<concept id="GUID-C7B748CF-D080-5448-98AF-1634D821E713" xml:lang="en"><title> SMP Example</title><shortdesc>This example demonstrates symmetric multiprocessing using

multiple threads. </shortdesc><prolog><metadata><keywords/></metadata></prolog><conbody>

<p>This example shows the differences between code running in a unicore

environment and in a multiprocessor environment.</p>

<section id="GUID-7AC2BE12-4F01-4140-8379-672475DE7681"><title>Download</title> <p>Click on the following link to download the example: <xref href="guid-6013a680-57f9-415b-8851-c4fa63356636/zips/guid-577596d8-76a2-4a41-aa0f-deefe5638e23.zip" scope="external">SmpExample.zip</xref></p><p>Click: <xref href="guid-6013a680-57f9-415b-8851-c4fa63356636/guid-577596d8-76a2-4a41-aa0f-deefe5638e23.html" scope="peer">browse</xref> to view the example code. </p> </section>

<section id="GUID-7A70847E-40A2-4267-9BD4-36F88F48C84A"><title>Class

summary</title><ul>

<section id="GUID-BE13F08D-09FD-41DF-9210-D46EBECADA69"><title>Design

and implementation</title> <p>The example creates two databases named

DBforSMP1.db and DBforSMP2.db and three threads: </p> <ul>

<li id="GUID-F4EE1C84-FE30-5C99-872B-1881D852CAEA"><p> <codeph>WriterThread1</codeph>: It creates the DBforSMP1.db database and writes integers to it.

The thread has the lowest priority amongst all the threads. </p> </li>

<li id="GUID-7B2D0419-E069-59FA-A6CE-70EBC203F656"><p> <codeph>WriterThread2</codeph>: It creates the DBforSMP2.db database and writes some integers to

it. The thread has highest priority amongst all the threads. </p> </li>

<li id="GUID-7C9584F2-E6C8-5A02-B4C2-4845CF4BAA36"><p> <codeph>ReaderThread</codeph>: It reads the two databases and prints the output to the console.

It has normal priority. </p> </li>

</ul> <p>In a unicore enviroment, the threads are executed according

to their priorities. This means first <codeph> WriterThread2</codeph>, then <codeph>ReaderThread</codeph> and finally <codeph>WriterThread1</codeph> is executed. Hence <codeph>ReaderThead</codeph> can only print the

contents of the DBforSMP2.db in the console because DBforSMP1 is empty.

in an SMP environment, threads run simultaneously and <codeph>ReaderThread</codeph> can print the contents of both the databases. </p> </section>

<section id="GUID-3686F166-C4DD-4269-9B9C-F0027F53547E"><title>Building

and configuring</title> <p>To build the example: </p> <ul>

<li id="GUID-037442C7-2436-5152-9A3E-9DC0B4283A9C-GENID-1-12-1-18-1-1-9-1-4-1-25-1-30-1-3-5-3-1"><p>The example

builds an executable called <filepath>smpexample.exe</filepath> in

the standard location. </p> </li>

<li id="GUID-B7C849B9-0885-5253-A78B-71854F86D772-GENID-1-12-1-18-1-1-9-1-4-1-25-1-30-1-3-5-3-2"><p>You can build

the example from your IDE or the command line. </p> <p>If you use

an IDE, import the <filepath>bld.inf</filepath> file of the example

into your IDE, and use the build command of the IDE. </p> <p>If you

use the command line, open a command prompt, and set the current directory

to the source code directory of the example. You can then build the

example with the SBSv1 build tools with the following commands: </p> <p><userinput>bldmake bldfiles</userinput> </p> <p><userinput>abld

build</userinput> </p> </li>

<section id="GUID-F773B722-7612-4BAF-99CB-84E39CE10C40"><title>Running

the example</title> <p>To test the example in an SMP environment,

add <codeph>crazyscheduling on</codeph>in the <filepath>epoc32\data\epoc.ini</filepath> file. This emulates SMP behaviour in the techview emulator. Execute

the example in both unicore and multiprocessor environment and see

the difference. Run the example following the instructions provided

in the console. </p> </section>