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.
1266 lines
40 KiB
1266 lines
40 KiB
2 years ago
|
<!-------- @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/4.7 [en] (X11; U; SunOS 5.6 sun4m) [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: Communication Utilities</title>
|
||
|
|
||
|
</head>
|
||
|
<body bgcolor="#FFFFFF">
|
||
|
|
||
|
<div align=right><b><i><a href="ug.html"> Zoltan User's Guide</a>
|
||
|
| <a href="ug_util_dd.html">Next</a> | <a href="ug_util_mem.html">Previous</a></i></b></div>
|
||
|
<!------------------------------------------------------------------------->
|
||
|
|
||
|
<h2>
|
||
|
Unstructured Communication Utilities</h2>
|
||
|
The unstructured communication package provides a simple interface for
|
||
|
doing complicated patterns of point-to-point communication, such as those
|
||
|
associated with data remapping. This package consists of a few simple functions
|
||
|
which create or modify communication <i>plans</i>, perform communication,
|
||
|
and destroy communication plans upon completion. The package is descended
|
||
|
from software first developed by Steve Plimpton and Bruce Hendrickson,
|
||
|
and has proved useful in a variety of different applications. For this
|
||
|
reason, it is maintained as a separate library and can be used independently
|
||
|
from Zoltan.
|
||
|
<p>In a prototypical usage of the unstructured communication package each
|
||
|
processor has some objects to send to other processors, but no processor
|
||
|
knows what messages it will receive. A call to
|
||
|
<b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>
|
||
|
produces a data structure called a communication <i>plan</i> which encapsulates
|
||
|
the basic information about the communication operation. The plan does
|
||
|
not know anything about the types of objects being transferred, only the
|
||
|
number of them. So the same plan can be used repeatedly to transfer different
|
||
|
types of data as long as the number of objects in the transfers remains
|
||
|
the same. The actual size of objects isn't specified until the call to
|
||
|
<b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>
|
||
|
which performs the data transfer.
|
||
|
<p>The plan which is produced by
|
||
|
<b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>
|
||
|
assumes that all the objects are of the same size. If this is not true,
|
||
|
then a call to <b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize</a></b> can
|
||
|
specify the actual size of each object, and the plan is augmented appropriately.
|
||
|
<b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize</a></b>
|
||
|
can be invoked repeatedly on the same plan to specify varying sizes for
|
||
|
different data transfer operations.
|
||
|
<p>Although a friendlier interface may be added in the future, for now
|
||
|
the data to be sent must be passed to
|
||
|
<b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>
|
||
|
as a packed buffer in which the objects are stored consecutively. This
|
||
|
probably requires the application to pull the data out of native data structures
|
||
|
and place in into the buffer. The destination of each object is specified
|
||
|
by the <i>proclist</i> argument to <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.
|
||
|
Some flexibility is supported by allowing <i>proclist</i> to contain negative
|
||
|
values, indicating that the corresponding objects are not to be sent. The
|
||
|
communication operations allow for any object to be sent to any destination
|
||
|
processor. However, if the objects are grouped in such a way that all those
|
||
|
being sent to a particular processor are consecutive, the time and
|
||
|
memory of an additional copy is avoided.
|
||
|
<p> Function
|
||
|
<b><a href="#Zoltan_Comm_Do_Reverse">Zoltan_Comm_Do_Reverse</a></b>
|
||
|
reverses the communication plan to send back messages to the originators.
|
||
|
<p> To allow overlap between communication and processing, POST and WAIT
|
||
|
variants of
|
||
|
<b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>
|
||
|
and
|
||
|
<b><a href="#Zoltan_Comm_Do_Reverse">Zoltan_Comm_Do_Reverse</a></b>
|
||
|
are provided. Communication is initiated by the POST function
|
||
|
(<b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do_Post</a></b> or
|
||
|
<b><a href="#Zoltan_Comm_Do_Reverse">Zoltan_Comm_Do_Reverse_Post</a></b>);
|
||
|
incoming messages are posted and outgoing messages are sent.
|
||
|
Then the user can continue processing.
|
||
|
After the processing is complete, the corresponding WAIT function
|
||
|
(<b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do_Wait</a></b> or
|
||
|
<b><a href="#Zoltan_Comm_Do_Reverse">Zoltan_Comm_Do_Reverse_Wait</a></b>)
|
||
|
is called to wait for all incoming messages to be received.
|
||
|
For convenience,
|
||
|
these functions use the same calling
|
||
|
arguments as
|
||
|
<b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b> and
|
||
|
<b><a href="#Zoltan_Comm_Do_Reverse">Zoltan_Comm_Do_Reverse</a></b>.
|
||
|
|
||
|
<p>All the functions in the unstructured communication library return integer
|
||
|
<a href="../ug_html/ug_interface.html#Error Codes">error codes</a> identical to
|
||
|
those used by Zoltan.
|
||
|
<p>
|
||
|
The C++ interface to the unstructured communication utility is found
|
||
|
in the <I>zoltan_comm_cpp.h</I> header file which defines the
|
||
|
<B>Zoltan_Comm</B> class.
|
||
|
<p>
|
||
|
A Fortran90 interface is not yet available.
|
||
|
|
||
|
|
||
|
<p>
|
||
|
<hr>
|
||
|
<table>
|
||
|
<tr VALIGN=TOP>
|
||
|
<td WIDTH="50%"><b>Source code location:</b></td>
|
||
|
|
||
|
<td WIDTH="50%"><i>Utilities/Communication</i></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr VALIGN=TOP>
|
||
|
<td><b>C Function prototypes file:</b></td>
|
||
|
|
||
|
<td><i>Utilities/Communication/zoltan_comm.h</i>
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr VALIGN=TOP>
|
||
|
<td><b>C++ class definition:</b></td>
|
||
|
|
||
|
<td><i>Utilities/Communication/zoltan_comm_cpp.h</i>
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr VALIGN=TOP>
|
||
|
<td><b>Library name:</b></td>
|
||
|
|
||
|
<td>libzoltan_comm.a</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr VALIGN=TOP>
|
||
|
<td><b>Other libraries used by this library:</b></td>
|
||
|
|
||
|
<td>libmpi.a, <a href="ug_util_mem.html">libzoltan_mem.a</a>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr VALIGN=TOP>
|
||
|
<td COLSPAN="2"><b>High Level Routines:</b>
|
||
|
<blockquote><b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a>: </b>computes
|
||
|
a communication plan for sending objects to destination processors.
|
||
|
<br><b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a>: </b>uses a communication
|
||
|
plan to send data objects to destination processors. The POST and WAIT variants
|
||
|
are
|
||
|
<br> <b><a href="#Zoltan_Comm_Do_Post">Zoltan_Comm_Do_Post</a></b> and
|
||
|
<br> <b><a href="#Zoltan_Comm_Do_Wait">Zoltan_Comm_Do_Wait</a>. </b>
|
||
|
<br><b><a href="#Zoltan_Comm_Do_Reverse">Zoltan_Comm_Do_Reverse</a>: </b>
|
||
|
performs the reverse (opposite) communication of <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>.
|
||
|
The POST and WAIT variants are
|
||
|
<br> <b><a href="#Zoltan_Comm_Do_Reverse_Post">Zoltan_Comm_Do_Reverse_Post</a></b> and
|
||
|
<br> <b><a href="#Zoltan_Comm_Do_Reverse_Wait">Zoltan_Comm_Do_Reverse_Wait</a>.</b>
|
||
|
|
||
|
<br><b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize</a>: </b>augments
|
||
|
the plan to allow objects to be of variable sizes.
|
||
|
<br><b><a href="#Zoltan_Comm_Copy">Zoltan_Comm_Copy</a>:</b> create a new
|
||
|
communication plan and copy an existing one to it.
|
||
|
<br><b><a href="#Zoltan_Comm_Copy_To">Zoltan_Comm_Copy_To</a>:</b> copy
|
||
|
one existing communication plan to another.
|
||
|
<br><b><a href="#Zoltan_Comm_Destroy">Zoltan_Comm_Destroy</a>:</b> free memory
|
||
|
associated with a communication plan. </blockquote>
|
||
|
<b>Low Level Routines:</b>
|
||
|
<blockquote><b><a href="#Zoltan_Comm_Exchange_Sizes">Zoltan_Comm_Exchange_Sizes</a>:
|
||
|
</b>updates the sizes of the messages each processor will receive.
|
||
|
<br><b><a href="#Zoltan_Comm_Invert_Map">Zoltan_Comm_Invert_Map</a>: </b>given
|
||
|
a set of messages each processor wants to send, determines the set of messages
|
||
|
each processor needs to receive.
|
||
|
<br><b><a href="#Zoltan_Comm_Sort_Ints">Zoltan_Comm_Sort_Ints</a>: </b>sorts
|
||
|
an array of integer values.
|
||
|
<br><b><a href="#Zoltan_Comm_Info">Zoltan_Comm_Info</a>: </b>returns information about a communication plan.
|
||
|
<br><b><a href="#Zoltan_Comm_Invert_Plan">Zoltan_Comm_Invert_Plan</a>: </b>given a communication plan, converts the plan into a plan for the reverse communication.
|
||
|
</blockquote>
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr VALIGN=TOP>
|
||
|
<td COLSPAN="2"><b>Use in Zoltan:</b>
|
||
|
<blockquote>The Zoltan library uses the unstructured communication package
|
||
|
in its migration tools and in some of the load-balancing algorithms. For
|
||
|
example, in <b><a href="../ug_html/ug_interface_mig.html#Zoltan_Migrate">Zoltan_Migrate</a></b>,
|
||
|
<b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>
|
||
|
is used to develop a communication map for sending objects to be exported
|
||
|
to their new destination processors. The sizes of the exported objects are
|
||
|
obtained and the communication map is augmented with a call to
|
||
|
<a href="#Zoltan_Comm_Resize"><b>Zoltan_Comm_Resize</b></a>.
|
||
|
The data for the objects is packed
|
||
|
into a communication buffer and sent to the other processors through a
|
||
|
call to <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>. After the received
|
||
|
objects are unpacked, the communication plan is no longer needed, and it
|
||
|
is deallocated by a call to
|
||
|
<b><a href="#Zoltan_Comm_Destroy">Zoltan_Comm_Destroy</a></b>.
|
||
|
Zoltan developers use the package whenever possible, as improvements made
|
||
|
to the package (for example, support for heterogeneous architectures)
|
||
|
automatically propagate to the algorithms. </blockquote>
|
||
|
</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr> <hr>
|
||
|
<b>C:</b><br>
|
||
|
<a NAME="Zoltan_Comm_Create"></a>
|
||
|
int <b>Zoltan_Comm_Create</b>(
|
||
|
struct Zoltan_Comm_Obj **<i>plan</i>,
|
||
|
int <i>nsend</i>,
|
||
|
int *<i>proclist</i>,
|
||
|
MPI_Comm <i>comm</i>,
|
||
|
int <i>tag</i>,
|
||
|
int *<i>nreturn</i>);
|
||
|
<br>
|
||
|
<b>C++:</b><br>
|
||
|
<b>Zoltan_Comm</b>(
|
||
|
const int & <i>nsend</i>,
|
||
|
int *<i>proclist</i>,
|
||
|
const MPI_Comm & <i>comm</i>,
|
||
|
const int & <i>tag</i>,
|
||
|
int *<i>nreturn</i>);
|
||
|
<br>
|
||
|
or
|
||
|
<br>
|
||
|
<b>Zoltan_Comm</b>();
|
||
|
<br>
|
||
|
<b>Zoltan_Comm::Create</b>(
|
||
|
const int & <i>nsend</i>,
|
||
|
int *<i>proclist</i>,
|
||
|
const MPI_Comm & <i>comm</i>,
|
||
|
const int & <i>tag</i>,
|
||
|
int *<i>nreturn</i>);
|
||
|
<hr>The <b>Zoltan_Comm_Create</b> function sets up the communication plan in
|
||
|
the unstructured communication package. Its input is a count of objects
|
||
|
to be sent to other processors, a list of the processors to which the objects
|
||
|
should be sent (repetitions are allowed), and an MPI communicator and tag.
|
||
|
It allocates and builds a communication plan that describes to which processors
|
||
|
data will be sent and from which processors data will be received. It also
|
||
|
computes the amount of data to be sent to and received from each processor.
|
||
|
It returns the number of objects to be received by the processor and a
|
||
|
pointer to the communication plan it created. The communication
|
||
|
plan is then used by calls to <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>
|
||
|
to perform the actual communication.
|
||
|
<br>
|
||
|
<table WIDTH="100%" NOSAVE >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> <i>plan</i></td>
|
||
|
|
||
|
<td>A pointer to the communication plan created by <b>Zoltan_Comm_Create</b>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nsend</i></td>
|
||
|
|
||
|
<td>The number of objects to be sent to other processors.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> proclist</i></td>
|
||
|
|
||
|
<td>An array of size <i>nsend</i> of destination processor numbers for
|
||
|
each of the objects to be sent.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> comm</i></td>
|
||
|
|
||
|
<td>The MPI communicator for the unstructured communication.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> <i>tag</i></td>
|
||
|
|
||
|
<td>A tag for MPI communication.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP> <i>nreturn</i></td>
|
||
|
|
||
|
<td>Upon return, the number of objects to be received by the processor.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
<p>
|
||
|
In the C++ interface to the communication utility, the communication plan
|
||
|
is represented by a <B>Zoltan_Comm</B> object. It is created when the
|
||
|
<B>Zoltan_Comm</B> constructor executes. There are two constructors. The
|
||
|
first one listed above uses parameters to initialize the plan. The
|
||
|
second constructor does not, but the plan can subsequently be initialized
|
||
|
with a call to <B>Zoltan_Comm::Create()</B>.
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Do"></a>
|
||
|
<a NAME="Zoltan_Comm_Do_Post"></a>
|
||
|
<a NAME="Zoltan_Comm_Do_Wait"></a>
|
||
|
<hr>
|
||
|
<b>C:</b><br>
|
||
|
int <b>Zoltan_Comm_Do</b>(struct Zoltan_Comm_Obj *<i>plan</i>,
|
||
|
int <i>tag</i>, char *<i>send_data</i>, int <i>nbytes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm_Do_Post</b>(struct Zoltan_Comm_Obj *<i>plan</i>,
|
||
|
int <i>tag</i>, char *<i>send_data</i>, int <i>nbytes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm_Do_Wait</b>(struct Zoltan_Comm_Obj *<i>plan</i>,
|
||
|
int <i>tag</i>, char *<i>send_data</i>, int <i>nbytes</i>, char *<i>recvbuf</i>);
|
||
|
<br><b>C++:</b><br>
|
||
|
int <b>Zoltan_Comm::Do</b>(const int & <i>tag</i>, char *<i>send_data</i>, const int & <i>nbytes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm::Do_Post</b>(const int & <i>tag</i>, char *<i>send_data</i>, const int & <i>nbytes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm::Do_Wait</b>(const int & <i>tag</i>, char *<i>send_data</i>, const int & <i>nbytes</i>, char *<i>recvbuf</i>);
|
||
|
|
||
|
<hr>The <b>Zoltan_Comm_Do</b> function performs the communication described
|
||
|
in a communication plan built by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.
|
||
|
Using the plan, it takes a buffer of object data to be sent and the size
|
||
|
(in bytes) of each object's data in that buffer and sends the data to other
|
||
|
processors.
|
||
|
<b>Zoltan_Comm_Do</b> also receives object data from
|
||
|
other processors and stores it in a receive buffer.
|
||
|
The receive buffer must be allocated by the code
|
||
|
calling <b>Zoltan_Comm_Do</b> using the number of received objects returned
|
||
|
by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b> or <b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize</a></b>.
|
||
|
If the objects have variable sizes, then <b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize</a></b>
|
||
|
must be called before <b>Zoltan_Comm_Do</b>.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> plan</i></td>
|
||
|
|
||
|
<td>A pointer to a communication plan built by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> <i>tag</i></td>
|
||
|
|
||
|
<td>An MPI message tag.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> send_data</i></td>
|
||
|
|
||
|
<td>A buffer filled with object data to be sent to other processors.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nbytes</i></td>
|
||
|
|
||
|
<td>The size (in bytes) of the data for one object, or the scale factor
|
||
|
if the objects have variable sizes. (See <b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize
|
||
|
</a></b>for
|
||
|
more details.)</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP> <i>recvbuf</i></td>
|
||
|
|
||
|
<td>Upon return, a buffer filled with object data received from other processors.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Do_Reverse"></a>
|
||
|
<a NAME="Zoltan_Comm_Do_Reverse_Post"></a>
|
||
|
<a NAME="Zoltan_Comm_Do_Reverse_Wait"></a>
|
||
|
<hr>
|
||
|
<b>C:</b><br>
|
||
|
int <b>Zoltan_Comm_Do_Reverse</b>(
|
||
|
struct Zoltan_Comm_Obj *<i>plan</i>, int <i>tag</i>, char *<i>send_data</i>, int <i>nbytes</i>,
|
||
|
int *<i>sizes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm_Do_Reverse_Post</b>(
|
||
|
struct Zoltan_Comm_Obj *<i>plan</i>, int <i>tag</i>, char *<i>send_data</i>, int <i>nbytes</i>,
|
||
|
int *<i>sizes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm_Do_Reverse_Wait</b>(
|
||
|
struct Zoltan_Comm_Obj *<i>plan</i>, int <i>tag</i>, char *<i>send_data</i>, int <i>nbytes</i>,
|
||
|
int *<i>sizes</i>, char *<i>recvbuf</i>);
|
||
|
<br><b>C++:</b><br>
|
||
|
int <b>Zoltan_Comm::Do_Reverse</b>(
|
||
|
const int & <i>tag</i>, char *<i>send_data</i>, const int & <i>nbytes</i>,
|
||
|
int *<i>sizes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm::Do_Reverse_Post</b>(
|
||
|
const int & <i>tag</i>, char *<i>send_data</i>, const int & <i>nbytes</i>,
|
||
|
int *<i>sizes</i>, char *<i>recvbuf</i>);
|
||
|
<br>
|
||
|
int <b>Zoltan_Comm::Do_Reverse_Wait</b>(
|
||
|
const int & <i>tag</i>, char *<i>send_data</i>, const int & <i>nbytes</i>,
|
||
|
int *<i>sizes</i>, char *<i>recvbuf</i>);
|
||
|
<hr>The <b>Zoltan_Comm_Do_Reverse</b> function performs communication based
|
||
|
on a communication plan built by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.
|
||
|
But unlike <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>, this routine performs
|
||
|
the reverse of the communication pattern. Specifically, all sends in the
|
||
|
plan are treated as receives and vice versa. <b>Zoltan_Comm_Do_Reverse</b>
|
||
|
is particularly well suited to return updated data objects to their originating
|
||
|
processors when the objects were initially transferred via
|
||
|
<b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> plan</i></td>
|
||
|
|
||
|
<td>A pointer to a communication plan built by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> <i>tag</i></td>
|
||
|
|
||
|
<td>An MPI message tag to be used by this routine.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> send_data</i></td>
|
||
|
|
||
|
<td>A buffer filled with object data to be sent to other processors.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nbytes</i></td>
|
||
|
|
||
|
<td>The size (in bytes) of the data associated with an object, or the scale
|
||
|
factor if the objects have variable sizes.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> sizes</i></td>
|
||
|
|
||
|
<td>If not NULL, this input array specifies the size of all the data objects
|
||
|
being transferred. This argument is passed directly to <b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize</a></b>.
|
||
|
This array has length equal to the <i>nsend</i> value passed to
|
||
|
<b>Zoltan_Comm_Create</b>.
|
||
|
But note that for <b>Zoltan_Comm_Do_Reverse</b> this array describes the sizes
|
||
|
of the values being received, not sent.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP> <i>recvbuf</i></td>
|
||
|
|
||
|
<td>Upon return, a buffer filled with object data received from other processors.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Resize"></a>
|
||
|
<hr>
|
||
|
<b>C:</b><br>
|
||
|
int <b>Zoltan_Comm_Resize</b>(
|
||
|
struct Zoltan_Comm_Obj *<i>plan</i>,
|
||
|
int *<i>sizes</i>, int <i>tag</i> , int *<i>total_recv_size</i>);
|
||
|
<br><b>C++:</b><br>
|
||
|
int <b>Zoltan_Comm::Resize</b>(
|
||
|
int *<i>sizes</i>, const int & <i>tag</i> , int *<i>total_recv_size</i>);
|
||
|
<hr>If the objects being communicated are of variable sizes, then the plan
|
||
|
produced by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b> is incomplete.
|
||
|
This routine allows the plan to be augmented to allow for variable sizes.
|
||
|
<b>Zoltan_Comm_Resize</b>
|
||
|
can be invoked repeatedly on the same plan to specify different object
|
||
|
sizes associated with different data transfers.
|
||
|
<br>
|
||
|
<table WIDTH="100%" NOSAVE >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> plan</i></td>
|
||
|
|
||
|
<td>A communication plan built by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> sizes</i></td>
|
||
|
|
||
|
<td>An input array of length equal to the <i>nsend</i> argument in the
|
||
|
call to <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b> which generated
|
||
|
the <i>plan</i>. Each entry in the array is the size of the corresponding
|
||
|
object to be sent. If <i>sizes</i> is <i>NULL </i>(on all processors),
|
||
|
the objects are considered to be the same size. Note that the true
|
||
|
size of a message will be scaled by the <i>nbytes</i> argument to <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> tag</i></td>
|
||
|
|
||
|
<td>A message tag to be used for communication within this routine, based
|
||
|
upon the communicator in <i>plan</i>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr VALIGN=TOP NOSAVE>
|
||
|
<td> <i>total_recv_size</i></td>
|
||
|
|
||
|
<td NOSAVE>Sum of the sizes of the incoming messages. To get the actual
|
||
|
size (in bytes), you need to scale by the <i>nbytes</i> argument to <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Copy"></a>
|
||
|
<hr>
|
||
|
<b>C:</b> struct Zoltan_Comm_Obj *<b>Zoltan_Comm_Copy</b>(
|
||
|
struct Zoltan_Comm_Obj *<i>plan</i>);
|
||
|
<br><b>C++:</b> <b>Zoltan_Comm(const Zoltan_Comm &plan);</b>
|
||
|
<hr><b>Zoltan_Comm_Copy</b> creates a new Zoltan_Comm_Obj structure
|
||
|
and copies the existing <I>plan</I> to it.
|
||
|
The corresponding C++ method is the <b>Zoltan_Comm</b> copy
|
||
|
constructor.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> plan</i></td>
|
||
|
|
||
|
<td>A pointer to the communication plan to be copied to the new
|
||
|
Zoltan_Comm_Obj structure. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> struct Zoltan_Comm_Obj *</td>
|
||
|
|
||
|
<td>the newly created plan, or NULL on error.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Copy_To"></a>
|
||
|
<hr>
|
||
|
<b>C:</b> int <b>Zoltan_Comm_Copy_To</b>(
|
||
|
struct Zoltan_Comm_Obj **<i>to</i>,
|
||
|
struct Zoltan_Comm_Obj *<i>from</i>);
|
||
|
<br><b>C++:</b> Zoltan_Comm & <b>operator= </b>(const Zoltan_Comm &plan);</b>
|
||
|
<hr><b>Zoltan_Comm_Copy_To</b> copies one existing communication plan
|
||
|
to another.
|
||
|
The corresponding C++ method is the <b>Zoltan_Comm</b> copy operator.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> to</i></td>
|
||
|
<td>A pointer to a pointer to the communication plan that will
|
||
|
be copied to. We destroy the plan first, and set the pointer to
|
||
|
the plan to NULL, before proceeding with the copy.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> from</i></td>
|
||
|
<td>A pointer the communication plan that we will make a copy of.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Destroy"></a>
|
||
|
<hr>
|
||
|
<b>C:</b> int <b>Zoltan_Comm_Destroy</b>(
|
||
|
struct Zoltan_Comm_Obj **<i>plan</i>);
|
||
|
<br><b>C++:</b> <b>~Zoltan_Comm();</b>
|
||
|
<hr>The <b>Zoltan_Comm_Destroy</b> function frees all memory associated with
|
||
|
a communication plan created by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.
|
||
|
The C++ <b>Zoltan_Comm</b> object does not have an explicit <b>Destroy</b> method.
|
||
|
It is deallocated when its destructor is called.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> plan</i></td>
|
||
|
|
||
|
<td>A pointer to a communication plan built by <b><a href="#Zoltan_Comm_Create">Zoltan_Comm_Create</a></b>.
|
||
|
Upon return, <i>plan</i> is set to NULL. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Exchange_Sizes"></a>
|
||
|
<hr>
|
||
|
<b>C:</b><br>
|
||
|
int <b>Zoltan_Comm_Exchange_Sizes</b>(
|
||
|
int *<i>sizes_to</i>, int *<i>procs_to</i>, int <i>nsends</i>, int <i>self_msg</i>,
|
||
|
int *<i>sizes_from</i>, int *<i>procs_from</i>, int <i>nrecvs</i>, int
|
||
|
*<i>total_recv_size</i>, int <i>my_proc</i>, int <i>tag</i>, MPI_Comm <i>comm</i>
|
||
|
);
|
||
|
<br><b>C++:</b><br>
|
||
|
static int <b>Zoltan_Comm::Exchange_Sizes</b>(
|
||
|
int *<i>sizes_to</i>, int *<i>procs_to</i>,
|
||
|
const int & <i>nsends</i>, const int & <i>self_msg</i>,
|
||
|
int *<i>sizes_from</i>, int *<i>procs_from</i>, const int & <i>nrecvs</i>,
|
||
|
int *<i>total_recv_size</i>, const int & <i>my_proc</i>,
|
||
|
const int & <i>tag</i>, const MPI_Comm & <i>comm</i>
|
||
|
);
|
||
|
<hr>This routine is used by <b><a href="#Zoltan_Comm_Resize">Zoltan_Comm_Resize</a></b>
|
||
|
to update the sizes of the messages each processor is expecting to receive.
|
||
|
The processors already know who will send them messages, but if variable
|
||
|
sized objects are being communicated, then the sizes of the messages are
|
||
|
recomputed and exchanged via this routine.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> sizes_to</i></td>
|
||
|
|
||
|
<td>Input array with the size of each message to be sent. Note
|
||
|
that the actual number of bytes in the message is the product of this value
|
||
|
and the <i>nbytes</i> argument to <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> procs_to</i></td>
|
||
|
|
||
|
<td>Input array with the destination processor for each of the messages
|
||
|
to be sent.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nsends</i></td>
|
||
|
|
||
|
<td>Input argument with the number of messages to be sent. (Length
|
||
|
of the <i>procs_to</i> array.)</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> self_msg</i></td>
|
||
|
|
||
|
<td>Input argument indicating whether a processor has data for itself (=1) or not
|
||
|
(=0) within the <i>procs_to</i> and <i>lengths_to</i> arrays.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> sizes_from</i></td>
|
||
|
|
||
|
<td>Returned array with the size of each message that will be received. Note
|
||
|
that the actual number of bytes in the message is the product of this value
|
||
|
and the <i>nbytes</i> argument to <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> procs_from</i></td>
|
||
|
|
||
|
<td>Returned array of processors from which data will be received.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nrecvs</i></td>
|
||
|
|
||
|
<td>Returned value with number of messages to be received. (length of <i>procs_from</i>
|
||
|
array.)</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> total_recv_size</i></td>
|
||
|
|
||
|
<td>The total size of all the messages to be received. As above, the actual
|
||
|
number of bytes will be scaled by the <i>nbytes</i> argument to <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b>.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> my_proc</i></td>
|
||
|
|
||
|
<td>The processor's ID in the <i>comm</i> communicator. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> tag</i></td>
|
||
|
|
||
|
<td>A message tag which can be used by this routine.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> comm</i></td>
|
||
|
|
||
|
<td>MPI Communicator for the processor numbering in the <i>procs</i> arrays.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Invert_Map"></a>
|
||
|
<hr>
|
||
|
<b>C:</b><br>
|
||
|
int <b>Zoltan_Comm_Invert_Map</b>( int
|
||
|
*<i>lengths_to</i>, int *<i> procs_to</i>, int <i>nsends</i>, int <i>self_msg</i>,
|
||
|
int **<i> lengths_from</i>, int **<i> procs_from</i>, int *<i> nrecvs</i>,
|
||
|
int <i>my_proc</i>, int <i>nprocs</i>, int <i>out_of_mem</i>, int <i>tag</i>,
|
||
|
MPI_Comm <i>comm</i> );
|
||
|
<br><b>C++:</b><br>
|
||
|
static int <b>Zoltan_Comm::Invert_Map</b>( int
|
||
|
*<i>lengths_to</i>, int *<i> procs_to</i>,
|
||
|
const int & <i>nsends</i>, const int & <i>self_msg</i>,
|
||
|
int * &<i> lengths_from</i>, int * &<i> procs_from</i>, int &<i> nrecvs</i>,
|
||
|
const int & <i>my_proc</i>, const int & <i>nprocs</i>,
|
||
|
const int & <i>out_of_mem</i>, const int & <i>tag</i>,
|
||
|
const MPI_Comm & <i>comm</i> );
|
||
|
<hr>The <b>Zoltan_Comm_Invert_Map</b> function is a low level communication
|
||
|
routine. It is useful when a processor knows to whom it needs to send data,
|
||
|
but not from whom it needs to receive data. Each processor provides to this
|
||
|
routine a set of lengths and destinations for the messages it wants to
|
||
|
send. The routine then returns the set of lengths and origins for the messages
|
||
|
a processor will receive. Note that by inverting the interpretation
|
||
|
of
|
||
|
<i>to</i> and <i>from</i> in these arguments, the routine can be used
|
||
|
to do the opposite: knowing how much data to receive and from which processors, it can
|
||
|
compute how much data to send and to which processors.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> lengths_to</i></td>
|
||
|
|
||
|
<td>Input array with the number of values in each of the messages to be sent.
|
||
|
Note that the actual size of each value is not specified until
|
||
|
the <b><a href="#Zoltan_Comm_Do">Zoltan_Comm_Do</a></b> routine is invoked.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> procs_to</i></td>
|
||
|
|
||
|
<td>Input array with the destination processor for each of the messages
|
||
|
to be sent.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nsends</i></td>
|
||
|
|
||
|
<td>Input argument with the number of messages to be sent. (Length
|
||
|
of the <i>lengths_to</i> and <i>procs_to</i> arrays.)</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> self_msg</i></td>
|
||
|
|
||
|
<td>Input argument indicating whether a processor has data for itself (=1) or not
|
||
|
(=0) within the <i>procs_to</i> and <i>lengths_to</i> arrays.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> lengths_from</i></td>
|
||
|
|
||
|
<td>Returned array with lengths of messages to be received.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> procs_from</i></td>
|
||
|
|
||
|
<td>Returned array of processors from which data will be received.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nrecvs</i></td>
|
||
|
|
||
|
<td>Returned value with number of messages to be received (lengths of
|
||
|
<i>lengths_from</i>
|
||
|
and <i>procs_from</i> arrays).</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> my_proc</i></td>
|
||
|
|
||
|
<td>The processor's ID in the <i>comm</i> communicator. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nprocs</i></td>
|
||
|
|
||
|
<td>Number of processors in the <i>comm</i> communicator. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> out_of_mem</i></td>
|
||
|
|
||
|
<td>Since it has a barrier operation, this routine is a convenient time
|
||
|
to tell all the processors that one of them is out of memory. This input
|
||
|
argument is 0 if the processor is OK, and 1 if the processor has failed
|
||
|
in a malloc call. All the processors will return with a code of <b>COMM_MEMERR</b>
|
||
|
if any of them is out of memory.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> tag</i></td>
|
||
|
|
||
|
<td>A message tag which can be used by this routine.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> comm</i></td>
|
||
|
|
||
|
<td>MPI Communicator for the processor numbering in the <i>procs</i> arrays.</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<p>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Sort_Ints"></a>
|
||
|
<hr>
|
||
|
int <b>Zoltan_Comm_Sort_Ints</b>( int *<i>vals_sort</i>,
|
||
|
int *<i>vals_other</i>, int <i>nvals</i>);
|
||
|
<hr>As its name suggests, the <b>Zoltan_Comm_Sort_Ints</b> function sorts a
|
||
|
set of integers via the quicksort algorithm. The integers are reordered
|
||
|
from lowest to highest, and a second array of integers is reordered in
|
||
|
the same fashion. This second array can be used to return the permutation
|
||
|
associated with the sort operation. There is no C++ interface to this
|
||
|
function. You can use the C function instead.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> vals_sort</i></td>
|
||
|
|
||
|
<td>The array of integers to be sorted. This array is permuted into sorted
|
||
|
order. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> vals_other</i></td>
|
||
|
|
||
|
<td>Another array of integers which is permuted identically to <i>vals_sort</i>. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nvals</i></td>
|
||
|
|
||
|
<td>The number of values in the two integer arrays. </td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Info"></a>
|
||
|
<hr>
|
||
|
<b>C:</b><br>
|
||
|
int <b>Zoltan_Comm_Info</b>(
|
||
|
struct Zoltan_Comm_Obj *<i>plan</i>,
|
||
|
int *<i>nsends</i>,
|
||
|
int *<i>send_procs</i>,
|
||
|
int *<i>send_lengths</i>,
|
||
|
int *<i>send_nvals</i>,
|
||
|
int *<i>send_max_size</i>,
|
||
|
int *<i>send_list</i>,
|
||
|
int *<i>nrecvs</i>,
|
||
|
int *<i>recv_procs</i>,
|
||
|
int *<i>recv_lengths</i>,
|
||
|
int *<i>recv_nvals</i>,
|
||
|
int *<i>recv_total_size</i>,
|
||
|
int *<i>recv_list</i>,
|
||
|
int *<i>self_msg</i>);
|
||
|
<br><b>C++:</b><br>
|
||
|
int <b>Zoltan_Comm::Info</b>(
|
||
|
int *<i>nsends</i>,
|
||
|
int *<i>send_procs</i>,
|
||
|
int *<i>send_lengths</i>,
|
||
|
int *<i>send_nvals</i>,
|
||
|
int *<i>send_max_size</i>,
|
||
|
int *<i>send_list</i>,
|
||
|
int *<i>nrecvs</i>,
|
||
|
int *<i>recv_procs</i>,
|
||
|
int *<i>recv_lengths</i>,
|
||
|
int *<i>recv_nvals</i>,
|
||
|
int *<i>recv_total_size</i>,
|
||
|
int *<i>recv_list</i>,
|
||
|
int *<i>self_msg</i>) const;
|
||
|
<hr><b>Zoltan_Comm_Info</b> returns information about a communication plan.
|
||
|
All arguments, except the <i>plan</i> itself, may be NULL; values are
|
||
|
returned only for non-NULL arguments.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i>
|
||
|
plan</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Communication data structure created by
|
||
|
<a href="#Zoltan_Comm_Create"><b>Zoltan_Comm_Create</b></a>.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nsends</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the number of processors to which messages are sent;
|
||
|
does not include self-messages.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> send_procs</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, a list of processors to which messages are sent;
|
||
|
self-messages are included.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> send_lengths</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the number of values to be sent to each processor
|
||
|
in <i>send_procs</i>.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> send_nvals</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the total number of values to send.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> send_max_size</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the maximum size of a message to be sent; does not
|
||
|
include self-messages.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> send_list</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the processor assignment of each value to be sent.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> nrecvs</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the number of processors from which to receive messages;
|
||
|
does not include self-messages.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> recv_procs</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, a list of processors from which messages are
|
||
|
received; includes self-messages.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> recv_lengths</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the number of values to be received from each
|
||
|
processor in <i>recv_procs</i>.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> recv_nvals</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the total number of values to receive.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> recv_total_size</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the total size of items to be received.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> recv_list</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the processor assignments of each value to be received.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i> self_msg</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Upon return, the number of self-messages.
|
||
|
</td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<a NAME="Zoltan_Comm_Invert_Plan"></a>
|
||
|
<hr>
|
||
|
<b>C:</b> int <b>Zoltan_Comm_Invert_Plan</b>(
|
||
|
struct Zoltan_Comm_Obj **<i>plan</i>
|
||
|
);
|
||
|
<br><b>C++:</b> int <b>Zoltan_Comm::Invert_Plan</b>();
|
||
|
|
||
|
<hr>Given a communication plan, <b>Zoltan_Comm_Invert_Plan</b>
|
||
|
alters the plan to make it the plan for the reverse communication.
|
||
|
Information in the input plan is replaced by information for the
|
||
|
reverse-communication plan. All receives in the reverse-communication
|
||
|
plan are blocked; thus, using the inverted plan does not produce the
|
||
|
same results as
|
||
|
<a href="#Zoltan_Comm_Do_Reverse"><b>Zoltan_Comm_Do_Reverse</b></a>.
|
||
|
If an error occurs within <b>Zoltan_Comm_Invert_Plan</b>, the original plan
|
||
|
is returned unaltered.
|
||
|
<br>
|
||
|
<table WIDTH="100%" >
|
||
|
<tr VALIGN=TOP>
|
||
|
<td VALIGN=TOP WIDTH="20%"><b>Arguments:</b></td>
|
||
|
|
||
|
<td WIDTH="80%"></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td VALIGN=TOP><i>
|
||
|
plan</i></td>
|
||
|
<td VALIGN=TOP>
|
||
|
Communication data structure created by
|
||
|
<a href="#Zoltan_Comm_Create"><b>Zoltan_Comm_Create</b></a>; the contents
|
||
|
of this plan are irretrievably modified by <b>Zoltan_Comm_Invert_Plan</b>.
|
||
|
</td>
|
||
|
</tr>
|
||
|
<tr>
|
||
|
<td><b>Returned Value:</b></td>
|
||
|
|
||
|
<td></td>
|
||
|
</tr>
|
||
|
|
||
|
<tr>
|
||
|
<td> int</td>
|
||
|
|
||
|
<td>Error code.</td>
|
||
|
</tr>
|
||
|
</table>
|
||
|
|
||
|
<!------------------------------------------------------------------------->
|
||
|
<hr>
|
||
|
<br>[<a href="ug.html">Table of Contents</a> | <a href="ug_util_dd.html">Next:
|
||
|
Distributed Directory Utility</a> | <a href="ug_util_mem.html">Previous:
|
||
|
Memory Management Utilities</a> | <a href="https://www.sandia.gov/general/privacy-security/index.html">Privacy and Security</a>]
|
||
|
</body>
|
||
|
</html>
|