Talk:Prefetcher

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Citation already provided in "Incorrect tweaking of the prefetcher" section[edit]

If you view the link that is currently cited in the "Incorrect tweaking of the prefetcher" section numberd 3 you can see the following text - "So, first off, it is a bad idea to periodically clean out that folder as some tech sites suggest. For one thing, XP will just re-create that data anyways; secondly, it trims the files anyways if there's ever more than 128 of them so that it doesn't needlessly consume space. So not only is deleting the directory totally unnecessary, but you're also putting a temporary dent in your PC's performance." Can we remove the citation needed flag? —Preceding unsigned comment added by Ciper (talkcontribs) 05:33, 14 November 2007 (UTC)[reply]

Misinformation[edit]

Being new to wiki-editing, I have no clue what I should do, but this article is either extremely confusing or in blatant disregard of fact.

Disabling the prefetcher does in fact reduce OS boot times (kernel load time), but increases application load time. While I am unable to cite a source, anyone who has followed the steps (disabling in registry, deleting select/all files in the prefetch folder) will notice that their boot times will decrease. People disputing this often include the load time for explorer.exe (the Windows GUI shell) which is noticeably increased. After a fresh install, this benefit of disabling prefetching is not immediately recognizable, but as the computer ages, more and more applications are run. over time, this adds up and slows down Windows' boot if prefetch is enabled because not only does prefetch need to load more and more installed applications, but it will also load into memory applications that will only be used once (such as the setup executables). On top of this, uninstalling unused applications does NOT remove the prefetched file and the prefetcher will also load those.

Especially with the performance of today's PCs, there really is no need for the Windows prefetcher as it currently operates.

--Hieyeck 05:27, 9 May 2007 (UTC)[reply]

You're lying, Hieyeck! Read these: [1], [2], [3] and [4] (read the section under Prefetch in the last article) -124.43.192.19 12:37, 19 June 2007 (UTC)[reply]

I wiped the prefetch folder for fun, and the boot time didn't change. This was surprising, because everyone told me i'd be in for a boot time of a couple millennia. Once Vista is loaded, the only apps that load very slightly slower are Word and Photoshop. Excel takes a few seconds with or without prefetch, but half of the startup time is due to loading EUROTOOL.XLA which takes 3 seconds (seriously). Everything else starts up pretty much instantly either way, or the app is a game and most of the startup time is due to a slow DVD check or connection to a server anyway. Does this thing even do anything? YMMV if your hardware is slower and/or you run apps that actually take more than 5 seconds to load. Brother Laz 17:37, 28 October 2007 (UTC)[reply]

Flushing the prefetch cache?[edit]

"A second myth is that the user should delete the prefetch folder contents to speed up the computer. If this is done, Windows will need to re-create all the prefetch files again, thereby slowing down Windows during boot and program starts until the prefetch files are created."

If the prefetch files have become horribly out of date, wouldn't this speed up the computer on the second and subsequent boots? Or does the Windows prefetcher keep the information up to date even with changes to the system? --Damian Yerrick (talk | stalk) 19:54, 9 October 2007 (UTC)[reply]

The latter. -- simxp (talk) 15:28, 12 December 2007 (UTC)[reply]

Deleting the prefetch folder is an 'official' step on the Dell Support website for speeding up the performance of a computer, as I was on their website following some of the steps yesterday. So I'm presuming they think it will help - don't ask me to explain why though - got no idea! Mike Burden (talk) 09:20, 1 August 2008 (UTC)[reply]

Expert cleanup and clarifications needed[edit]

I added cleanup, fact and clarifyme tags within the article, with html comments explaining what I think is needed. One example is that the article mentions "Windows" without specifying which versions are meant. Another is the use of "Defrag.exe" options without explanation of the "-b" option (not listed in http://support.microsoft.com/kb/283080/en-us for XP and option not present in my Vista Home Premium). -84.223.72.181 (talk) 13:58, 24 April 2008 (UTC)[reply]

  • Here is a possible citation for the 'defrag [drive] -b' option:
http://www.langa.com/newsletters/2003/2003-09-29.htm#5

WindowsXP runs this about every 3 days but if you need to optimize manually all you need to do is run a Command window and type defrag c: -b

(This link is broken - simicich) Simicich (talk)

166.70.81.97 (talk) 20:04, 12 November 2008 (UTC)[reply]

Patent[edit]

I'm glad it's patented. So that the techs won't show on my Linux machine, that is. Rursus dixit. (mbork3!) 19:41, 3 July 2010 (UTC)[reply]

citation needed on /B flag[edit]

Someone has marked this Citation Needed. I see that if I run defrag with /b it works - that is,

Invoking boot optimization on WIN7 (C:)...

but it is not in the help. It is not in the official writeup although someone said that the version of defrag in Windows Server 2008 R2 exposed it in the /? help for a short while.

http://technet.microsoft.com/en-us/library/cc731650(WS.10).aspx

I have no idea if this constitutes a citation (or perhaps would be useful as a citation), but I suspect not in either case. How do you provide a citation for an undocumented flag? The existence of the flag is obviously true, but is it more important to include and spread the knowledge or to provide a cite and remove unsupported material? I believe that we should provide knowledge first (especially when it is easy to demonstrate that it is accurate, as it is here), but what do I know?

BTW, the langa.com newsletter link is broken.

Simicich (talk) 06:06, 13 April 2011 (UTC)[reply]