<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:series="http://unfoldingneurons.com/"
		>
<channel>
	<title>Comments on: Defragging The Windows Page File</title>
	<atom:link href="http://www.pcmech.com/article/defragging-the-windows-page-file/feed/" rel="self" type="application/rss+xml" />
	<link>http://www.pcmech.com/article/defragging-the-windows-page-file/</link>
	<description>Helping Normal People Get Their Geek On</description>
	<lastBuildDate>Mon, 23 Nov 2009 01:37:34 -0500</lastBuildDate>
	<generator>http://wordpress.org/?v=2.8.5</generator>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
		<item>
		<title>By: GC</title>
		<link>http://www.pcmech.com/article/defragging-the-windows-page-file/comment-page-1/#comment-26897</link>
		<dc:creator>GC</dc:creator>
		<pubDate>Fri, 26 Jun 2009 22:46:18 +0000</pubDate>
		<guid isPermaLink="false">http://www.pcmech.com/?p=7059#comment-26897</guid>
		<description>Be wary of actually deleting the page file before defragmenting old system disks.

My experience (and that of others posting elsewhere) has been that this results in a page file that is has much worse fragmentation (e.g., &gt; 1000 fragments, I saw one commenter say theirs had &gt;26,000 fragments).  

I suspect that what is happening is that the disk defragmenter allocates all the large blocks to existing files, so when you allocate a new page file after defragmentation, the new page file just gets the leftover pieces.

Three guesses on why there are only small pieces left over (may be a combination of all three):
1. on old disks with many OS system packs and hotfixes, system files become spread all over the disk, and the disk defragmenter can&#039;t move files that are in use, and so the spaces in between them are small.
2. on old disks there may be sectors that have gone bad spread all over the disk.  Can&#039;t move those.
3. disk defragmenter and PageDefrag just defragment existing files, not free space.  (In fact they may keep free space fragmented on purpose, so there will be space nearby when any file is extended.)

(I had hoped PageDefrag would move all the OS files [green slivers in disk defragmenter] together in larger blocks, but it doesn&#039;t move them.)

Instead of actually deleting the page file, stop using the paging file without actually deleting it, so disk defragmenter can try to defragment it.</description>
		<content:encoded><![CDATA[<p>Be wary of actually deleting the page file before defragmenting old system disks.</p>
<p>My experience (and that of others posting elsewhere) has been that this results in a page file that is has much worse fragmentation (e.g., &gt; 1000 fragments, I saw one commenter say theirs had &gt;26,000 fragments).  </p>
<p>I suspect that what is happening is that the disk defragmenter allocates all the large blocks to existing files, so when you allocate a new page file after defragmentation, the new page file just gets the leftover pieces.</p>
<p>Three guesses on why there are only small pieces left over (may be a combination of all three):<br />
1. on old disks with many OS system packs and hotfixes, system files become spread all over the disk, and the disk defragmenter can&#8217;t move files that are in use, and so the spaces in between them are small.<br />
2. on old disks there may be sectors that have gone bad spread all over the disk.  Can&#8217;t move those.<br />
3. disk defragmenter and PageDefrag just defragment existing files, not free space.  (In fact they may keep free space fragmented on purpose, so there will be space nearby when any file is extended.)</p>
<p>(I had hoped PageDefrag would move all the OS files [green slivers in disk defragmenter] together in larger blocks, but it doesn&#8217;t move them.)</p>
<p>Instead of actually deleting the page file, stop using the paging file without actually deleting it, so disk defragmenter can try to defragment it.</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Larry Miller</title>
		<link>http://www.pcmech.com/article/defragging-the-windows-page-file/comment-page-1/#comment-22426</link>
		<dc:creator>Larry Miller</dc:creator>
		<pubDate>Mon, 23 Mar 2009 04:50:02 +0000</pubDate>
		<guid isPermaLink="false">http://www.pcmech.com/?p=7059#comment-22426</guid>
		<description>The problems caused by a fragmented pagefile have been greatly exagerated. Fragmentation impairs performance when a file is read or written serially. This NEVER happens with the pagefile. Pagefile IO is rarely more than 64K at a time and usually mixed with other paging activity. Yes, other paging activity, the large majority of paging does not use the pagefile. Pagefile fragmentation will not normally effect performance except in extreme cases. An extreme case would be hundreds of fragments.

The initial portion of the pagefile will not become fragmented over time. Only when the system is forced to expand the pagefile will this occur. Windows will not do this without announcing this with a warning dialog. If you have not seen these warnings, the pagefile has not been expanded. In any event any pagefile expansion will be deleted on a reboot.

The proper way to prevent pagefile expansion is to set the initial size so this need never occur. Trying to do this by lowering the maximum setting is totally wrong and will severly impair performance or cause application failures.

If pagefile fragmentation is a problem then make sure the initial size is adequate and then use pagedefrag. Do this once and you will never have a fragmented pagefile again, even after months of heavy continuous use.

Larry Miller 
Microsoft MCSA</description>
		<content:encoded><![CDATA[<p>The problems caused by a fragmented pagefile have been greatly exagerated. Fragmentation impairs performance when a file is read or written serially. This NEVER happens with the pagefile. Pagefile IO is rarely more than 64K at a time and usually mixed with other paging activity. Yes, other paging activity, the large majority of paging does not use the pagefile. Pagefile fragmentation will not normally effect performance except in extreme cases. An extreme case would be hundreds of fragments.</p>
<p>The initial portion of the pagefile will not become fragmented over time. Only when the system is forced to expand the pagefile will this occur. Windows will not do this without announcing this with a warning dialog. If you have not seen these warnings, the pagefile has not been expanded. In any event any pagefile expansion will be deleted on a reboot.</p>
<p>The proper way to prevent pagefile expansion is to set the initial size so this need never occur. Trying to do this by lowering the maximum setting is totally wrong and will severly impair performance or cause application failures.</p>
<p>If pagefile fragmentation is a problem then make sure the initial size is adequate and then use pagedefrag. Do this once and you will never have a fragmented pagefile again, even after months of heavy continuous use.</p>
<p>Larry Miller<br />
Microsoft MCSA</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: SuperFluid</title>
		<link>http://www.pcmech.com/article/defragging-the-windows-page-file/comment-page-1/#comment-20215</link>
		<dc:creator>SuperFluid</dc:creator>
		<pubDate>Tue, 03 Feb 2009 03:07:02 +0000</pubDate>
		<guid isPermaLink="false">http://www.pcmech.com/?p=7059#comment-20215</guid>
		<description>Pagedefrag has a hardcoded error in it that shows its face if you tune your disks and filesystems for maximum performance by moving your eventlogs off of C: Drive. So, if you moved your registry hives or logs from their default area to gain performance, a hardcode that existed for pagefile.sys that APK fixed for Dr. Russinovich the author of this program, still exists in it, albeit for the registry hives and eventlogs of the operating system and applications.

(Yes, this is doable, &amp; e.g. -&gt; via this registry path area + its variables of HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Eventlog\Application so you can move the application event log file to another disk to unburden the main C drive of that task)

Thus, effectively helping to speed up its operations. 

Credits Alexander Peter Kowalski on that idea from Windows IT Pro magazine article where he helped Mark Russinovich out pointing out this fact to him and that his program also had the pagefile.sys location hardcoded as well, but, even though he alerted Dr. Russinovich of this at that article on Windows IT Pro magazine forums, I do not think it has been fixed for the registry hives or the eventlogs (although Dr. Russinovich fixed them for pagefile.sys at Mr. Kowalski&#039;s suggestions and reguest, using the methods Mr. Kowalski suggested).</description>
		<content:encoded><![CDATA[<p>Pagedefrag has a hardcoded error in it that shows its face if you tune your disks and filesystems for maximum performance by moving your eventlogs off of C: Drive. So, if you moved your registry hives or logs from their default area to gain performance, a hardcode that existed for pagefile.sys that APK fixed for Dr. Russinovich the author of this program, still exists in it, albeit for the registry hives and eventlogs of the operating system and applications.</p>
<p>(Yes, this is doable, &amp; e.g. -&gt; via this registry path area + its variables of HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Eventlog\Application so you can move the application event log file to another disk to unburden the main C drive of that task)</p>
<p>Thus, effectively helping to speed up its operations. </p>
<p>Credits Alexander Peter Kowalski on that idea from Windows IT Pro magazine article where he helped Mark Russinovich out pointing out this fact to him and that his program also had the pagefile.sys location hardcoded as well, but, even though he alerted Dr. Russinovich of this at that article on Windows IT Pro magazine forums, I do not think it has been fixed for the registry hives or the eventlogs (although Dr. Russinovich fixed them for pagefile.sys at Mr. Kowalski&#8217;s suggestions and reguest, using the methods Mr. Kowalski suggested).</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Drew</title>
		<link>http://www.pcmech.com/article/defragging-the-windows-page-file/comment-page-1/#comment-12312</link>
		<dc:creator>Drew</dc:creator>
		<pubDate>Sat, 19 Jul 2008 16:24:09 +0000</pubDate>
		<guid isPermaLink="false">http://www.pcmech.com/?p=7059#comment-12312</guid>
		<description>Great post! I&#039;ve been using Page Defrag since about a month after it came out and love it. It increases a slight amount to your systems boot time while it does the defrag but only around 10-15 seconds, pending your PC speed.
It&#039;s a great little tool and works great IMO</description>
		<content:encoded><![CDATA[<p>Great post! I&#8217;ve been using Page Defrag since about a month after it came out and love it. It increases a slight amount to your systems boot time while it does the defrag but only around 10-15 seconds, pending your PC speed.<br />
It&#8217;s a great little tool and works great IMO</p>
]]></content:encoded>
	</item>
	<item>
		<title>By: Sharron</title>
		<link>http://www.pcmech.com/article/defragging-the-windows-page-file/comment-page-1/#comment-12305</link>
		<dc:creator>Sharron</dc:creator>
		<pubDate>Sat, 19 Jul 2008 14:16:20 +0000</pubDate>
		<guid isPermaLink="false">http://www.pcmech.com/?p=7059#comment-12305</guid>
		<description>I&#039;m an XP user and having had a taste of Vista I decided to stay with XP: Hence I don&#039;t know whether or not the following programs&#039; functions are catered for by default in Vista, though I suspect not.

Worth mentioning, with this in mind, are another two freeware programs; those being NTRegOpt and Erunt. http://www.larshederer.homepage.t-online.de/erunt/

These two download in a single package: Erunt creates a backup of your system registry including open hives at startup and saves it to your a file in user account where it can be accessed from the command prompt in the Recovery Console. 

NTRegOpt optimises the open hives of yout NT registry that most defragmenters can&#039;t get at by defragmaning and compacting.

Another point worth mentioning is that if you use Diskeeper (2007 or 2008) there is a boot-time runonce defragmentation option available from the program&#039;s GUI which also defragments your MFT in addition to paging file.

Also in Diskeeeper is an option to pad out your MFT. (This is done manually with Diskeeper 2007 and automatically by Diskeeper 2008.) What this means is that when windows if first installed it creates a master file table with enough space for x many entries: This isn&#039;t always large enough; and as the disk fills up it can cause the MFT to fragment causing system instability. Diskeeper resizes the MFT based upon an approximation of the expected number of files plus some, calculated by a complicated algorithm. This avoids MFT fragmentation as the disk fills up.</description>
		<content:encoded><![CDATA[<p>I&#8217;m an XP user and having had a taste of Vista I decided to stay with XP: Hence I don&#8217;t know whether or not the following programs&#8217; functions are catered for by default in Vista, though I suspect not.</p>
<p>Worth mentioning, with this in mind, are another two freeware programs; those being NTRegOpt and Erunt. <a href="http://www.larshederer.homepage.t-online.de/erunt/" rel="nofollow">http://www.larshederer.homepage.t-online.de/erunt/</a></p>
<p>These two download in a single package: Erunt creates a backup of your system registry including open hives at startup and saves it to your a file in user account where it can be accessed from the command prompt in the Recovery Console. </p>
<p>NTRegOpt optimises the open hives of yout NT registry that most defragmenters can&#8217;t get at by defragmaning and compacting.</p>
<p>Another point worth mentioning is that if you use Diskeeper (2007 or 2008) there is a boot-time runonce defragmentation option available from the program&#8217;s GUI which also defragments your MFT in addition to paging file.</p>
<p>Also in Diskeeeper is an option to pad out your MFT. (This is done manually with Diskeeper 2007 and automatically by Diskeeper 2008.) What this means is that when windows if first installed it creates a master file table with enough space for x many entries: This isn&#8217;t always large enough; and as the disk fills up it can cause the MFT to fragment causing system instability. Diskeeper resizes the MFT based upon an approximation of the expected number of files plus some, calculated by a complicated algorithm. This avoids MFT fragmentation as the disk fills up.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
