Cloned SEACAS for EXODUS library with extra build files for internal package management.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 

970 lines
39 KiB

<!-------- @HEADER
!
! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!
! Zoltan Toolkit for Load-balancing, Partitioning, Ordering and Coloring
! Copyright 2012 Sandia Corporation
!
! Under the terms of Contract DE-AC04-94AL85000 with Sandia Corporation,
! the U.S. Government retains certain rights in this software.
!
! Redistribution and use in source and binary forms, with or without
! modification, are permitted provided that the following conditions are
! met:
!
! 1. Redistributions of source code must retain the above copyright
! notice, this list of conditions and the following disclaimer.
!
! 2. Redistributions in binary form must reproduce the above copyright
! notice, this list of conditions and the following disclaimer in the
! documentation and/or other materials provided with the distribution.
!
! 3. Neither the name of the Corporation nor the names of the
! contributors may be used to endorse or promote products derived from
! this software without specific prior written permission.
!
! THIS SOFTWARE IS PROVIDED BY SANDIA CORPORATION "AS IS" AND ANY
! EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
! IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
! PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL SANDIA CORPORATION OR THE
! CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
! EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
! PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
! PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
! LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
! NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
! SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
!
! Questions? Contact Karen Devine kddevin@sandia.gov
! Erik Boman egboman@sandia.gov
!
! !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
!
! @HEADER
------->
<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<meta name="GENERATOR" content="Mozilla/4.76 [en] (X11; U; Linux 2.4.2-2smp i686) [Netscape]">
<meta name="sandia.approval_type" content="formal">
<meta name="sandia.approved" content="SAND2007-4748W">
<meta name="author" content="Zoltan PI">
<title>Zoltan User's Guide: Load-Balancing Interface</title>
</head>
<body bgcolor="#FFFFFF">
<div align=right><b><i><a href="ug.html">Zoltan User's Guide</a>&nbsp;
|&nbsp; <a href="ug_interface_augment.html">Next</a>&nbsp; |&nbsp; <a href="ug_interface_init.html">Previous</a></i></b></div>
<h2>
<a NAME="Load-Balancing Functions"></a>Load-Balancing Functions</h2>
The following functions are the load-balancing interface functions in the
Zoltan library; their descriptions are included below.
<blockquote><b><a href="#Zoltan_LB_Partition">Zoltan_LB_Partition</a></b>
<br><b><a href="#Zoltan_LB_Set_Part_Sizes">Zoltan_LB_Set_Part_Sizes</a></b>
<br><b><a href="#Zoltan_LB_Eval">Zoltan_LB_Eval</a></b>
<br><b><a href="#Zoltan_LB_Free_Part">Zoltan_LB_Free_Part</a></b></blockquote>
For <a href="ug_backward.html">backward compatibility</a> with previous
versions of Zoltan, the following functions are also maintained. These
functions are applicable only when the number of parts to be generated
is equal to the number of processors on which the parts are computed.
That is, these functions assume "parts" and "processors" are synonymous.
<blockquote>
<b><a href="#Zoltan_LB_Balance">Zoltan_LB_Balance</a></b>
<br><b><a href="#Zoltan_LB_Free_Data">Zoltan_LB_Free_Data</a></b>
</blockquote>
Descriptions of algorithm-specific interface functions are included with
the documentation of their associated algorithms.
Algorithm-specific functions include:
<blockquote>
<b><a href="ug_alg_rcb.html#Zoltan_RCB_Box">Zoltan_RCB_Box</a></b>
</blockquote>
<!------------------------------------------------------------------------->
<hr WIDTH="100%"><a NAME="Zoltan_LB_Partition"></a>
<hr WIDTH="100%">
<table WIDTH="100%" NOSAVE >
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C:</td>
<td WIDTH="85%">int <b>Zoltan_LB_Partition</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; struct <b>Zoltan_Struct</b> *<i>zz</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>changes</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_gid_entries</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_lid_entries</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_import</i>,&nbsp;
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>import_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>import_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>import_procs</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>import_to_part</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_export</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>export_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>export_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>export_procs</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>export_to_part</i>);</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP WIDTH="15%" NOSAVE>FORTRAN:</td>
<td>FUNCTION <b>Zoltan_LB_Partition</b>(<i>zz, changes, num_gid_entries,
num_lid_entries, num_import, import_global_ids, import_local_ids, import_procs,
import_to_part, num_export, export_global_ids, export_local_ids, export_procs,
export_to_part</i>)&nbsp;
<br>INTEGER(Zoltan_INT) :: Zoltan_LB_Partition&nbsp;
<br>TYPE(Zoltan_Struct), INTENT(IN) :: zz&nbsp;
<br>LOGICAL, INTENT(OUT) :: changes&nbsp;
<br>INTEGER(Zoltan_INT), INTENT(OUT) :: num_gid_entries, num_lid_entries
<br>INTEGER(Zoltan_INT), INTENT(OUT) :: num_import, num_export&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_global_ids, export_global_ids&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_local_ids, export_local_ids&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_procs, export_procs
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_to_part, export_to_part</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C++:</td>
<td WIDTH="85%">int <b>Zoltan::LB_Partition</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int &<i>changes</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int &<i>num_gid_entries</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int &<i>num_lid_entries</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int &<i>num_import</i>,&nbsp;
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
&<i>import_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
&<i>import_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int * &<i>import_procs</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int * &<i>import_to_part</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int &<i>num_export</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
&<i>export_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
&<i>export_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int * &<i>export_procs</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int * &<i>export_to_part</i>);</td>
</tr>
</table>
<hr WIDTH="100%"><b>Zoltan_LB_Partition</b> invokes the load-balancing
routine specified by the <i><a href="ug_alg.html#LB_METHOD">LB_METHOD</a></i>
parameter. The number of parts it generates is specified by the <i><a href="ug_alg.html#NUM_GLOBAL_PARTS">NUM_GLOBAL_PARTS</a></i>
or
<i><a href="ug_alg.html#NUM_LOCAL_PARTS">NUM_LOCAL_PARTS</a></i>
parameters. Results of the partitioning are returned in lists of objects
to be imported into and exported from parts on this processor.
Objects are included in these lists if <i>either</i> their part
assignment or their processor assignment is changed by the new decomposition.
If an application requests multiple parts on a single processor, these
lists may include objects whose part assignment is changing, but whose
processor assignment is unchanged.
<p>
Returned arrays are allocated in Zoltan; applications
should not allocate these arrays before calling <b>Zoltan_LB_Partition</b>.
The arrays are later freed through calls to <b><a href="#Zoltan_LB_Free_Part">Zoltan_LB_Free_Part</a></b>.
<br>&nbsp;
<table WIDTH="100%" NOSAVE >
<tr>
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
<td WIDTH="80%"></td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; zz</i></td>
<td>Pointer to the Zoltan structure, created by <b><a href="ug_interface_init.html#Zoltan_Create">Zoltan_Create</a></b>,
to be used in this invocation of the load-balancing routine.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; changes</i></td>
<td>Set to 1 or .TRUE. if the decomposition was changed by the load-balancing
method; 0 or .FALSE. otherwise.</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>&nbsp;&nbsp;&nbsp; <i>num_gid_entries</i></td>
<td>Upon return, the number of array entries used to describe a single
global ID.&nbsp; This value is the maximum value over all processors of
the parameter <a href="ug_param.html#NUM_GID_ENTRIES">NUM_GID_ENTRIES</a>.</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>&nbsp;&nbsp;&nbsp; <i>num_lid_entries</i></td>
<td>Upon return, the number of array entries used to describe a single
local ID.&nbsp; This value is the maximum value over all processors of
the parameter <a href="ug_param.html#NUM_LID_ENTRIES">NUM_LID_ENTRIES</a>.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; num_import&nbsp;</i></td>
<td>Upon return, the number of objects
that are newly assigned to this processor or to parts on this processor
(i.e., the number of objects being imported from different parts to
parts on this processor).
If the value returned
is -1, no import information has been returned and all import arrays below
are NULL. (The
<a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a> parameter
determines whether import lists are returned).</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; import_global_ids</i></td>
<td>Upon return, an array of <i>num_import</i> global IDs of objects to
be imported to parts on this processor.
<br>(size = <i>num_import</i> * <i>num_gid_entries</i>)</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; import_local_ids</i></td>
<td>Upon return, an array of&nbsp; <i>num_import</i> local IDs of objects
to be imported to parts on this processor.
<br>(size = <i>num_import</i> * <i>num_lid_entries</i>)</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; import_procs</i></td>
<td>Upon return, an array of size <i>num_import</i> listing the processor
IDs of the processors that owned the imported objects in the previous decomposition
(i.e., the source processors).</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; import_to_part</i></td>
<td>Upon return, an array of size <i>num_import</i> listing the parts
to which the imported objects are being imported.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; num_export</i></td>
<td>Upon return, this value of this count and the following lists
depends on the value of
the <a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a> parameter:
<UL>
<LI>It is the count of objects on this processor
that are newly assigned to other processors or to other parts
on this processor, if
<a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a>
is "EXPORT" or "EXPORT AND IMPORT".
<LI>It is the count of all objects on this processor,
if
<a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a>
is "PARTS" (or "PART ASSIGNMENTS").
<LI>It is -1 if the value of
<a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a>
indicates that either no lists are to be returned, or only import lists
are to be returned.
If the value returned
is -1, no export information has been returned and all export arrays below
are NULL .
</UL>
</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; export_global_ids</i></td>
<td>Upon return, an array of <i>num_export</i> global IDs of objects to
be exported from parts on this processor
(if <a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a>
is equal to "EXPORT" or "EXPORT AND IMPORT"),
or an array of <i>num_export</i> global IDs
for every object on this processor
(if <a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a>
is equal to "PARTS" or "PART ASSIGNMENTS"),
.&nbsp;
<br>(size = <i>num_export</i> * <i>num_gid_entries</i>)</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; export_local_ids</i></td>
<td>Upon return, an array of <i>num_export</i> local IDs associated
with the global IDs returned in <i>export_global_ids</i>
<br>(size = <i>num_export</i> * <i>num_lid_entries</i>)</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; export_procs</i></td>
<td>Upon return, an array of size <i>num_export </i>listing the processor
ID of the processor to which each object is now assigned
(i.e., the destination processor).
If <a href="ug_alg.html#RETURN_LISTS">RETURN_LISTS</a>
is equal to "PARTS" or "PART ASSIGNMENTS", this list includes all objects, otherwise
it only includes the objects which are moving to a new part and/or process.
</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; export_to_part</i></td>
<td>Upon return, an array of size <i>num_export</i> listing the parts
to which the objects are assigned under the new partition.</td>
</tr>
<tr>
<td><b>Returned Value:</b></td>
<td></td>
</tr>
<tr>
<td VALIGN=TOP>&nbsp;&nbsp;&nbsp; int</td>
<td><a href="ug_interface.html#Error Codes">Error code</a>.</td>
</tr>
</table>
<p><!------------------------------------------------------------------------->
<hr WIDTH="100%"><a NAME="Zoltan_LB_Set_Part_Sizes"></a>
<hr WIDTH="100%">
<table WIDTH="100%" NOSAVE >
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C:</td>
<td WIDTH="85%">int <b>Zoltan_LB_Set_Part_Sizes</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; struct <b>Zoltan_Struct</b> *<i>zz</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int <i>global_num,</i>
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int <i>len</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>part_ids,</i>
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>wgt_idx</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; float *<i>part_sizes</i>);</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP WIDTH="15%" NOSAVE>FORTRAN:</td>
<td WIDTH="85%">
function <b>Zoltan_LB_Set_Part_Sizes</b>(
<i>zz,global_part,len,partids,wgtidx,partsizes</i>)
<br>integer(Zoltan_INT) :: Zoltan_LB_Set_Part_Sizes
<br>type(Zoltan_Struct) INTENT(IN) zz
<br>integer(Zoltan_INT) INTENT(IN) global_part,len,partids(*),wgtidx(*)
<br>real(Zoltan_FLOAT) INTENT(IN) partsizes(*)
</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C++:</td>
<td WIDTH="85%">int <b>Zoltan::LB_Set_Part_Sizes</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; const int &<i>global_num,</i>
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; const int &<i>len</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>part_ids,</i>
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>wgt_idx</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; float *<i>part_sizes</i>);</td>
</tr>
</table>
<hr WIDTH="100%"><b>Zoltan_LB_Set_Part_Sizes</b> is used to specify the
desired part sizes in Zoltan. By default, Zoltan assumes that all
parts should be of equal size.&nbsp; With <b>Zoltan_LB_Set_Part_Sizes</b>,
one can specify the relative (not absolute) sizes of the parts. For
example, if two parts are requested and the desired sizes are 1 and
2, that means that the first part will be assigned approximately one
third of the total load. If the sizes were instead given as 1/3 and 2/3,
respectively, the result would be exactly the same. Note that if there
are multiple weights per object, one can (must) specify the part size
for each weight dimension independently.
<br>&nbsp;
<table WIDTH="100%" NOSAVE >
<tr>
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
<td WIDTH="80%"></td>
</tr>
<tr>
<td><i>&nbsp;&nbsp;&nbsp; zz</i></td>
<td>Pointer to the Zoltan structure created by <b><a href="#Zoltan_Create">Zoltan_Create</a></b>.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp; global_num</i></td>
<td>Set to 1 if global part numbers are given, 0 otherwise (local
part numbers).</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp; len</i></td>
<td>Length of the next three input arrays.</td>
</tr>
<tr>
<td>&nbsp;&nbsp; <i>part_ids</i></td>
<td>Array of part numbers, either global or local. (Part numbers
are integers starting from 0.)</td>
</tr>
<tr VALIGN=TOP NOSAVE>
<td>&nbsp;&nbsp; <i>vwgt_idx</i></td>
<td NOSAVE>Array of weight indices (between 0 and OBJ_WEIGHT_DIM-1). This
array should contain all zeros when there is only one weight per object.</td>
</tr>
<tr VALIGN=TOP NOSAVE>
<td>&nbsp;&nbsp; <i>part_sizes</i></td>
<td NOSAVE>Relative values for part sizes; <i>part_sizes[i]</i> is
the desired relative size of the <i>vwgt_idx[i]</i>'th weight of part
<i>part_ids[i].</i></td>
</tr>
<tr>
<td><b>Returned Value:</b></td>
<td></td>
</tr>
<tr>
<td VALIGN=TOP>&nbsp;&nbsp;&nbsp; int</td>
<td><a href="ug_interface.html#Error Codes">Error code</a>.</td>
</tr>
<tr>
<td></td>
<td></td>
</tr>
<tr>
<td></td>
<td></td>
</tr>
</table>
<!------------------------------------------------------------------------->
<hr WIDTH="100%"><a NAME="Zoltan_LB_Eval"></a>
<hr WIDTH="100%">
<table WIDTH="100%" NOSAVE >
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C:</td>
<td WIDTH="85%">int <b>Zoltan_LB_Eval</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; struct <b>Zoltan_Struct</b> *<i>zz</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int <i>print_stats</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ZOLTAN_BALANCE_EVAL *<i>obj_info</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ZOLTAN_GRAPH_EVAL *<i>graph_info</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ZOLTAN_HG_EVAL *<i>hg_info</i>);&nbsp;</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP WIDTH="15%" NOSAVE>FORTRAN:</td>
<td>FUNCTION <b>Zoltan_LB_Eval</b>(<i>zz, print_stats</i>)&nbsp;
<br>INTEGER(Zoltan_INT) :: Zoltan_LB_Eval
<br>TYPE(Zoltan_Struct), INTENT(IN) :: zz&nbsp;
<br>LOGICAL, INTENT(IN) :: print_stats&nbsp;
</tr>
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C++:</td>
<td WIDTH="85%">int <b>Zoltan::LB_Eval</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; const int &<i>print_stats</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ZOLTAN_BALANCE_EVAL *<i>obj_info</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ZOLTAN_GRAPH_EVAL *<i>graph_info</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; ZOLTAN_HG_EVAL *<i>hg_info</i>);&nbsp;</td>
</tr>
</table>
<hr WIDTH="100%"><b>Zoltan_LB_Eval </b>evaluates the quality of a decomposition.
The quality metrics of interest differ depending on how you are using Zoltan.
If you are partitioning points in space using one of Zoltan's geometric methods,
you will want to know the weighted balance of objects across parts. However
if you are partitioning a graph, you will want to know about edges that have
vertices in more than one part. <b>Zoltan_LB_Eval </b> can write three
different structures with these differing metrics.
<p>
<table WIDTH="100%" NOSAVE >
<tr>
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
<td WIDTH="80%"></td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; zz</i></td>
<td>Pointer to the Zoltan structure.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; print_stats</i></td>
<td>If <i>print_stats</i>>0 (.TRUE. in Fortran), print the quality metrics to <i>stdout.</i></td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; obj_info</i></td>
<td>If <i>obj_info</i>is non-NULL, write object balance values to the <a href="#ZOLTAN_BALANCE_EVAL">ZOLTAN_BALANCE_EVAL</a> structure.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; graph_info</i></td>
<td>If <i>graph_info</i>is non-NULL, write graph partition metrics to the <a href="#ZOLTAN_GRAPH_EVAL">ZOLTAN_GRAPH_EVAL</a> structure.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; hg_info</i></td>
<td>If <i>hg_info</i>is non-NULL, write hypergraph partition metrics to the <a href="#ZOLTAN_HG_EVAL">ZOLTAN_HG_EVAL</a> structure.</td>
</tr>
<tr>
<td VALIGN=TOP><b>Returned Value:</b></td>
<td></td>
</tr>
<tr>
<td VALIGN=TOP>&nbsp;&nbsp;&nbsp; int</td>
<td><a href="ug_interface.html#Error Codes">Error code.</a></td>
</tr>
</table>
<p>
The EVAL structures are defined in <i>zoltan/src/include/zoltan_eval.h</i>.
Several of the fields in the EVAL structures are arrays of values.
The arrays contain values for
<ol>
<li>the total for the local process
<li>the total across all parts
<li>the minimum across all parts
<li>the maximum across all parts
<li>the average across all parts
</ol>
<p>
in that order. The corresponding macros that refer to these fields are:
<p>
<code>
#define EVAL_LOCAL_SUM 0<br>
#define EVAL_GLOBAL_SUM 1<br>
#define EVAL_GLOBAL_MIN 2<br>
#define EVAL_GLOBAL_MAX 3<br>
#define EVAL_GLOBAL_AVG 4<br>
</code>
<p>
<a NAME="ZOLTAN_BALANCE_EVAL"></a>
The ZOLTAN_BALANCE_EVAL structure contains the following fields, and
would be of interest if you are doing geometric partitioning:
<p>
<UL>
<li><strong>obj_imbalance</strong>: imbalance in count of objects across parts, scaled by
requested part sizes.
<li><strong>imbalance</strong>: imbalance in weight of objects across parts, scaled by
requested part sizes.
<li><strong>nobj</strong>: an array containing the number of objects on the local <em>process</em>, the total number of objects, the minimum number of objects in a <em>part</em>, the maximum number in a partitition, and the average number of objects across the parts.
<li><strong>obj_wgt</strong>: an array containing the sum of the weight of objects on the local <em>process</em>, the total weight across all processes, the minimum weight in a <em>part</em>, the maximum weight in a partitition, and the average weight of objects across the parts.
<li><strong>xtra_imbalance</strong>: if the <a href="ug_param.html#OBJ_WEIGHT_DIM">OBJ_WEIGHT_DIM</a> exceeds one, the <strong>obj_imbalance</strong> value for the extra weights is in this array.
<li><strong>xtra_obj_wgt</strong>: if the <a href="ug_param.html#OBJ_WEIGHT_DIM">OBJ_WEIGHT_DIM</a> exceeds one, the <strong>obj_wgt</strong> array for the extra weights is in this array.
</UL>
<p>
<a NAME="ZOLTAN_GRAPH_EVAL"></a>
The ZOLTAN_GRAPH_EVAL structure contains the following fields, and
would be of interest if you are doing graph partitioning:
<p>
<UL>
<li><strong>cutl</strong>: an array containing what is known at the CUTL or the ConCut measure of the graph, the first element is not set (this is the local process total, which has no meaning), the next is the sum of the CUTL across parts, then the minimum CUTL across parts, then the maximum CUTL across parts, finally the average CUTL across parts.
<li><strong>cutn</strong>: an array containing what is known at the CUTN or the NetCut measure of the graph, the first element is not set, the next is the sum of the CUTL across parts, then the minimum CUTL across parts, then the maximum CUTL across parts, finally the average CUTL across parts.
<li><strong>cuts</strong>: an array counting the number of cut edges (the first element again is not set) across all parts, the minimum across parts of cut edges, the maximum, and then the average
<li><strong>cut_weight</strong>: an array summing the weight of cut edges (the first element again is not set) across all parts, the minimum across parts, the maximum, and then the average
<li><strong>nnborparts</strong>: an array which counts the number of neighboring parts that each part has, the first element again is not set, the next is the sum across all parts, the minimum across parts, the maximum, and then the average
<li><strong>obj_imbalance</strong>: imbalance in count of objects across parts, scaled by
requested part sizes.
<li><strong>imbalance</strong>: imbalance in weight of objects across parts, scaled by
requested part sizes.
<li><strong>nobj</strong>: an array containing the number of objects on the local <em>process</em>, the total number of objects, the minimum number of objects in a <em>part</em>, the maximum number in a partitition, and the average number of objects across the parts.
<li><strong>obj_wgt</strong>: an array containing the sum of the weight of objects on the local <em>process</em>, the total weight across all processes, the minimum weight in a <em>part</em>, the maximum weight in a partitition, and the average weight of objects across the parts.
<li><strong>num_boundary</strong>: an array which counts the number of objects in a part that have at least one remote neighbor, the first element is not set, the next is the sum across parts, the next is the minimum count in any part, the next is the maximum, and then the average for all parts
<li><strong>xtra_imbalance</strong>: if the <a href="ug_param.html#OBJ_WEIGHT_DIM">OBJ_WEIGHT_DIM</a> exceeds one, the <strong>obj_imbalance</strong> value for the extra weights is in this array.
<li><strong>xtra_obj_wgt</strong>: if the <a href="ug_param.html#OBJ_WEIGHT_DIM">OBJ_WEIGHT_DIM</a> exceeds one, the <strong>obj_wgt</strong> array for the extra weights is in this array.
<li><strong>xtra_cut_wgt</strong>: if the <a href="ug_param.html#EDGE_WEIGHT_DIM">EDGE_WEIGHT_DIM</a> exceeds one, the <strong>cut_wgt</strong> array for the each extra weight is in this array.
</UL>
<p>
<p>
<a NAME="ZOLTAN_HG_EVAL"></a>
The ZOLTAN_HG_EVAL structure contains the following fields, and
would be of interest if you are doing hypergraph partitioning:
<p>
<UL>
<li><strong>obj_imbalance</strong>: imbalance in count of objects across parts, scaled by
requested part sizes.
<li><strong>imbalance</strong>: imbalance in weight of objects across parts, scaled by
requested part sizes.
<li><strong>cutl</strong>: an array containing what is known at the CUTL or the ConCut measure of the hypergraph, the first element is not set (this is the local process total, which has no meaning), the next is the sum of the CUTL across parts, then the minimum CUTL across parts, then the maximum CUTL across parts, finally the average CUTL across parts.
<li><strong>cutn</strong>: an array containing what is known at the CUTN or the NetCut measure of the hypergraph, the first element is not set, the next is the sum of the CUTL across parts, then the minimum CUTL across parts, then the maximum CUTL across parts, finally the average CUTL across parts.
<li><strong>nobj</strong>: an array containing the number of objects on the local <em>process</em>, t
he total number of objects, the minimum number of objects in a <em>part</em>, the maximum number
in a partitition, and the average number of objects across the parts.
<li><strong>obj_wgt</strong>: an array containing the sum of the weight of objects on the local <em>p
rocess</em>, the total weight across all processes, the minimum weight in a <em>part</em>, the m
aximum weight in a part, and the average weight of objects across the parts.
<li><strong>xtra_imbalance</strong>: if the <a href="ug_param.html#OBJ_WEIGHT_DIM">OBJ_WEIGHT_DIM</a>
exceeds one, the <strong>obj_imbalance</strong> value for the extra weights is in this array.
<li><strong>xtra_obj_wgt</strong>: if the <a href="ug_param.html#OBJ_WEIGHT_DIM">OBJ_WEIGHT_DIM</a> e
xceeds one, the <strong>obj_wgt</strong> array for the extra weights is in this array.
</UL>
<p>
<table WIDTH="100%" NOSAVE >
<tr VALIGN=TOP NOSAVE>
<td><b>Query functions:</b>
</td>
</tr>
<tr>
<td>
<br>&nbsp;&nbsp;&nbsp;&nbsp; Required:</td>
<td NOSAVE>
<br><a href="ug_query_lb.html#ZOLTAN_NUM_OBJ_FN">ZOLTAN_NUM_OBJ_FN</a>
and <a href="ug_query_lb.html#ZOLTAN_OBJ_LIST_FN">ZOLTAN_OBJ_LIST_FN</a>
</td>
</tr>
<tr VALIGN=TOP NOSAVE>
<td>&nbsp;&nbsp;&nbsp;&nbsp; Optional:</td>
<td NOSAVE>
<a href="ug_query_lb.html#Graph-based Functions">Graph-Based</a> functions are required
for writing the <a href="#ZOLTAN_GRAPH_EVAL">ZOLTAN_GRAPH_EVAL</a> structure.
<br>
<a href="ug_query_lb.html#Hypergraph-based Functions">Hypergraph-Based</a> functions will be used
for writing the <a href="#ZOLTAN_HG_EVAL">ZOLTAN_HG_EVAL</a> if they are available,
otherwise Zoltan will create a hypergraph from the
<a href="ug_query_lb.html#Graph-based Functions">graph-Based</a> functions.
<br>
If <a href="ug_interface_lb.html#Zoltan_LB_Set_Part_Sizes">Zoltan_LB_Set_Part_Sizes</a> has been
called, the part sizes set by the <a href="ug_interface_lb.html#Zoltan_LB_Set_Part_Sizes">Zoltan_LB_Set_Part_Sizes</a>
will be used in calculating imbalances.
</td>
</tr>
</table>
<p>Note that the sum of <i>ncuts</i> over all processors is actually twice
the number of edges cut in the graph (because each edge is counted twice).
The same principle holds for <i>cut_wgt.</i><i></i>
<p>There are a few improvements in Zoltan_LB_Eval&nbsp; in Zoltan version
1.5 (or higher). First, the balance data are computed with respect to both
processors and parts (if applicable). Second, the desired part
sizes (as set by Zoltan_LB_Set_Part_Sizes)&nbsp; are taken into account
when computing the imbalance.
<p>Known bug: If a part is spread across several processors, the computed
cut information (<i>ncuts </i>and <i>cut_wgt)</i> may be incorrect (too
high).
<p><!------------------------------------------------------------------------->
<hr WIDTH="100%"><a NAME="Zoltan_LB_Free_Part"></a>
<hr WIDTH="100%">
<table WIDTH="100%" NOSAVE >
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C:</td>
<td WIDTH="85%">int <b>Zoltan_LB_Free_Part</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>procs</i>,&nbsp;
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>to_part</i>);</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP WIDTH="15%" NOSAVE>FORTRAN:</td>
<td>FUNCTION <b>Zoltan_LB_Free_Part</b>(<i>global_ids, local_ids, procs,
to_part</i>)
<br>INTEGER(Zoltan_INT) :: Zoltan_LB_Free_Part&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: global_ids
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: local_ids
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: procs, to_part</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C++:</td>
<td WIDTH="85%">int <b>Zoltan::LB_Free_Part</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>procs</i>,&nbsp;
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>to_part</i>);</td>
</tr>
</table>
<hr WIDTH="100%"><b>Zoltan_LB_Free_Part</b> frees the memory allocated
by Zoltan to return the results of <b><a href="#Zoltan_LB_Partition">Zoltan_LB_Partition</a></b>
or
<b><a href="ug_interface_mig.html#Zoltan_Invert_Lists">Zoltan_Invert_Lists</a></b>.
Memory pointed to by the arguments is freed and the arguments are set to
NULL in C and C++ or nullified in Fortran. NULL arguments may be passed to <b>Zoltan_LB_Free_Part</b>.
Typically, <b>Zoltan_LB_Free_Part</b> is called twice: once for the import lists, and once for the export lists.<br>
Note that this function does not destroy the Zoltan data structure itself;
it is deallocated through a call to <b><a href="ug_interface_init.html#Zoltan_Destroy">Zoltan_Destroy</a></b>
in C and Fortran and by the object destructor in C++.
<br>
<table WIDTH="100%" >
<tr>
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
<td WIDTH="80%"></td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; global_ids</i></td>
<td>An array containing the global IDs of objects.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; local_ids</i></td>
<td>An array containing the local IDs of objects.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; procs</i></td>
<td>An array containing processor IDs.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; to_part</i></td>
<td>An array containing part numbers.</td>
</tr>
<tr>
<td VALIGN=TOP><b>Returned Value:</b></td>
<td></td>
</tr>
<tr>
<td VALIGN=TOP>&nbsp;&nbsp;&nbsp; int</td>
<td><a href="ug_interface.html#Error Codes">Error code</a>.&nbsp;</td>
</tr>
</table>
<!------------------------------------------------------------------------->
<hr WIDTH="100%"><a NAME="Zoltan_LB_Balance"></a>
<hr WIDTH="100%">
<table WIDTH="100%" NOSAVE >
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C:</td>
<td WIDTH="85%">int <b>Zoltan_LB_Balance</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; struct <b>Zoltan_Struct</b> *<i>zz</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>changes</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_gid_entries</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_lid_entries</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_import</i>,&nbsp;
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>import_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>import_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>import_procs</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int *<i>num_export</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>export_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>export_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>export_procs</i>);&nbsp;</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP WIDTH="15%" NOSAVE>FORTRAN:</td>
<td>FUNCTION <b>Zoltan_LB_Balance</b>(<i>zz, changes, num_gid_entries,
num_lid_entries, num_import, import_global_ids, import_local_ids, import_procs,
num_export, export_global_ids, export_local_ids, export_procs</i>)&nbsp;
<br>INTEGER(Zoltan_INT) :: Zoltan_LB_Balance&nbsp;
<br>TYPE(Zoltan_Struct), INTENT(IN) :: zz&nbsp;
<br>LOGICAL, INTENT(OUT) :: changes&nbsp;
<br>INTEGER(Zoltan_INT), INTENT(OUT) :: num_gid_entries, num_lid_entries
<br>INTEGER(Zoltan_INT), INTENT(OUT) :: num_import, num_export&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_global_ids, export_global_ids&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_local_ids, export_local_ids&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_procs, export_procs&nbsp;</td>
</tr>
</table>
<hr WIDTH="100%"><b>Zoltan_LB_Balance</b> is a wrapper around
<b><a href="#Zoltan_LB_Partition">Zoltan_LB_Partition</a></b>
that excludes the part assignment results. <b>Zoltan_LB_Balance</b>
assumes the number of parts is equal to the number of processors;
thus, the part assignment is equivalent to the processor assignment.
Results of the partitioning are returned in lists of objects to be imported
and exported. These arrays are allocated in Zoltan; applications should
not allocate these arrays before calling <b>Zoltan_LB_Balance</b>. The
arrays are later freed through calls to <b><a href="#Zoltan_LB_Free_Data">Zoltan_LB_Free_Data</a></b>
or
<b><a href="#Zoltan_LB_Free_Part">Zoltan_LB_Free_Part</a></b>.
<br>&nbsp;
<table WIDTH="100%" NOSAVE >
<tr>
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
<td WIDTH="80%"></td>
</tr>
<tr>
<td VALIGN=TOP></td>
<td>All arguments are analogous to those in <b><a href="#Zoltan_LB_Partition">Zoltan_LB_Partition</a></b>.
Part-assignment arguments <i>import_to_part</i> and <i>export_to_part</i>
are not included, as processor and parts numbers are considered to
be the same in <b>Zoltan_LB_Balance</b>.</td>
</tr>
<tr>
<td><b>Returned Value:</b></td>
<td></td>
</tr>
<tr>
<td VALIGN=TOP>&nbsp;&nbsp;&nbsp; int</td>
<td><a href="ug_interface.html#Error Codes">Error code</a>.</td>
</tr>
</table>
<p><!------------------------------------------------------------------------->
<hr WIDTH="100%"><a NAME="Zoltan_LB_Free_Data"></a>
<hr WIDTH="100%">
<table WIDTH="100%" NOSAVE >
<tr NOSAVE>
<td VALIGN=TOP NOSAVE>C:</td>
<td WIDTH="85%">int <b>Zoltan_LB_Free_Data</b> (
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>import_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>import_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>import_procs</i>,&nbsp;
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>export_global_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; <b><a href="ug_usage.html#Data Types for Object IDs">ZOLTAN_ID_PTR</a></b>
*<i>export_local_ids</i>,
<br>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; int **<i>export_procs</i>);&nbsp;</td>
</tr>
<tr NOSAVE>
<td VALIGN=TOP WIDTH="15%" NOSAVE>FORTRAN:</td>
<td>FUNCTION <b>Zoltan_LB_Free_Data</b>(<i>import_global_ids, import_local_ids,
import_procs, export_global_ids, export_local_ids, export_procs</i>)&nbsp;
<br>INTEGER(Zoltan_INT) :: Zoltan_LB_Free_Data&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_global_ids, export_global_ids&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_local_ids, export_local_ids&nbsp;
<br>INTEGER(Zoltan_INT), POINTER, DIMENSION(:) :: import_procs, export_procs&nbsp;</td>
</tr>
</table>
<hr WIDTH="100%"><b>Zoltan_LB_Free_Data</b> frees the memory allocated
by the Zoltan to return the results of <b><a href="#Zoltan_LB_Balance">Zoltan_LB_Balance</a></b>
or
<b><a href="ug_interface_mig.html#Zoltan_Compute_Destinations">Zoltan_Compute_Destinations</a></b>.
Memory pointed to by the arguments is freed and the arguments are set to
NULL in C or nullified in Fortran. NULL arguments may be passed to <b>Zoltan_LB_Free_Data</b>.
Note that this function does not destroy the Zoltan data structure itself;
it is deallocated through a call to <b><a href="ug_interface_init.html#Zoltan_Destroy">Zoltan_Destroy</a></b>.
<br>&nbsp;
<table WIDTH="100%" >
<tr>
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
<td WIDTH="80%"></td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; import_global_ids</i></td>
<td>The array containing the global IDs of objects imported to this processor.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; import_local_ids</i></td>
<td>The array containing the local IDs of objects imported to this processor.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; import_procs</i></td>
<td>The array containing the processor IDs of the processors that owned
the imported objects in the previous decomposition (i.e., the source processors).</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; export_global_ids&nbsp;</i></td>
<td>The array containing the global IDs of objects exported from this processor.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; export_local_ids</i></td>
<td>The array containing the local IDs of objects exported from this processor.</td>
</tr>
<tr>
<td VALIGN=TOP><i>&nbsp;&nbsp;&nbsp; export_procs</i></td>
<td>The array containing the processor IDs of processors that own the exported
objects in the new decomposition (i.e., the destination processors).</td>
</tr>
<tr>
<td VALIGN=TOP><b>Returned Value:</b></td>
<td></td>
</tr>
<tr>
<td VALIGN=TOP>&nbsp;&nbsp;&nbsp; int</td>
<td><a href="ug_interface.html#Error Codes">Error code</a>.&nbsp;</td>
</tr>
</table>
<!-------------------------------------------------------------------------><!------------------------------------------------------------------------->
<hr WIDTH="100%">[<a href="ug.html">Table of Contents</a>&nbsp; | <a href="ug_interface_augment.html">Next:&nbsp;
Functions for Augmenting a Decomposition</a>&nbsp; |&nbsp; <a href="ug_interface_init.html">Previous:&nbsp;
Initialization Functions</a>&nbsp; |&nbsp; <a href="https://www.sandia.gov/general/privacy-security/index.html">Privacy and Security</a>]
</body>
</html>