<?xml version="1.0" encoding="UTF-8" ?>
<?xml-stylesheet type="text/xsl" href="http://blogs.msdn.com/utility/FeedStylesheets/rss.xsl" media="screen"?><rss version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:slash="http://purl.org/rss/1.0/modules/slash/" xmlns:wfw="http://wellformedweb.org/CommentAPI/"><channel><title>So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx</link><description>UPDATE 5/5/2004: I posted a correction to this post here . Sorry if there was any inconvenience. Raymond Chen posted a fascinating comment on the dangers that paging causes server applications the other day. One of the people commenting asked one of the</description><dc:language>en-US</dc:language><generator>CommunityServer 2.1 SP1 (Build: 61025.2)</generator><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#92046</link><pubDate>Thu, 18 Mar 2004 17:04:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:92046</guid><dc:creator>Mat Hall</dc:creator><description>Thanks!  That's answered my question from Raymond's blog...  (And now I've added yet another thing to read to my list; at this rate I'll never get any work done!)</description></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#92058</link><pubDate>Thu, 18 Mar 2004 17:17:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:92058</guid><dc:creator>Mat Hall</dc:creator><description>Forgot the &amp;lt;a href=&lt;a target="_new" href="http://yesihaveabeard.blogspot.com/2004_03_01_yesihaveabeard_archive.html#107963321132301660&amp;gt;trackback.&amp;lt;/a&amp;gt;"&gt;http://yesihaveabeard.blogspot.com/2004_03_01_yesihaveabeard_archive.html#107963321132301660&amp;gt;trackback.&amp;lt;/a&amp;gt;&lt;/a&gt;</description></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#94480</link><pubDate>Tue, 23 Mar 2004 12:38:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:94480</guid><dc:creator>Marco Russo</dc:creator><description>Larry, I think that there is a generally wrong assumptions taken on paging file size. There are some &amp;quot;magic formula&amp;quot; about the right size of paging file, but mostly are based upon hardware (and memory) we had in the 90s. Now it is common to have a 512Mb RAM PC, I work all day with a 1.5Gb notebook and with a 2Gb desktop. In these cases could be not so smart to have a paging file of 2 or 3 Gb (like some of these &amp;quot;magic formula&amp;quot; suggest you). Sometime it's better to avoid having no paging file (option you have starting from Windows XP) but a reasonable 256/512Mb is a right size even if you have Gb of RAM.</description></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#94634</link><pubDate>Tue, 23 Mar 2004 17:34:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:94634</guid><dc:creator>Larry Osterman</dc:creator><description>Marco, remember that the paging file has to be big enough to hold a snapshot of all of physical RAM (for the memory dump system recovery option).  That's the minimum paging file size the system sets.&lt;br&gt;&lt;br&gt;It may be incorrect, but that's where the numbers come from.&lt;br&gt;</description></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#94910</link><pubDate>Tue, 23 Mar 2004 23:09:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:94910</guid><dc:creator>IS</dc:creator><description>It's not 100% on topic and this is just IMHO and AFAIK(I am not an NT guy). I am not trying to flame anyone.&lt;br&gt;NT's paging algorithms are somewhat suboptimal. NT was designed from scratch and authors put lots of untested conceptual ideas into it. Some are very nice and help memory management big time (some depending on hardware). For instance page coloring was sort of modern feature at the time and it does help to reduce cache conflicts. There are innovations like that outside of MM too. The best known one would be the  registry. But this is totally OT, I wanted to talk about the MM.&lt;br&gt;NT decides which pages to clean (sync with backing store) based on working set model. Most of other common OSes use some kind of LRU approximation. Some systems have WS as a second advisory algorithm, but LRU normally takes a priority.&lt;br&gt;Problem with WS is that is has hard time adopting to changing load conditions. Hence horrable performance under NT when someone decides to run a large query. Computer almost goes offline. All the foreground memory pushed to backing store and once query ends, it's pretty much reload back into the same frames same data.&lt;br&gt;Anyone cares to comment?&lt;br&gt;&lt;br&gt;PS. Personally, I don't understand why did not MS hire 3/4 MM programmers/consultants and fixed this. It should probably take 3 months for alpha quality code.</description></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#94911</link><pubDate>Tue, 23 Mar 2004 23:15:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:94911</guid><dc:creator>Larry Osterman</dc:creator><description>IS, I deleted the first of your dup'ed comments, I'm assuming your browser timed out, since there didn't appear to be any difference between the two.&lt;br&gt;</description></item><item><title>re: paging</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#94922</link><pubDate>Tue, 23 Mar 2004 23:55:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:94922</guid><dc:creator>IS</dc:creator><description>Yes, thank you. I am not sure what happened, but it was not intention to post the same thing twice.&lt;br&gt;Hope everyone understands, I am not trying to flame/blame. I think, I even understand why paging was written the way it was. I was simply curious whether anyone cares to comment on my observations and/or why have not this been addressed [yet].</description></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#94925</link><pubDate>Wed, 24 Mar 2004 00:01:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:94925</guid><dc:creator>Larry Osterman</dc:creator><description>Unfortunately I can't answer your comment :(  As I said above, I'm not a MM guy.  I'm a networking/security/email/lotsaotherstuff guy but not MM :(&lt;br&gt;</description></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#94968</link><pubDate>Wed, 24 Mar 2004 01:17:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:94968</guid><dc:creator>Rick</dc:creator><description>Is there any reason to set the paging file greater than 4095 (ie. the maximum addressable memory space)?  If so, why?</description></item><item><title>Mea Culpa (it's corrections time).</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#126534</link><pubDate>Wed, 05 May 2004 20:29:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:126534</guid><dc:creator>Larry Osterman's WebLog</dc:creator><description /></item><item><title>re: So why does NT require such a wonking great big paging file on my machine?</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#132053</link><pubDate>Fri, 14 May 2004 16:51:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:132053</guid><dc:creator>Dunc</dc:creator><description>&amp;quot;Is there any reason to set the paging file greater than 4095&amp;quot; -- because EACH APP gets the same virtual memory model.</description></item><item><title>More office hijinks</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#153913</link><pubDate>Sat, 12 Jun 2004 01:30:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:153913</guid><dc:creator>Larry Osterman's WebLog</dc:creator><description /></item><item><title>re: Summary of the recent spate of /3GB articles</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#218883</link><pubDate>Mon, 23 Aug 2004 18:17:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:218883</guid><dc:creator>The Old New Thing</dc:creator><description /></item><item><title>re: Still more misinformation about virtual memory and paging files</title><link>http://blogs.msdn.com/larryosterman/archive/2004/03/18/92010.aspx#402449</link><pubDate>Sat, 26 Mar 2005 06:34:00 GMT</pubDate><guid isPermaLink="false">91d46819-8472-40ad-a661-2c78acb4018c:402449</guid><dc:creator>Larry Osterman's WebLog</dc:creator><description /></item></channel></rss>