plugins/org.eclipse.mat.ui.help/tasks/configure_mat.dita - mat/org.eclipse.mat - Git at Google

 <?xml version="1.0" encoding="UTF-8"?>
 <!--
     Copyright (c) 2008, 2019 SAP AG and IBM Corporation.
     All rights reserved. This program and the accompanying materials
     are made available under the terms of the Eclipse Public License v1.0
     which accompanies this distribution, and is available at
     http://www.eclipse.org/legal/epl-v10.html

     Contributors:
         SAP AG - initial API and implementation
         IBM Corporation - preference page and OQL
  -->
 <!DOCTYPE task PUBLIC "-//OASIS//DTD DITA Task//EN" "task.dtd" >
 <task id="task_configure_mat" xml:lang="en-us">
 	<title>Memory Analyzer Configuration</title>
 	<prolog>
 		<copyright>
 			<copyryear year=""></copyryear>
 			<copyrholder>
 				Copyright (c) 2008, 2019 SAP AG and others.
 			    All rights reserved. This program and the accompanying materials
 			    are made available under the terms of the Eclipse Public License v1.0
 			    which accompanies this distribution, and is available at
 			    http://www.eclipse.org/legal/epl-v10.html
 			</copyrholder>
 		</copyright>
 	</prolog>

 	<taskbody>
 		<context>
 			<p>
 				Well, analyzing big heap dumps can also require more heap space.
 				Give it
 				some more memory (possible by running on a 64-bit machine):
 			</p>

 			<codeph>MemoryAnalyzer.exe -vmargs -Xmx4g</codeph>

 			<p>Alternatively, edit the MemoryAnalyzer.ini to
 				contain:</p>

 			<p>
 				<codeblock>-vmargs
 -Xmx4g</codeblock>
 			</p>

 			<p>
 				For more details, check out the section
 				<xref format="html"
 					href="http://help.eclipse.org/topic/org.eclipse.platform.doc.user/tasks/running_eclipse.htm">Running Eclipse</xref>
 				in the Help Center. It also contains more details if you are running
 				on Mac OS X.
 			</p>

 			<p>If you are running the Memory Analyzer inside your Eclipse SDK,
 				you need to edit the eclipse.ini file.</p>

 			<p>The memory intensive parts are the parsing of the dump and building
 				of the dominator tree. Try parsing the heap
 				dump from the command line, perhaps on a bigger machine. The dumps
 				and index files can then be copied to a more convenient machine.
 				Once the dump has been parsed, it usually can
 				be opened with less memory in the GUI.
 				As a rough estimate if the number of objects is N and the number of
 				classes C, it might take at least T bytes to parse and build the dominator tree
 				where:
 				<lines>T &#8776; N * 28.25 + C * 1000 + P</lines>
 				P is the space used by the DTFJ or HPROF parsers. For a PHD file, the space could be:
 				<lines>P &#8776; C * 1000</lines>
 				Memory Analyzer uses additional memory for caching index files, so
 				performance will be better if there is more memory available than the minimum
 				required to parse a dump.
 				</p>
 				<p>Memory Analyzer has an architectural limit of 2<sup>31</sup> - 3 objects,
 				a current limit of 2<sup>31</sup> - 8 = 2,147,483,640 objects, but has not been
 				tested with that many objects. The current record is a heap dump file of 70Gbytes
 				containing 1,160,000,000 objects, which was opened with Memory Analyzer running
 				with a 64Gbyte heap.</p>

 			<p>The preference page is opened via a menu option.
 				<menucascade>
 					<uicontrol>Window</uicontrol>
 					<uicontrol>Preferences</uicontrol>
 				</menucascade>
 			</p>
 			<p id="preferences">The preferences page allows various aspects of Memory Analyzer to be controlled.
 			<image href="../mimes/preferences.png">
 				<alt>preferences page showing general options with DTFJ subpage</alt>
 			</image>
 			<dl>
 			<dlentry>
 				<dt>Keep unreachable objects</dt>
 				<dd>Objects that appear unreachable from GC roots are not discarded in an early stage of
 				Memory Analyzer processing, but are retained for further analysis.</dd>
 			</dlentry>
 			<dlentry>
 				<dt>Hide the getting started wizard</dt>
 				<dd>Controls whether to display a wizard for leak suspects, top components after opening
 				a snapshot.</dd>
 			</dlentry>
 			<dlentry>
 				<dt>Hide popup query help</dt>
 				<dd>Do not display the help panel underneath the query wizard, unless F1 or the help button
 				is pressed.</dd>
 			</dlentry>
 			<dlentry>
 				<dt>Hide the Welcome screen on launch</dt>
 				<dd>The welcome page has the 'Overview' and 'Tutorials' tabs and the 'Workbench' button.
 				The welcome page is closed by going to the workbench.
 				Selecting this option means that when Memory Analyzer is started the workbench is displayed
 				first.
 				The welcome page can be reopened from the workbench by:
 					<menucascade>
 						<uicontrol>Help</uicontrol>
 						<uicontrol>Welcome</uicontrol>
 					</menucascade>
 				</dd>
 			</dlentry>
 			<dlentry>
 				<dt>Bytes Display</dt>
 				<dd>
 					<p>
 						There is a option (from MAT 1.5 onwards) to display bytes in B, KB, MB, GB, or Smart formats.
 						The default is to always display in Bytes format to match the previous MAT behavior and not cause any confusion.
 						The option can be changed in the Eclipse preference dialog or using <option>-Dbytes_display=(bytes|kilobytes|megabytes|gigabytes|smart)</option>.
 					</p>
 					<image href="../mimes/size_units_cfg.png">
 						<alt>Preferences dialog to configure size units</alt>
 					</image>
 					<dl>
 						<dlentry>
 							<dt>Bytes (B)</dt>
 							<dd>Memory counted in single bytes</dd>
 						</dlentry>
 						<dlentry>
 							<dt>Kilobytes (KB)</dt>
 							<dd>Memory counted in units of 1,024 bytes</dd>
 						</dlentry>
 						<dlentry>
 							<dt>Megabytes (MB)</dt>
 							<dd>Memory counted in units of 1,048,576 bytes</dd>
 						</dlentry>
 						<dlentry>
 							<dt>Gigabytes (GB)</dt>
 							<dd>Memory counted in units of 1,073,741,824‬ bytes</dd>
 						</dlentry>
 						<dlentry>
 							<dt>Smart</dt>
 							<dd>In Smart mode, if the value is a gigabyte or more, display in gigabytes; similarly for megabytes and kilobytes; otherwise, display in bytes.
 								<image href="../mimes/size_units_sample.png">
 									<alt>A histogram showing entries with different size units</alt>
 								</image>
 							</dd>
 						</dlentry>
 					</dl>
 				</dd>
 			</dlentry>
 			<dlentry id="dtfj_preferences">
 				<dt>
 					<menucascade>
 						<uicontrol>Memory Analyzer</uicontrol>
 						<uicontrol>DTFJ Parser</uicontrol>
 						<uicontrol>Stack frames</uicontrol>
 					</menucascade>
 				</dt>
 				<dd>
 					Normally stack frames are just shown in the Thread overview and stacks query.
 					This option allows stack frames to be treated as objects which are then visible like any other Java object in a Memory Analyzer view.
 					<dl>
 					<dlentry>
 					<dt>Normal
 					</dt>
 					<dd>Stack frames are just shown in the Thread overview and stacks query.
 					</dd>
 					</dlentry>
 					<dlentry>
 					<dt>Only stack frames as pseudo-objects
 					</dt>
 					<dd>
 					Stack frames are of type <codeph>&lt;stack frame&gt;</codeph>, size 0, with method names and source file
 and line numbers via fields and a name resolver.
 					<p>This is useful when looking at paths to and from objects via local variables as the stack frames are visible in
 					the paths to GC roots queries.</p>
 					</dd>
 					</dlentry>
 					<dlentry>
 					<dt>Stack frames as pseudo-objects and running methods as pseudo-classes
 					</dt>
 					<dd>Stack frames of are of type <codeph>packageName.className.methodName(Signature)ReturnType</codeph> extending <codeph>&lt;method&gt;</codeph> representing the method being executed,
 of size the stack frame size, with source file and line number via fields and a name resolver,
 and those methods are pseudo-classes of type <codeph>&lt;method type&gt;</codeph> of size 0.
 					<p>This can be useful to find out which methods are currently running and how much stack space they take up.
 					To examine running methods then take the histogram view, filter by '\(', then sort by instances or instance size.
 					</p>
 					</dd>
 					</dlentry>
 					<dlentry>
 					<dt>Stack frames as pseudo-objects and all methods as pseudo-classes
 					</dt>
 					<dd>Stack frames of are of type <codeph>packageName.className.methodName(Signature)ReturnType</codeph> extending <codeph>&lt;method&gt;</codeph>codeph> representing the method being executed,
 of size the stack frame size, with source file and line number via fields and a name resolver,
 and all methods are pseudo-classes of type <codeph>&lt;method type&gt;</codeph> of size based on the JITted and byte code sizes. The method sizes are then not part of the class size.
 					<p>This can be useful to find out which methods have large JITted or byte code sizes. They can be viewed by going to the histogram view,
 					then selecting <codeph>&lt;method type&gt;</codeph> and listing objects.
 					</p>
 					</dd>
 					</dlentry>
 					</dl>

 				</dd>
 			</dlentry>
 			<dlentry>
 				<dt>
 					<menucascade>
 						<uicontrol>Memory Analyzer</uicontrol>
 						<uicontrol>DTFJ Parser</uicontrol>
 						<uicontrol>Suppress Class native memory size calculations (IBM system dumps)</uicontrol>
 					</menucascade>
 				</dt>
 				<dd>
 					By default, when MAT parses IBM system dumps, the size of classes includes some of the amount of native memory in the Java process (but outside of the Java heap) which is related to those classes such as native memory for bytecode and JIT compiled code for the class methods. Check this option to disable this calculation and only report Java heap usage.
 				</dd>
 			</dlentry>
 					<dlentry id="hprof_preferences">
 						<dt>
 							<menucascade>
 								<uicontrol>Memory Analyzer</uicontrol>
 								<uicontrol>HPROF Parser</uicontrol>
 								<uicontrol>Parser Strictness</uicontrol>
 							</menucascade>
 						</dt>
 						<dd>
 							By default, the HPROF parser runs in strict mode. This means that
 							any deviation in the file from the
 							<xref format="html"
 								href="http://java.net/downloads/heap-snapshot/hprof-binary-format.html">HPROF specification</xref>
 							is treated as a severe error, an exception is thrown, and dump
 							processing stops. There is one exception to this rule which is
 							that we can reliably workaround the observed issue in
 							<xref format="html"
 								href="https://bugs.eclipse.org/bugs/show_bug.cgi?id=404679">bug #404679</xref>
 							.

 							<p>This
 								is a change in behavior from previous
 								releases when a
 								warning
 								was
 								shown in the error log and processing
 								continued. This
 								default
 								change was made to alert the user to a
 								potential problem
 								either
 								with the file itself or a bug in the JVM
 								or in MAT. You may
 								choose
 								to change the strictness of the parser:</p>
 							<dl>
 								<dlentry>
 									<dt>Strict
 									</dt>
 									<dd>Default. Any deviation from the HPROF specification causes
 										an exception to be thrown and processing to stop (with the
 										exception of bug #404679). This option may be specified on the
 										command line with -DhprofStrictnessStop=true
 									</dd>
 								</dlentry>
 								<dlentry>
 									<dt>Warning
 									</dt>
 									<dd>
 										Choose this option to revert to the old behavior where a
 										warning is printed to the error log and processing continues.
 										This option is not recommended because the warning is probably
 										a sign of a problem. Please open a bug report instead. This
 										option may be specified on the
 										command line with
 										-DhprofStrictnessWarning=true
 									</dd>
 								</dlentry>
 							</dl>

 						</dd>
 					</dlentry>
 			<dlentry>
 				<dt>
 					<menucascade>
 						<uicontrol>General</uicontrol>
 						<uicontrol>Appearance</uicontrol>
 						<uicontrol>Colors and Fonts</uicontrol>
 						<uicontrol>Memory Analyzer</uicontrol>
 						<uicontrol>OQL comment color</uicontrol>
 					</menucascade>
 				</dt>
 				<dd>Modify the color of comments in the Object Query Language (OQL) studio.</dd>
 			</dlentry>
 			<dlentry>
 				<dt>
 					<menucascade>
 						<uicontrol>General</uicontrol>
 						<uicontrol>Appearance</uicontrol>
 						<uicontrol>Colors and Fonts</uicontrol>
 						<uicontrol>Memory Analyzer</uicontrol>
 						<uicontrol>OQL keyword color</uicontrol>
 					</menucascade>
 				</dt>
 				<dd>Modify the color of keywords in the Object Query Language (OQL) studio.</dd>
 			</dlentry>
 			</dl>
 			</p>
 		</context>
 	</taskbody>
 </task>
	<?xml version="1.0" encoding="UTF-8"?>
	<!--
	Copyright (c) 2008, 2019 SAP AG and IBM Corporation.
	All rights reserved. This program and the accompanying materials
	are made available under the terms of the Eclipse Public License v1.0
	which accompanies this distribution, and is available at
	http://www.eclipse.org/legal/epl-v10.html

	Contributors:
	SAP AG - initial API and implementation
	IBM Corporation - preference page and OQL
	-->
	<!DOCTYPE task PUBLIC "-//OASIS//DTD DITA Task//EN" "task.dtd" >
	<task id="task_configure_mat" xml:lang="en-us">
	<title>Memory Analyzer Configuration</title>
	<prolog>
	<copyright>
	<copyryear year=""></copyryear>
	<copyrholder>
	Copyright (c) 2008, 2019 SAP AG and others.
	All rights reserved. This program and the accompanying materials
	are made available under the terms of the Eclipse Public License v1.0
	which accompanies this distribution, and is available at
	http://www.eclipse.org/legal/epl-v10.html
	</copyrholder>
	</copyright>
	</prolog>

	<taskbody>
	<context>
	<p>
	Well, analyzing big heap dumps can also require more heap space.
	Give it
	some more memory (possible by running on a 64-bit machine):
	</p>

	<codeph>MemoryAnalyzer.exe -vmargs -Xmx4g</codeph>

	<p>Alternatively, edit the MemoryAnalyzer.ini to
	contain:</p>

	<p>
	<codeblock>-vmargs
	-Xmx4g</codeblock>
	</p>

	<p>
	For more details, check out the section
	<xref format="html"
	href="http://help.eclipse.org/topic/org.eclipse.platform.doc.user/tasks/running_eclipse.htm">Running Eclipse</xref>
	in the Help Center. It also contains more details if you are running
	on Mac OS X.
	</p>

	<p>If you are running the Memory Analyzer inside your Eclipse SDK,
	you need to edit the eclipse.ini file.</p>

	<p>The memory intensive parts are the parsing of the dump and building
	of the dominator tree. Try parsing the heap
	dump from the command line, perhaps on a bigger machine. The dumps
	and index files can then be copied to a more convenient machine.
	Once the dump has been parsed, it usually can
	be opened with less memory in the GUI.
	As a rough estimate if the number of objects is N and the number of
	classes C, it might take at least T bytes to parse and build the dominator tree
	where:
	<lines>T ≈ N * 28.25 + C * 1000 + P</lines>
	P is the space used by the DTFJ or HPROF parsers. For a PHD file, the space could be:
	<lines>P ≈ C * 1000</lines>
	Memory Analyzer uses additional memory for caching index files, so
	performance will be better if there is more memory available than the minimum
	required to parse a dump.
	</p>
	<p>Memory Analyzer has an architectural limit of 2<sup>31</sup> - 3 objects,
	a current limit of 2<sup>31</sup> - 8 = 2,147,483,640 objects, but has not been
	tested with that many objects. The current record is a heap dump file of 70Gbytes
	containing 1,160,000,000 objects, which was opened with Memory Analyzer running
	with a 64Gbyte heap.</p>

	<p>The preference page is opened via a menu option.
	<menucascade>
	<uicontrol>Window</uicontrol>
	<uicontrol>Preferences</uicontrol>
	</menucascade>
	</p>
	<p id="preferences">The preferences page allows various aspects of Memory Analyzer to be controlled.
	<image href="../mimes/preferences.png">
	<alt>preferences page showing general options with DTFJ subpage</alt>
	</image>
	<dl>
	<dlentry>
	<dt>Keep unreachable objects</dt>
	<dd>Objects that appear unreachable from GC roots are not discarded in an early stage of
	Memory Analyzer processing, but are retained for further analysis.</dd>
	</dlentry>
	<dlentry>
	<dt>Hide the getting started wizard</dt>
	<dd>Controls whether to display a wizard for leak suspects, top components after opening
	a snapshot.</dd>
	</dlentry>
	<dlentry>
	<dt>Hide popup query help</dt>
	<dd>Do not display the help panel underneath the query wizard, unless F1 or the help button
	is pressed.</dd>
	</dlentry>
	<dlentry>
	<dt>Hide the Welcome screen on launch</dt>
	<dd>The welcome page has the 'Overview' and 'Tutorials' tabs and the 'Workbench' button.
	The welcome page is closed by going to the workbench.
	Selecting this option means that when Memory Analyzer is started the workbench is displayed
	first.
	The welcome page can be reopened from the workbench by:
	<menucascade>
	<uicontrol>Help</uicontrol>
	<uicontrol>Welcome</uicontrol>
	</menucascade>
	</dd>
	</dlentry>
	<dlentry>
	<dt>Bytes Display</dt>
	<dd>
	<p>
	There is a option (from MAT 1.5 onwards) to display bytes in B, KB, MB, GB, or Smart formats.
	The default is to always display in Bytes format to match the previous MAT behavior and not cause any confusion.
	The option can be changed in the Eclipse preference dialog or using <option>-Dbytes_display=(bytes\|kilobytes\|megabytes\|gigabytes\|smart)</option>.
	</p>
	<image href="../mimes/size_units_cfg.png">
	<alt>Preferences dialog to configure size units</alt>
	</image>
	<dl>
	<dlentry>
	<dt>Bytes (B)</dt>
	<dd>Memory counted in single bytes</dd>
	</dlentry>
	<dlentry>
	<dt>Kilobytes (KB)</dt>
	<dd>Memory counted in units of 1,024 bytes</dd>
	</dlentry>
	<dlentry>
	<dt>Megabytes (MB)</dt>
	<dd>Memory counted in units of 1,048,576 bytes</dd>
	</dlentry>
	<dlentry>
	<dt>Gigabytes (GB)</dt>
	<dd>Memory counted in units of 1,073,741,824‬ bytes</dd>
	</dlentry>
	<dlentry>
	<dt>Smart</dt>
	<dd>In Smart mode, if the value is a gigabyte or more, display in gigabytes; similarly for megabytes and kilobytes; otherwise, display in bytes.
	<image href="../mimes/size_units_sample.png">
	<alt>A histogram showing entries with different size units</alt>
	</image>
	</dd>
	</dlentry>
	</dl>
	</dd>
	</dlentry>
	<dlentry id="dtfj_preferences">
	<dt>
	<menucascade>
	<uicontrol>Memory Analyzer</uicontrol>
	<uicontrol>DTFJ Parser</uicontrol>
	<uicontrol>Stack frames</uicontrol>
	</menucascade>
	</dt>
	<dd>
	Normally stack frames are just shown in the Thread overview and stacks query.
	This option allows stack frames to be treated as objects which are then visible like any other Java object in a Memory Analyzer view.
	<dl>
	<dlentry>
	<dt>Normal
	</dt>
	<dd>Stack frames are just shown in the Thread overview and stacks query.
	</dd>
	</dlentry>
	<dlentry>
	<dt>Only stack frames as pseudo-objects
	</dt>
	<dd>
	Stack frames are of type <codeph><stack frame></codeph>, size 0, with method names and source file
	and line numbers via fields and a name resolver.
	<p>This is useful when looking at paths to and from objects via local variables as the stack frames are visible in
	the paths to GC roots queries.</p>
	</dd>
	</dlentry>
	<dlentry>
	<dt>Stack frames as pseudo-objects and running methods as pseudo-classes
	</dt>
	<dd>Stack frames of are of type <codeph>packageName.className.methodName(Signature)ReturnType</codeph> extending <codeph><method></codeph> representing the method being executed,
	of size the stack frame size, with source file and line number via fields and a name resolver,
	and those methods are pseudo-classes of type <codeph><method type></codeph> of size 0.
	<p>This can be useful to find out which methods are currently running and how much stack space they take up.
	To examine running methods then take the histogram view, filter by '\(', then sort by instances or instance size.
	</p>
	</dd>
	</dlentry>
	<dlentry>
	<dt>Stack frames as pseudo-objects and all methods as pseudo-classes
	</dt>
	<dd>Stack frames of are of type <codeph>packageName.className.methodName(Signature)ReturnType</codeph> extending <codeph><method></codeph>codeph> representing the method being executed,
	of size the stack frame size, with source file and line number via fields and a name resolver,
	and all methods are pseudo-classes of type <codeph><method type></codeph> of size based on the JITted and byte code sizes. The method sizes are then not part of the class size.
	<p>This can be useful to find out which methods have large JITted or byte code sizes. They can be viewed by going to the histogram view,
	then selecting <codeph><method type></codeph> and listing objects.
	</p>
	</dd>
	</dlentry>
	</dl>

	</dd>
	</dlentry>
	<dlentry>
	<dt>
	<menucascade>
	<uicontrol>Memory Analyzer</uicontrol>
	<uicontrol>DTFJ Parser</uicontrol>
	<uicontrol>Suppress Class native memory size calculations (IBM system dumps)</uicontrol>
	</menucascade>
	</dt>
	<dd>
	By default, when MAT parses IBM system dumps, the size of classes includes some of the amount of native memory in the Java process (but outside of the Java heap) which is related to those classes such as native memory for bytecode and JIT compiled code for the class methods. Check this option to disable this calculation and only report Java heap usage.
	</dd>
	</dlentry>
	<dlentry id="hprof_preferences">
	<dt>
	<menucascade>
	<uicontrol>Memory Analyzer</uicontrol>
	<uicontrol>HPROF Parser</uicontrol>
	<uicontrol>Parser Strictness</uicontrol>
	</menucascade>
	</dt>
	<dd>
	By default, the HPROF parser runs in strict mode. This means that
	any deviation in the file from the
	<xref format="html"
	href="http://java.net/downloads/heap-snapshot/hprof-binary-format.html">HPROF specification</xref>
	is treated as a severe error, an exception is thrown, and dump
	processing stops. There is one exception to this rule which is
	that we can reliably workaround the observed issue in
	<xref format="html"
	href="https://bugs.eclipse.org/bugs/show_bug.cgi?id=404679">bug #404679</xref>
	.

	<p>This
	is a change in behavior from previous
	releases when a
	warning
	was
	shown in the error log and processing
	continued. This
	default
	change was made to alert the user to a
	potential problem
	either
	with the file itself or a bug in the JVM
	or in MAT. You may
	choose
	to change the strictness of the parser:</p>
	<dl>
	<dlentry>
	<dt>Strict
	</dt>
	<dd>Default. Any deviation from the HPROF specification causes
	an exception to be thrown and processing to stop (with the
	exception of bug #404679). This option may be specified on the
	command line with -DhprofStrictnessStop=true
	</dd>
	</dlentry>
	<dlentry>
	<dt>Warning
	</dt>
	<dd>
	Choose this option to revert to the old behavior where a
	warning is printed to the error log and processing continues.
	This option is not recommended because the warning is probably
	a sign of a problem. Please open a bug report instead. This
	option may be specified on the
	command line with
	-DhprofStrictnessWarning=true
	</dd>
	</dlentry>
	</dl>

	</dd>
	</dlentry>
	<dlentry>
	<dt>
	<menucascade>
	<uicontrol>General</uicontrol>
	<uicontrol>Appearance</uicontrol>
	<uicontrol>Colors and Fonts</uicontrol>
	<uicontrol>Memory Analyzer</uicontrol>
	<uicontrol>OQL comment color</uicontrol>
	</menucascade>
	</dt>
	<dd>Modify the color of comments in the Object Query Language (OQL) studio.</dd>
	</dlentry>
	<dlentry>
	<dt>
	<menucascade>
	<uicontrol>General</uicontrol>
	<uicontrol>Appearance</uicontrol>
	<uicontrol>Colors and Fonts</uicontrol>
	<uicontrol>Memory Analyzer</uicontrol>
	<uicontrol>OQL keyword color</uicontrol>
	</menucascade>
	</dt>
	<dd>Modify the color of keywords in the Object Query Language (OQL) studio.</dd>
	</dlentry>
	</dl>
	</p>
	</context>
	</taskbody>
	</task>