⭐ 欢迎来到虫虫下载站！ | 📦 资源下载 📁 资源专辑 ℹ️ 关于我们
⭐ 虫虫下载站
📤 上传资源

📄 node76.html

📁 Design and building parallel program
💻 HTML
字号:
🔍 
<html><!DOCTYPE HTML PUBLIC "-//W3O//DTD W3 HTML 2.0//EN">
<!Converted with LaTeX2HTML 95.1 (Fri Jan 20 1995) by Nikos Drakos (nikos@cbl.leeds.ac.uk), CBLU, University of Leeds >
<HEAD>
<TITLE>6.9 Modularity</TITLE>
</HEAD>
<BODY>
<meta name="description" value="6.9 Modularity">
<meta name="keywords" value="book">
<meta name="resource-type" value="document">
<meta name="distribution" value="global">
<P>
 <BR> <HR><a href="msgs0.htm#2" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/tppmsgs/msgs0.htm#2"><img ALIGN=MIDDLE src="asm_color_tiny.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/asm_color_tiny.gif" alt="[DBPP]"></a>    <A NAME=tex2html2859 HREF="node75.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node75.html"><IMG ALIGN=MIDDLE ALT="previous" SRC="previous_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/previous_motif.gif"></A> <A NAME=tex2html2867 HREF="node77.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node77.html"><IMG ALIGN=MIDDLE ALT="next" SRC="next_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/next_motif.gif"></A> <A NAME=tex2html2865 HREF="node67.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node67.html"><IMG ALIGN=MIDDLE ALT="up" SRC="up_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/up_motif.gif"></A> <A NAME=tex2html2869 HREF="node1.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node1.html"><IMG ALIGN=MIDDLE ALT="contents" SRC="contents_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/contents_motif.gif"></A> <A NAME=tex2html2870 HREF="node133.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node133.html"><IMG ALIGN=MIDDLE ALT="index" SRC="index_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/index_motif.gif"></A> <a href="msgs0.htm#3" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/tppmsgs/msgs0.htm#3"><img ALIGN=MIDDLE src="search_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/search_motif.gif" alt="[Search]"></a>   <BR>
<B> Next:</B> <A NAME=tex2html2868 HREF="node77.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node77.html">6.10 Performance Issues</A>
<B>Up:</B> <A NAME=tex2html2866 HREF="node67.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node67.html">6 Fortran M</A>
<B> Previous:</B> <A NAME=tex2html2860 HREF="node75.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node75.html">6.8 Mapping</A>
<BR><HR><P>
<H1><A NAME=SECTION03390000000000000000>6.9 Modularity</A></H1>
<P>
<A NAME=secfmmod>&#160;</A>
<P>
Example <A HREF="node75.html#exccmx" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node75.html#exccmx">6.6</A> showed how FM constructs can be used to
<A NAME=9581>&#160;</A>
implement parallel and concurrent composition.  The basic ideas are
straightforward. A process (task) is defined for each component of a
multicomponent program. These processes are then composed using a
process block, with channels and/or mergers connecting ports in the
different components.  Mapping constructs control resource allocation
and hence determine whether program components execute on the same or
different processors.
<P>
In this section, we discuss the techniques used to implement <em>
<A NAME=9582>&#160;</A>
sequential
 </em> composition in FM programs.  Recall that in
<A NAME=9583>&#160;</A>
sequential composition, different program components execute in
sequence on all processors (Section <A HREF="node41.html#secseqcomp" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node41.html#secseqcomp">4.2.2</A>).  These
program components may themselves communicate and synchronize, but
they cannot create new tasks.  Hence, each process executes the same
program, and the entire computation moves sequentially from one
parallel operation to the next.  This is the SPMD programming model
discussed in Section <A HREF="node9.html#sec1other" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node9.html#sec1other">1.3.2</A>.
<P>
An FM implementation of an SPMD program comprises two components: initialization and execution.  The
<em> initialization
 </em> component creates the SPMD process structure
<A NAME=9587>&#160;</A>
and the communication structures (channels and mergers) required by
<A NAME=9588>&#160;</A>
the program components called in the sequential composition.  The
initialization typically comprises some channel and merger creation
code followed by a process do-loop that creates one process per
processor, with the appropriate ports passed as arguments.
<P>
The <em> execution
 </em> component implements the program executed by
each process created by the process do-loop.  This program comprises a
sequence of calls to the program components involved in the sequential
composition.  These components can perform internal communication
using the communication structure established during the setup phase
of the computation.
<P>
The important point to understand about this structure is that once
appropriate communication structures have been defined in the
initialization phase, sequential composition can be used directly in
the main body of the program, without any special constructs.
<P>
<BR><HR>
<b> Example <IMG BORDER=0 ALIGN=BOTTOM ALT="" SRC="img925.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/img925.gif">.<IMG BORDER=0 ALIGN=BOTTOM ALT="" SRC="img924.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/img924.gif">    Sequential Composition</b>:<A NAME=exfmseq>&#160;</A>
<P>
<A NAME=9592>&#160;</A>
We consider the SPMD finite difference computation used to illustrate
<A NAME=9593>&#160;</A>
sequential composition in Section <A HREF="node41.html#secseqcomp" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node41.html#secseqcomp">4.2.2</A>.  This
computation is structured as a sequence of calls to a finite
difference routine that performs nearest-neighbor communication and a
reduction routine used to detect termination; the latter routine
performs global communication.
<A NAME=9595>&#160;</A>
The basic logic executed by each process in this computation is as
<A NAME=9596>&#160;</A>
follows, where <tt> localgrid</tt> is the local component of the grid data
structure.
<P>

<PRE><TT> 
		<tt> subroutine node</tt>
<P>
		<tt> real localgrid(N,N/P)</tt>
<P>
		<tt> while(.not. done) do</tt>
<P>
				<tt> call finite_difference(localgrid, localmax)</tt>
<P>
				<tt> call global_maximum(localmax, globmax)</tt>
<P>
				<tt> if(globmax .lt. threshold) done = .true.</tt>
<P>
		<tt> enddo</tt>
<P>
		<tt> end</tt>
<P>
</TT></PRE>

<P>
Programs <A HREF="node76.html#progfmfdi" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node76.html#progfmfdi">6.10</A> and <A HREF="node76.html#progfmfdx" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node76.html#progfmfdx">6.11</A> implement the
initialization and execution components of an FM implementation of
this program.  Because the program components invoked in the
sequential composition perform both nearest-neighbor and hypercube
communication, the initialization code first creates channels
implementing these two communication structures.  (The function <tt>
xor</tt> is used as in Algorithm <A HREF="node124.html#algbutalg" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node124.html#algbutalg">11.1</A>.)  Then, a process
do-loop creates <tt> P</tt> instances of the <tt> node</tt> process, passing
each instance the appropriate ports.
<P>
The execution code comprises the process <tt> node</tt>, which calls first
the procedure <tt> finite_difference</tt> and then the procedure <tt>
global_maximum</tt>.  These procedures use the ports passed as arguments
to perform internal communication.
<P>
The various port variables passed as arguments in the execution
component make for a rather clumsy program.  The program can be made
tidier by storing the ports in <tt> PROCESS COMMON</tt> with a series of
<tt> MOVEPORT</tt> calls prior to calling user code.  Each process can
then execute the simpler code presented at the beginning of this
example.
<P>
<BR><HR>
<P>
<P><A NAME=progfmfdi>&#160;</A><IMG BORDER=0 ALIGN=BOTTOM ALT="" SRC="img926.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/img926.gif"><P>
<P>
<P><A NAME=progfmfdx>&#160;</A><IMG BORDER=0 ALIGN=BOTTOM ALT="" SRC="img927.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/img927.gif"><P>
<P>
<BR><HR>
<b> Example <IMG BORDER=0 ALIGN=BOTTOM ALT="" SRC="img929.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/img929.gif">.<IMG BORDER=0 ALIGN=BOTTOM ALT="" SRC="img928.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/img928.gif">    Message-passing Library</b>:<A NAME=exfmmp>&#160;</A>
<P>
<A NAME=9686>&#160;</A>
The initialization code in Program <A HREF="node76.html#progfmfdi" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node76.html#progfmfdi">6.10</A> defined a
separate communication structure for each component in the sequential
composition.  In this example, we present an alternative approach in
which the initialization component of an SPMD program creates a
single, general-purpose communication structure that can be used to
perform any communication.  In effect, we implement a message-passing
library that provides similar functionality to the MPI library described
in Chapter <A HREF="node94.html#chapmpi" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node94.html#chapmpi">8</A>.  We also use this example to show how
port variables can be stored in <tt> PROCESS COMMON</tt> to simplify program
interfaces.
<P>
Program <A HREF="node76.html#progfmmp" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node76.html#progfmmp">6.12</A> is a skeleton message-passing library that
implements just simple send and receive functions.  The main program
first creates <tt> P</tt> mergers to implement a fully connected network,
<A NAME=9692>&#160;</A>
and then creates <tt> P</tt> <tt> node</tt> processes, passing each a single
inport and <tt> P</tt> outports.  The <tt> node</tt> process stashes the
inport and outports in <tt> PROCESS COMMON</tt> and then calls the subroutine
<tt> compute</tt>, which performs the actual computation.  The <tt> PROCESS
COMMON</tt> is defined as follows (file <tt> mp_stashed.com</tt>).
<P>

<PRE><TT> 
		<tt> C</tt> 		 File <tt> &quot;mp_stashed.com&quot;</tt>
<P>
        				 <tt> PROCESS COMMON /stashed/ stashed_inp, stashed_outp</tt>
<P>
        				 <tt> INPORT  (integer n, character x(n)) stashed_inp</tt>
<P>
        				 <tt> OUTPORT (integer n, character x(n)) stashed_outp(P)</tt>
<P>
</TT></PRE>

<P>
<A NAME=9708>&#160;</A>
This initialization code makes it possible for the rest of the program
to call <tt> mp_send</tt> and <tt> mp_receive</tt> routines to send and
receive messages to and from other processes.
<P>
<BR><HR>
<P>
<P><A NAME=progfmmp>&#160;</A><IMG BORDER=0 ALIGN=BOTTOM ALT="" SRC="img930.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/img930.gif"><P>
<P>

<BR> <HR><a href="msgs0.htm#2" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/tppmsgs/msgs0.htm#2"><img ALIGN=MIDDLE src="asm_color_tiny.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/asm_color_tiny.gif" alt="[DBPP]"></a>    <A NAME=tex2html2859 HREF="node75.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node75.html"><IMG ALIGN=MIDDLE ALT="previous" SRC="previous_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/previous_motif.gif"></A> <A NAME=tex2html2867 HREF="node77.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node77.html"><IMG ALIGN=MIDDLE ALT="next" SRC="next_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/next_motif.gif"></A> <A NAME=tex2html2865 HREF="node67.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node67.html"><IMG ALIGN=MIDDLE ALT="up" SRC="up_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/up_motif.gif"></A> <A NAME=tex2html2869 HREF="node1.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node1.html"><IMG ALIGN=MIDDLE ALT="contents" SRC="contents_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/contents_motif.gif"></A> <A NAME=tex2html2870 HREF="node133.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node133.html"><IMG ALIGN=MIDDLE ALT="index" SRC="index_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/index_motif.gif"></A> <a href="msgs0.htm#3" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/tppmsgs/msgs0.htm#3"><img ALIGN=MIDDLE src="search_motif.gif" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/search_motif.gif" alt="[Search]"></a>   <BR>
<B> Next:</B> <A NAME=tex2html2868 HREF="node77.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node77.html">6.10 Performance Issues</A>
<B>Up:</B> <A NAME=tex2html2866 HREF="node67.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node67.html">6 Fortran M</A>
<B> Previous:</B> <A NAME=tex2html2860 HREF="node75.html" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/node75.html">6.8 Mapping</A>
<BR><HR><P>
<P><ADDRESS>
<I>&#169 Copyright 1995 by <A href="msgs0.htm#6" tppabs="http://www.dit.hcmut.edu.vn/books/system/par_anl/tppmsgs/msgs0.htm#6">Ian Foster</a></I>
</ADDRESS>
</BODY>

⌨️ 快捷键说明

复制代码 Ctrl + C
搜索代码 Ctrl + F
全屏模式 F11
切换主题 Ctrl + Shift + D
显示快捷键 ?
增大字号 Ctrl + =
减小字号 Ctrl + -