[Octopus-devel] Fwd: [Octopus-notify] svn commit: r8577 - in trunk/src: td by joseba

Joseba Alberdi joseba.alberdi at ehu.es
Thu Nov 17 20:41:00 WET 2011 

Hi David,

you don't have to. If the documentation is before the
function/subroutine/type it is OK to be !>
For example:
  !> Describes mesh distribution to nodes.
  !!
  !! Some general things:
  !! All members of type(mesh_t) are equal on all
  !! nodes when running parallel except
  !! - np, np_part
  !! - x, vol_pp
  !! These four are defined for all the points the node is responsible for.
  type mesh_t

But if first of all is the variable, we have to put !<
For instance:
   integer  :: np               !< Local number of points in mesh

Hope is clear
Thanks!
Joseba

2011/11/17 David Strubbe <dstrubbe at berkeley.edu>

> The mailing list doesn't like my message because it was too long. I have
> abbreviated it below.
>
> ---------- Forwarded message ----------
> From: David Strubbe <dstrubbe at berkeley.edu>
> Date: Thu, Nov 17, 2011 at 11:21 AM
> Subject: Re: [Octopus-notify] svn commit: r8577 - in trunk/src: td by
> joseba
> To: Octopus-devel at tddft.org
>
>
> Joseba,
>
> If you grep for '!>' you will find a lot of other instances in the code.
> Should they be changed too?
>
> David
>
> src/basic/multicomm.F90:22:!> Given a number n_index of indices and their
> ranges index_range(1:n_index),
> src/basic/multicomm.F90:107:  !> Stores all communicators and groups
> src/basic/multicomm.F90:132:  !> An all-pairs communication schedule for a
> given group.
> src/basic/multicomm.F90:159:  !> create index and domain communicators
> src/basic/multicomm.F90:425:    !> check if a balanced distribution of
> nodes will be used
> src/basic/multicomm.F90:698:  !> This routine uses the one-factorization
> (or near-one-factorization
> src/basic/multicomm.F90:796:  !> Function to divide the range of numbers
> from 1 to nobjs
> src/basic/multicomm.F90:873:  !> Function to divide the range of numbers
> from 1 to nobjs
> src/basic/command_line.F90:211:  !> Initializes the getopt machinery. Must
> be called before attempting
> src/basic/blacs_proc_grid.F90:68:  !> Initializes a blacs context from an
> MPI communicator with
> src/basic/blacs.F90:23:!> This module contains interfaces for BLACS
> routines
> src/basic/unit_system.F90:24:!> Atomic weights should be read in "atomic
> mass units" (u) (not to
> src/basic/unit_system.F90:228:  !> These routines output the
> unit-conversion factors, defined by
> src/basic/unit_system.F90:322:  !> This is a very primitive procedure that
> attempts to find out
> src/basic/profiling.F90:22:  !>/* To use the profiling module you simply
> have to define a
> src/basic/profiling.F90:194:  !> Create profiling subdirectory.
> src/basic/profiling.F90:399:  !> Initialize a profile object and add it to
> the list
> src/basic/profiling.F90:455:  !> Increment in counter and save entry time.
> src/basic/profiling.F90:503:  !> Increment out counter and sum up
> difference between entry
> src/basic/profiling.F90:757:  !> Write profiling results of each node to
> profiling.NNN/profiling.nnn
> src/basic/topology.F90:22:!> Given a number n_index of indices and their
> ranges index_range(1:n_index),
> src/basic/topology.F90:78:    !> this routine tries to guess the
> distribution of the processors
> src/basic/calc_mode.F90:81:    !> Defines that the current run mode
> requires division of states
> src/basic/calc_mode.F90:89:    !> Whether the current run mode requires
> divisions compatible with
>  src/basic/mpi_lib_inc.F90:22:!> Every node has incount (may vary from
> node to node) items (in
> src/grid/mesh_batch_inc.F90:498:!> This functions exchanges points of a
> mesh according to a certain
> src/grid/cube.F90:219:  !> True if global coordinates belong to this
> process. On output
> src/grid/cube.F90:220:  !> lxyz contains the local coordinates
> src/grid/cube.F90:236:  !> do the mapping between global and local points
> of the cube
> src/grid/mesh_cube_map.F90:49:  !>  DOCUMENTATION NEEDED
> src/grid/mesh_partition.F90:61:  !> Converts the mesh given by grid points
> into a graph. Each
> src/grid/fourier_space_inc.F90:59:!> The following routines convert the
> function between real space and Fourier space
> src/grid/fourier_space_inc.F90:190:!> The next two subroutines convert a
> function in Fourier space
> src/grid/derivatives.F90:111:    !> If the so-called variational
> discretization is used, this controls a
> src/grid/derivatives.F90:301:  !> Returns maximum extension of the stencil
> in spatial direction
> src/grid/derivatives.F90:351:  !> Returns the diagonal elements of the
> Laplacian, needed for preconditioning
> src/grid/batch.F90:105:    !> We also need a linear array with the states
> in order to calculate derivatives, etc.
> src/grid/batch.F90:109:    !> If the memory is contiguous, we can perform
> some operations faster.
> src/grid/batch.F90:281:  !> When we are interested in batches of 1D
> functions
> src/grid/batch.F90:406:  !> THREADSAFE
> src/grid/batch.F90:415:  !> THREADSAFE
> src/grid/boundaries_inc.F90:21:!> Updates ghost points of every node. A
> vector suitable
> src/grid/cube_function_inc.F90:101:!> The next two subroutines convert a
> function between the normal
> src/grid/cube_function_inc.F90:263:!> The next two subroutines convert a
> function between the normal
> src/grid/mesh_init.F90:486:!> When running parallel in domains, stencil
> and np_stencil
> src/grid/mesh_init.F90:829:  !> calculate the volume of integration
> src/grid/par_vec_inc.F90:20:!> Generally:
> src/grid/par_vec_inc.F90:87:!> v_local has to be of length
> np_local+np_ghost+np_bndry
> src/grid/par_vec_inc.F90:154:!> Reverse operation of Xvec_scatter.
> src/grid/par_vec_inc.F90:205:!> Like Xvec_gather but the result is gathered
> src/grid/par_vec_inc.F90:248:!> This function collects points from the
> array src in nodes and puts
> src/grid/mesh.F90:73:  !> Describes mesh distribution to nodes.
> src/grid/mesh.F90:110:  !> This data type defines a plane, and a regular
> grid defined on
> src/grid/mesh.F90:126:  !> This data type defines a line, and a regular
> grid defined on this
> src/grid/mesh.F90:150:  !> finds the dimension of a box doubled in the
> non-periodic dimensions
> src/grid/mesh.F90:232:  !> Finds out if a given point of a mesh belongs to
> the "border" of the
> src/grid/mesh.F90:334:  !> Returns the index of the point which is nearest
> to a given vector
> src/grid/mesh.F90:414:  !> mesh_gcutoff returns the "natural" band
> limitation of the
> src/grid/mesh.F90:450:  !> Read the mesh parameters from file that were
> written by mesh_dump.
> src/grid/mesh.F90:560:  !> Fill the lxyz and lxyz_inv arrays from a file
> src/grid/mesh.F90:592:  !> Extracts the point numbers of a rectangular
> subset spanned
> src/grid/mesh.F90:639:  !> Checks if the (x, y, z) indices of point are
> valid, i.e.
> src/grid/grid.F90:79:  !>
>
> ... etc.
>
> On Thu, Nov 17, 2011 at 3:39 AM, <octopus-svn at tddft.org> wrote:
>
>>      Author: joseba
>>        Date: Thu Nov 17 11:39:01 2011
>> New Revision: 8577
>>   Changeset: http://www.tddft.org/trac/octopus/changeset/8577
>>
>> Modified:
>>   trunk/src/basic/multicomm.F90
>>   trunk/src/td/pes.F90
>>
>> Log:
>> * Initilized the rank, if is called without OpenMP
>> * Added and corrected documentation. With doxygen the "<" or ">" has to
>> point to the source is documenting
>>
>>
>> _______________________________________________
>> Octopus-notify mailing list
>> Octopus-notify at tddft.org
>> http://www.tddft.org/mailman/listinfo/octopus-notify
>>
>
>
>
> _______________________________________________
> Octopus-devel mailing list
> Octopus-devel at tddft.org
> http://www.tddft.org/mailman/listinfo/octopus-devel
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.tddft.org/pipermail/octopus-devel/attachments/20111117/fc013c2d/attachment-0001.html>

More information about the Octopus-devel mailing list