_space" can not be used. For performance reasons it would be better to have the loops over g-vectros in the gather/scatter routines in new subprograms with the actual arrays (also the addressing) in the parameter list

History: JGH (29-Dec-2000) : Changes for parallel use JGH (13-Mar-2001) : added timing calls JGH (26-Feb-2003) : OpenMP enabled JGH (17-Nov-2007) : Removed mass arrays JGH (01-Dec-2007) : Removed and renamed routines 03.2008 [tlaino] : Splitting pw_types into pw_types and pw_methods