Warp2Search.net » News » October 2005 » CCleaner Cripples Application Load Times

Recently CCleaner has added an internet urban legend as a cleaning option, "Old Prefetch Data". Cleaning the Prefetch folder is an internet Myth that simply will not die due to the gross ignorance of many people in regards to how Windows XP Prefetching works. These same people generally recommend other bogus advice such as disabling Windows Prefetching completely and adding /Prefetch:1 to desktop shortcuts.



"Bottom line: You will NOT improve Windows performance by cleaning out the Prefetch folder. You will, in fact, degrade Windows performance by cleaning out the Prefetch folder."

CCleaner for the most part is a good application, it quickly and easily removes temporary and unused files from Windows. It has a nice interface that clearly shows what has been "cleaned". On neglected systems this can free hundreds of Megabytes of harddisk space. Apparently in the authors quest to clean everything and anything, he blindly ignored how Prefetching works...

Popular Technology.net

Domingo Unregistered

#62741 Posted on: 10/03/2005 07:55 PM He also seemed to have disabled the automatic windows TEMP folder cleaning in the latest release, too. Good program, though. I'm not sure about the prefetcher really affecting you much either way, though. Unless you're crusing XP with a PII or older, I doubt a 1/2 second load difference is noticeable anyway.

Anach Unregistered

#62744 Posted on: 10/03/2005 10:13 PM Its also like the memory recovery programs, that actually slow down your PC. I still have a hard time getting some people to stop using them. Or many various Anti-spyware apps that dont do anything, and remember the good ol' Evidence ellimiator. Mostly these things are spread by word of mouth by people who are just learning how to use their PC and the internet. Finding all these ads and cool programs to *Try before you buy*. Placebo Software. People think its working. So they are happy to use it and tell their friends.

PersianImmortal Unregistered

#62746 Posted on: 10/03/2005 10:30 PM Firstly, disabling prefetch and cleaning out your Prefetch folder does not 'cripple' Windows. Prefetch adds to your Windows bootup time (sometimes significantly) just so it can guess which applications you like to open often and give them a slight boost in opening times. So you had another 10 seconds to your bootup time, to save 1 second on the loading time for say the 5-10 apps you open the most often. In my experimentation over the years I've found that you can noticeable improve Windows performance with minimal impact on app loading times by disabling prefetch, setting Task Scheduler service to Disabled (prevents new prefetch data), and clearing out your prefetch folder.

ChrisW Unregistered

#62747 Posted on: 10/03/2005 10:36 PM I still find a lot of this Prefetch 'feature', to be a bit hit & miss. 100% quicker ? 100% slower if you clean out the Prefetch folder ? Bullshit. I've cleaned out this folder completely, and never saw any sort of drop in performance, in any way. Thing is, Prefetch, will prefetch a lot of stuff you use briefly. I regulary find lots of setup.exe's, installs, etc.. from drivers, programs, etc.. that I've downloaded from the web. I download these programs to a separate partition, then burn off the contents to DVD, and delete the folder. Now, tell me how still having these setup.exe's, installs, etc.. in the prefetch folder, can be a good thing ? Surely, trying to prefetch something that isn't there anymore, is NOT a good thing ? If you uninstall a program that has an entry in the Prefetch folder, it stays in that Prefetch folder. How is this a good thing ? Surely, it's better to check, on a regular basis, for entries that no longer exist ? Too many people on the net, particularly these 'Tweak' websites, really think they know everything.

Laboraffe Unregistered

#62748 Posted on: 10/03/2005 10:38 PM Commentary misses the point with knee-jerk reaction. The prefetch cleaning option is not intended to increase performance but simply remove a run history of applications. Users may not want it recorded that they ever ran pr0nmaster5000.exe or continue to do so from a removeable flash drive.

Cimlite Junior Member Posts: 6 Joined: 2005-11-12

#62749 Posted on: 10/03/2005 11:01 PM Best thing is to clean out the prefetch yourself and only keeping basic stuff. Things like explorer.exe and programs that you actually *do* use a lot you might want to keep in there. Removing explorer.exe from the prefetch does noticibly slowdown the responce time in windows. You can try that yourself easily, just move it to another location. That said, cleaning out old "junk" that you either don't have any more or do not use more than once a month is not a bad thing. Laboraffe, I've heard good things about "pr0nmaster5000.exe"... where exactly can I get it? ;P

Cellar_Dweller Senior Member Posts: 152 Joined: 2003-02-03

#62750 Posted on: 10/03/2005 11:08 PM I agree with some of what he is saying but, there is a real good reason to clear out your Prefetch folder and the biggest reason is: Spyway, Trojans, Adaware. Many recent trojans, virus, spyware etc. will put something within the prefetch (It only makes sense) so that no matter what you clean up and if you forget the prefetch, which MOST people do when cleaning up trojans and the like forgetting this is in XP will soon reboot to see everything they thought they had cleaned up has since returned... LOL.

Mastertech Junior Member Posts: 2 Joined: 2005-04-13

#62751 Posted on: 10/03/2005 11:13 PM I'm not sure about the prefetcher really affecting you much either way, though. Unless you're crusing XP with a PII or older, I doubt a 1/2 second load difference is noticeable anyway. It is NOT a 1/2 second difference. It can be 10 seconds or more for some applications. Firstly, disabling prefetch and cleaning out your Prefetch folder does not 'cripple' Windows. It cripples Windows Boot and Application Load Times. Prefetch adds to your Windows bootup time (sometimes significantly) just so it can guess which applications you like to open often and give them a slight boost in opening times. This is a lie, no it does not. It does not load anything relating to an application until you load it! You don't understand how it works. It is NOT a cache! So you had another 10 seconds to your bootup time, to save 1 second on the loading time for say the 5-10 apps you open the most often. It does not work this way and you have no idea what you are talking about. In my experimentation over the years I've found that you can noticeable improve Windows performance with minimal impact on app loading times by disabling prefetch, setting Task Scheduler service to Disabled (prevents new prefetch data), and clearing out your prefetch folder. You have no idea what you are talking about and have obviously never tested this. 100% quicker ? 100% slower if you clean out the Prefetch folder ? Bullshit. I've cleaned out this folder completely, and never saw any sort of drop in performance, in any way. That is because you have never timed it. Try it sometime and you will see. The article gives a way to test it for yourself. "Thing is, Prefetch, will prefetch a lot of stuff you use briefly. I regulary find lots of setup.exe's, installs, etc.. from drivers, programs, etc.. that I've downloaded from the web. I download these programs to a separate partition, then burn off the contents to DVD, and delete the folder." No it does not. It does not cache ANYTHING. You don't undertand how prefetching works is why you are confused. It only creates a .pf for any application that is launched at least 3 times. Now, tell me how still having these setup.exe's, installs, etc.. in the prefetch folder, can be a good thing ? Surely, trying to prefetch something that isn't there anymore, is NOT a good thing ? You don't understand. The .pf files for applications that are not there are NOT cached! They do nothing but take up 14kb of space. Eventually the prefetcher will delete these files in the mean time they do nothing. If you uninstall a program that has an entry in the Prefetch folder, it stays in that Prefetch folder. How is this a good thing ? How is it a bad thing? The files are so small and will not be used and eventually cleaned. Uninstalled applications are NOT prefetched! Surely, it's better to check, on a regular basis, for entries that no longer exist ? Why bother even with the folder full of items (128) it maybe takes up 5MB. Who cares it is not worth your time to bother. Windows does this automatically. The prefetch cleaning option is not intended to increase performance but simply remove a run history of applications Instead it cripples load times!!! Why? To save 1MB of disk space?M/b>

Mastertech Junior Member Posts: 2 Joined: 2005-04-13

#62752 Posted on: 10/03/2005 11:24 PM I agree with some of what he is saying but, there is a real good reason to clear out your Prefetch folder and the biggest reason is: Spyway, Trojans, Adaware. Many recent trojans, virus, spyware etc. will put something within the prefetch (It only makes sense) so that no matter what you clean up and if you forget the prefetch, which MOST people do when cleaning up trojans and the like forgetting this is in XP will soon reboot to see everything they thought they had cleaned up has since returned... LOL. This is another myth. The .pf files are NOT executeable and are the only things in the Prefetch folder besides the layout.ini. So if a virus did put something in there an AV scanner will find it since it would not be a regular .pf file. Take a look for yourself. Cleaning the rest of the folders contents makes NO sense. Best thing is to clean out the prefetch yourself and only keeping basic stuff. Things like explorer.exe and programs that you actually *do* use a lot you might want to keep in there. No you want to leave ANY file in there for an installed application. Otherwise when you go to launch it you will be crippling it's load time.

ChrisW Unregistered

#62754 Posted on: 10/04/2005 12:31 AM >>That is because you have never timed it. Try it sometime and you will see. The article gives a way to test it for yourself. Believe me, I've done plenty of testing. Explorer.exe, is really the only difference I saw. Boot times, never decreased, or increased. I have 2 Athlon 64 PCs, both with 2 Gig of RAM, and fast SATA drives. Both take around 15 seconds to get to the desktop with no activity, with a full or empty Prefetch folder. It really doesn't make any difference. >>It is NOT a 1/2 second difference. It can be 10 seconds or more for some applications. I have NEVER ever seen ANY application take 10 seconds to run, regardless of the prefetch.

PersianImmortal Unregistered

#62755 Posted on: 10/04/2005 01:14 AM Mastertech, I stand by my comments, which are derived from years of using XP on a variety of systems, and testing a range of options. I know prefetch is supposed to optimize boot layout, and speed up both boot times and application load times. That is the theory. However in practice, even after several weeks of allowing the Prefetcher to do its job unhindered (it needs 3 days to optimize layout and at least two application launches to optimize app launches), I found my bootup times were increasing, and my apps times were not noticeably any faster. I don't need a "methodology" for measuring the success or failure of a feature - I use my own observation skills. It is true to say that if you haven't disabled prefetching, clearing out the prefetch folder is bad, because Windows simply has to restart the process of recreating the files and analyzing things from scratch. However I suggest people try the disabled prefetch/disabled Task Scheduler scenario for themselves and see the difference. I found it a clear-cut case against using the prefetcher.

MajorTom Unregistered

#62756 Posted on: 10/04/2005 01:47 AM Besides which If you run CCleaner then go and check your prefetch folder you will see that it doesn't delete everything .... just "old" data as implied by "Old Prefetch Data".... ho hum

Mastertech Junior Member Posts: 2 Joined: 2005-04-13

#62757 Posted on: 10/04/2005 12:12 PM That said, it wouldn't "cripple" Windows turning it off; just make it as slow as previous versions of Windows are in those regards basically. AGAIN. The article says it "cripples" APPLICATION LOAD TIMES NOT WINDOWS! Mastertech, I stand by my comments, which are derived from years of using XP on a variety of systems, and testing a range of options. I've been doing this for over 15 years and work daily for a PC OEM. I can tell you flat out you never tested this. What is scarey is your offer "tweaking" advice. I know prefetch is supposed to optimize boot layout, and speed up both boot times and application load times. That is the theory. No it is not a theory but a fact. However in practice, even after several weeks of allowing the Prefetcher to do its job unhindered (it needs 3 days to optimize layout and at least two application launches to optimize app launches), I found my bootup times were increasing, and my apps times were not noticeably any faster. Whatever, you never tested or timed this. For all I know you also had the Task Scheduler service disabled and prefetching as well. Anyone can test this with a stop watch. I don't need a "methodology" for measuring the success or failure of a feature - I use my own observation skills. Your "observations skills" are a joke and not factual nor tested. Show me some numbers and a test I can reproduce here. You don't have them because you never tested this. I guarantee you any "test" you come up will show prefetching will improve load times or I will point out what you are doing wrong and why it was not working. ever I suggest people try the disabled prefetch/disabled Task Scheduler scenario for themselves and see the difference. I found it a clear-cut case against using the prefetcher. This article gives a clear way to test this. Why not try it before spreading more bad advice that you don't even bother to test. You can't tell people to "notice" something. They need to time it. You haven't found anything but your ignorance in not testing or understanding how prefetching works. People like you that spread these myths is what is wrong with the Internet and why people have unoptimized systems with bad advice Besides which If you run CCleaner then go and check your prefetch folder you will see that it doesn't delete everything .... just "old" data as implied by "Old Prefetch Data".... ho hum You are clueless. The article clearly explains it deletes prefetch files over two weeks old and why this is stupid. If the NTFS last access date is turned off it deletes them all. PLEASE read the article and try to comprehend it.

PersianImmortal Unregistered

#62758 Posted on: 10/04/2005 12:40 PM Your rabid responses, and your 'Optimize XP' guide which is basically one giant ad for various utilities and sponsored links tells the story by itself mate. I allow people to make their own mind up instead of telling them they're idiots if they don't agree with me. I offer descriptions in my guides and encourage people to experiment. I don't ram "technical facts" down peoples' throats. Simply put: Windows is already optimized out of the box for the average user. The reason we tweak it is for users who want full control over the functions of their XP, and who don't just accept that MS knows best.

TSThomas Junior Member Posts: 2 Joined: 2003-05-10

#62759 Posted on: 10/04/2005 12:51 PM "AGAIN. The article says it "cripples" APPLICATION LOAD TIMES NOT WINDOWS!" Calm down a bit; re-read what I wrote. You seem to have misinterpreted it, I said just that "make it as slow as previous versions of Windows are *in those regards* basically".

PersianImmortal Unregistered

#62760 Posted on: 10/04/2005 01:40 PM The concept behind the prefetcher is simple: It preloads information in anticipation of loading certain applications at a later point. It preloads/pages this information at some earlier point - applications don't magically load faster when prefetch is enabled, a portion of them has been preloaded into memory at some earlier stage. In my experimentation, I timed Windows startup, and found that after several weeks of prefetch being enabled my startup time had increased by several seconds - noticeable and timed. It also seemed to be growing. People who follow my TweakGuides Tweaking Companion end up doing the following, which is not quite 'disabling the prefetcher', but removes the aspects of the prefetcher which blow out startup times: 1. Run Bootvis regularly to optimize boot layout. This is similar to have prefetcher automatically optimize the bootup, except it is the manual method. 2. Prefetcher registry setting is left at default of 3. 3. Disable Task Scheduler - this prevents new prefetch files from being created/recreated. 4. Delete the contents of the Prefetch Folder except for Layout.ini. This decreases bootup time. 5. Defrag the drive. Refernces for this are pages 54 and 157 of my guide, as well as various advice throughout. Now how about you test this scenario out, and tell me if your application times are significantly longer, and your bootup time has not decreased. I have proven it to myself time and time again, and will continue to provide this advice.

Fuzzywuzzy Unregistered

#62761 Posted on: 10/04/2005 02:26 PM Your "advice" does not tranfer to other systems, however. I have also tested the prefetch subsystem in XP on various systems, from lowly computers that barely had enough power to run explorer to state-of-the-art powerhouses. In all cases, turning off prefetch as you suggest LOWERS performance across the board. You do NOT gain anything by turning off prefetch. I have not seen any bootup time increase by using prefetch as it was designed. I don't understand the interest in bootup time, anyway. How often are you rebooting? If you're having to reboot that much, that indicates other problems with your computer system that are unrelated to prefetch. Perhaps you have messed up other systems as well, resulting in decreased stability. Your "advice" may seem appropriate for you, but for mass distribution to other people, it is inaccurate and potentially damaging.

Nemo Unregistered

#62763 Posted on: 10/04/2005 02:38 PM @ Mastertech said: "I've been doing this for over 15 years and work daily for a PC OEM." Well there you go, shot yourself in the foot there, 99.99% of people working for large OEMs know fook all about PC's. I've been woking with PCs (support + software Dev) for a lot longer than 15 years and everyone knows that.

TweakHound Unregistered

#62764 Posted on: 10/04/2005 04:24 PM TSThomas - Prefetching is on for a reason; it's faster. Disabling or clearing it is counterproductive with no real benefit - it doesn't save a noticeable amount of memory nor does it improve performance - it can't really. Well said!

Mastertech Junior Member Posts: 2 Joined: 2005-04-13

#62767 Posted on: 10/04/2005 10:07 PM Your rabid responses, and your 'Optimize XP' guide which is basically one giant ad for various utilities and sponsored links tells the story by itself mate. You're lack of knowledge on how prefetching works and your inclusion of numerous urban legend "tweaks" says it all. I offer descriptions in my guides and encourage people to experiment. I don't ram "technical facts" down peoples' throats. You are right, I test things and find out how they really work before misleading people and can back it up with facts, when they don't understand and actual evidence. Simply put: Windows is already optimized out of the box for the average user. The reason we tweak it is for users who want full control over the functions of their XP, and who don't just accept that MS knows best. Out of the box Windows has alot of fluff turned on that only slows the system down. The concept behind the prefetcher is simple: It preloads information in anticipation of loading certain applications at a later point. It preloads/pages this information at some earlier point - applications don't magically load faster when prefetch is enabled, a portion of them has been preloaded into memory at some earlier stage. You are 100% wrong. Prefetching loads NOTHING before hand. It traces an applications startup and after 3 times creates a .PF file. From then on out when you go to load that application, the first thing Windows does is look in the Prefetch folder and if a .pf file exists it loads all the files that are referenced in the file at the same time. Open a .PF file and look at it yourself. All it is, is a list of file locations. By default during an application startup some of those files might wait on another one before they are loaded. With prefetching enabled they are all loaded simultaneously, thus applications load faster. If you don't believe me ask Microsoft. Your lack of knowledge on the subject makes this scarey that you offer people incorrect advice. I suggest you speak with Microsoft.

Mastertech Junior Member Posts: 2 Joined: 2005-04-13

#62768 Posted on: 10/04/2005 10:10 PM @ Mastertech said: "I've been doing this for over 15 years and work daily for a PC OEM." Well there you go, shot yourself in the foot there, 99.99% of people working for large OEMs know fook all about PC's. I've been woking with PCs (support + software Dev) for a lot longer than 15 years and everyone knows that. As have I but I have only worked for a PC OEM for 15 years. #1 it is not a "large" OEM. #2 None of which changes how Prefetching works. Show me ONE document to support how you think Prefetching works as I can provide many to show how it actually does.

PersianImmortal Unregistered

#62774 Posted on: 10/05/2005 03:20 AM I see. So the link you yourself provide in your guide, and is a Microsoft-authored link is wrong, but you are right? Your understanding of Prefetch is clearly wrong. Prefetch must preload data at some point as part of its optimization of application load times, and that is borne out by both Microsoft sources, and prominent MS experts. As you say, it is scary that you are offering tweaking advice. Your performance figures are derived from Microsoft comparing a stock standard XP install running prefetch. However a customized XP install as per my guide, which doesn't actually disable prefetch (the Prefetch key remains =3 in the registry), but removes the .pf files, and disables Task Scheduler to prevent their recreation, results in faster boot times and no noticeable decline in app launch times. Try this yourself before continually refuting it, it does work. Prefetch caters to the average user who doesn't mind the fact that MS has chosen to give them bootup load times in return for shorter app load times. I am saying I don't want that tradeoff, and in my actual experience (not regurgitating technical information incorrectly from MS sites), I find that my method results in faster bootup and no noticeable reduction in app load times. Your rude, abusive, incorrect ranting has shown that you are more interested in getting on the latest "myth busting" bandwagon, rather than finding new ways of getting better performance in games and XP - which is what my site is devoted to. None of my tweaks hurt performance or system stability, and I continue to stand by that.

Mastertech Junior Member Posts: 2 Joined: 2005-04-13

#62775 Posted on: 10/05/2005 04:21 AM I see. So the link you yourself provide in your guide, and is a Microsoft-authored link is wrong, but you are right? Yes. Your understanding of Prefetch is clearly wrong. Prefetch must preload data at some point as part of its optimization of application load times, and that is borne out by both Microsoft sources, and prominent MS experts. As you say, it is scary that you are offering tweaking advice. It is NOT MY understanding but how Prefetching works. It does preload the data right after you click on the icon to launch the application. What you are implying is ridiculous. Files for 128 applications are preloaded at windows startup? Get a clue. Your performance figures are derived from Microsoft comparing a stock standard XP install running prefetch. My performance figures are derived from testing with a stopwatch. However a customized XP install as per my guide, which doesn't actually disable prefetch (the Prefetch key remains =3 in the registry), but removes the .pf files, and disables Task Scheduler to prevent their recreation, results in faster boot times and no noticeable decline in app launch times. Try this yourself before continually refuting it, it does work. No it doesn't and you're a fool for believing it does. Go do some real research and test things before posting BS. People like you are what is wrong with the internet. You provide BS advice consitently with out verifying it. Prefetch caters to the average user who doesn't mind the fact that MS has chosen to give them bootup load times in return for shorter app load times. You are so clueless on this you have no idea. Try timing this instead of guessing fool. I am saying I don't want that tradeoff, and in my actual experience (not regurgitating technical information incorrectly from MS sites), I find that my method results in faster bootup and no noticeable reduction in app load times. Right.... Your method is what you get when people with no knowledge of how something works doesn't bother to test it and just posts assumptions. Your rude, abusive, incorrect ranting has shown that you are more interested in getting on the latest "myth busting" bandwagon, rather than finding new ways of getting better performance in games and XP - which is what my site is devoted to. None of my tweaks hurt performance or system stability, and I continue to stand by that. I am interested in providing ACCURATE information not snake oil. Don't be a fool and try testing something before you make idiotic claims. Try to stop being ignorant and do some research.

PersianImmortal Unregistered

#62776 Posted on: 10/05/2005 05:43 AM Thanks for the discussion, and let's just agree to disagree. I have little if any regard for your work now - far less than I did before. Of all the tweakers I can name, Thomas McGuire is perhaps one of the last few for whom I have high regard, particularly for his contributions to game tweaking, as well as assorted tech sites like HardOCP which explore new discoveries. Using alarmist terms like "crippling windows" - when I know for a fact that is not the case - and flip-flopping when MS' own documents reveal that prefetch must preload data at some point (I would suggest boot time seems logical based on my observations). Good going. For the record, I have never had any problems admitting I am wrong, adjusting my guides accordingly when new facts come to light, or testing/user feedback tells me things are not working. Many of the myths you speak of I have already updated since the start of the year in my guide indicating that they are useless (but harmless in any case). At the same time, unlike your ranting, I have no problems with my readers using other guides and deciding for themselves what works and what doesn't. Unlike you I don't label my guide 'incompatible with other guides' to try to wrangle traffic my way. I don't accept your guide as fact, but it seems harmless enough. It's not the way I do things, but that's fine. The proof is in the pudding as they say. Everyone who has used my TGTC seems to be happy with the newfound speed and stability, and indeed on a wide range of systems I have installed and used for myself and friends, it's the same case. Fast loading times, fast app launch times, better performance in games and rock-solid stability. Take care.

ChrisW Unregistered

#62781 Posted on: 10/05/2005 11:52 AM >>I see. So the link you yourself provide in your guide, and is a Microsoft-authored link is wrong, but you are right? >>Yes. LMAO, your arrogance is just a joke, a bit like your website. >>My performance figures are derived from testing with a stopwatch. I don't know where to even start, here. other than find a way to stop myself laughing. >>Go do some real research and test things before posting BS. People like you are what is wrong with the internet. You provide BS advice consitently with out verifying it. Yes kids, lets all do our "Research" with a stopwatch. >>You are so clueless on this you have no idea. Try timing this instead of guessing fool. Try looking in the mirror, as you've just described yourself perfectly. FOOL. >>I am interested in providing ACCURATE information not snake oil. Don't be a fool and try testing something before you make idiotic claims. Try to stop being ignorant and do some research. Accurate information ? Lmao !! You have Norton Antivirus, and Trend bloody micro, under your anti-virus recommendations. ANY savvy person will know that Kaspersky is the best. And what about Anti-Vir ? For Free, it Certainly beats anything you've mentioned on your shit site. Your whole "Tweak" website is just a complete joke. A whole page of sheer arrogance, rather than helpful advice. I would go back a start again. If your ego swells anymore, you're likely to explode at any time. Although, this can't be a bad thing. Now, how would you put it, FOOL !!

Warp2Search.net » News » October 2005 » CCleaner Cripples Application Load Times