Windows Prefetching Reference

[Mastertech]

gdogg honestly you are the only one who has not tried testing this properly or working with me and I wonder why that is? Ibis clearly found he had no layout.ini file and no disk defragmenter installed. Whenever you want to get serious about this let me know.

All the Myths on the site are accurate, the one you pointed out is now worded more clearly. You are welcome to provide proof for anything you want.

I am begining to think you have something to do with CCleaner.

[gdogg]

your awfully parinoid

but what amazes me is you think i have something to do with crap cleaner. when i have hundreds of posts in these forums, its would be like me thinking you have something to do with M$ cause your new, are saying something that cleanly didn't improve boot times for a huge amount of us nlite users, not just me. honestly man, your myth isn't always 100% true, now, go reword it, to say, on my computer this myth is true, in these aricals i found this myth is true, but for some, the myth might just be my website.

and yet, you think my tests a year ago, and yesterday are all b$, come on man, get real, you havn't provided a single piece of evidence to back up your claims.

where are screeshots of bootvis plz.

and about the myth, it was completly bogus, and contradicted itself, til i told you about it.

not only that, but the way you now worded it, is so obvious in the first place

increasing system cache, doesn't improve workstation peroformance , DUH.

but b4 it said, increasing system cache, doesn't improve filesystem performance, but yes it did, and you even went on to say it did.

Edited October 14, 2005 by gdogg

[I like cleaning my prefetch]

I love this. Mastertech gets shot down again! Whats going to happen? I predict he will post the same thing he posted at the start of the topic. Thats how it goes. Don't even bother trying to sort anything out with the great "Mastertech" it get you nowhere but p***ed off.

[gdogg]

oh and one more thing

what website claims that increase file system cache improves workstation performance?

you should post links to sites, where these supposed myths exist.

you know, i got a new myth for your site

myth - suspend mode decreases boot times

fact - suspend mode in no which way or form will decrease boot times. If you fell for this myth, you were just gullible.

Why you might ask? because suspend is not booting, it is suspending and resuming.

Edited October 14, 2005 by gdogg

[Mastertech]

gdogg, you seem obsessed with not trying. How come with a few posts we found something out very critical about Ibis's setup yet you refuse to try? I find that very telling. Whenever you are ready to get serious let me know. Until then I will work with those that are. Interesting isn't it that the two people making all the outrageous claims haven't properly tested anything? At this point gdogg I cannot take anything you say with any credibility. You will have to rely on other members here to prove or disprove anything. I have a feeling quite a few nLite users have a broken prefetch.

increasing system cache, doesn't improve workstation peroformance , DUH.

but b4 it said, increasing system cache, doesn't improve filesystem performance, but yes it did, and you even went on to say it did.

Try stepping back for a minute (if that is possible). The myth page, especially the optimization myth section title "XP Myths" relate to #1 Myths Regarding Windows XP. Now no matter how the large system cache tweak myth was worded, the intent was to show that using that tweak on an XP installation will not improve performance. You can clearly see it will only improve a file servers performance. That is quoted directly from Microsoft. If you want to stay hung up on the wording then by all means knock yourself out. Either way it is clearer now.

Plenty of tweak sites/forums mention the LargeSystemCache tweak, they use there own wording of course and recommend enabling it for an XP installation.

[Alanoll]

I've just disabled posting on Mastertech and gdogg for their last few posts in this thread. I was not joking in my post before.

Keep it clean, or don't post.

[Kelsenellenelvian]

Hey Thanks Allanol!!

[Fridge-RaideR]

Yeh but as I said, under the 200mb it aint nescesary, my firefox launches in 1sec. MSN and other similair programs etc all very fast as well 0.5-2sec. But I just have 8 serviceses on, u got like 15+. So if u have many stuff, it helps, if u got less stuff, ur better of without. I think I dont have to test this anymore, wot I will test is game start up performance. But wots that 2sec longer waiting for a game, while the level loading in bf2 for example takes 30sec.. then that 2sec is nothing.

Under 200MB what? :S ISO? Mine's at ~135 or some such But I understand what you mean and I think you're right. The less you have to load in the bootup and the less programs you use after, the less Prefetch does for you. And yeah, I need those services started, or else I loose functionality with some things, I'm not after the smallest Windows possible, but the smallest I can get without loosing any functionality I can't be without. I could probably kill a few more, but I haven't really looked into it (got any tips you can share by just looking at what I have there?) I think my gaming system is more up to your standard and there I haven't planned on using Prefetch either, but for my workstation OS, I couldn't live without it. I do so much in that, as opposed to the Gaming OS where I only do one thing at a time.

Installed 200mb, My iso was around 110mb

[suryad]

Nice dude! My most shrunk down XP iso was around 150 mb. The recent one I had to create because of the Ryan pack not allowing me to update was around 200 megs...yes it is kind of bloated but I went a bit easy on stripping out stuff from XP like OUtlook Express just in case I will need it in the future when I install Office 2k3 on my comp.

[Ibis]

Is there anyone with a Windows installation that has the full usability of Prefetch still in it and could test so we get something to look at? My test are worthless, and so i gdoggs if he tested on a computer either without the standard disk defragmenter or the Task Scheduler service. Otherwise it's good I guess, but I don't think we can be sure and draw any conclusion until some more tests is done to see exactly what is needed for the full Prefetch. I'll start playing around in VPC and I'll post my results here. I'll begin with my current installation, just not remove the Defragmenter and see what happens.

[kegobeer]

My nLite installation has everything needed to do some good tests. I haven't removed any crucial components or critial apps. I verified my prefetch directory exists and has all necessary files. I hope to find some time this weekend; by then Mastertech should have posting enabled again.

Edited October 14, 2005 by kegobeer

[Bold_Fortune]

Is there anyone with a Windows installation that has the full usability of Prefetch still in it and could test so we get something to look at?

Ibis, this is partly for you, partly out of my own curiosity.

I do not use nLite, as may know. I do however delete many files from my XP installation. My Window directory is 185MB.

I also keep a full, XP installation image stored on another partition for testing purposes and comparisons. And I keep this image up-to-date same as I do the other.

Each installation is identical to the other. Each has the same programs installed. Each has the same settings and tweaks applied. Each is kept up with the latest drivers and Windows Critical Updates. Everything installed is exactly the same as the other. Every change made is exactly the same as the other.

There are only two exceptions to this:

1) Services settings

2) Files removed

I cannot run bootvis on my 185MB image. I have COM+ disabled and all it's associated files deleted.

I can run bootvis on my full installation, as COM+ is enabled and none of its files have been deleted. (Also, if it matters, Windows Defragmenter is enabled on this installation; where it is removed on the other.)

Here are the Services settings I have on the full-installation image:

http://www.bold-fortune.com/forums/index.p...&st=0#entry1912

And these are the only files I removed from the full-installation image:

http://www.bold-fortune.com/forums/index.php?showtopic=387

To figure out everything I have deleted on the 185MB installation...well, you'd kind of have to read my entire guide.

However, I can show you which Services I have enabled, disabled, and deleted, here:

http://www.bold-fortune.com/forums/index.php?showtopic=241

On each installation, my reboot time is a consistent 31 seconds (given an oddball, occasional 32 second reboot on each). I have always kept a curious check on this. (I was excited recently to discover that if I set Acronis Scheduler2 Service to manual, and not allow Acronis True Image Monitor to startup with my system, that it shaved 2 seconds off my reboot time.)

One more thing, I use PerfectDisk Defragmenter on both installations to manage my boot files, and I have always had Prefetch enabled (set to "3" in the registry) on both installations.

Now that we have those reference points established, let me show you my results.

With my full installation (reboot time being a consistent 31 seconds) when I run bootvis, my consistent reboot time remains at 31 seconds. If I then, set my Prefetch to "0" in the registry (disabled), my reboot time gains 2 seconds, making it now 33 seconds. Now, if I then keep that "0" setting, and purge the Prefetch folder of all it's files, and defragment using PerfectDisk, my reboot time goes down a second, to 32 seconds. So I gained 1 second on my reboot time on my full installation.

As I said, I cannot run bootvis on my 185MB installation. However, when I disable the Prefetch in the 185MB installation, my 31 second reboot time gains 1 second, making it 32 seconds. Then, leaving Prefetch disabled, and now purging the Prefetch folder of all its files, defragmenting using PerfectDisk, my reboot time reverts once again to its 31 seconds.

You know something? I really learned something with all this.

By having the Prefetch disabled, and the Prefetch folder purged of all its files, on my 185MB installation, there is absolutely no difference in my reboot times. On the other hand, should I do the same for my full installation image, it adds 1 second to my reboot time.

[Tarun]

I've just disabled posting on Mastertech and gdogg for their last few posts in this thread. I was not joking in my post before.

Keep it clean, or don't post.

Thank you.

That allows people to do their own research without biased opinions.

Also people, you don't NEED to reboot to get prefetch to start caching. Just have task Scheduler on and go to Start > Run and enter this code.

%windir%\system32\Rundll32.exe advapi32.dll,ProcessIdleTasks

Works like a charm.

I've also tested Prefetch on my system and others. It really doesn't make any difference except for a few nanoseconds.

[Zxian]

Thank you Bold! It's really awesome that you took the time to test this out.

I think that for most people (not the minority of us who use nLite), leaving the prefetcher alone is best. I'm really skeptical about the people who said that having the prefetcher on adds 60% to the boot time... On full installations of Home, Pro, MCE, Server 2003, etc etc etc I've always seen an improvement with the prefetcher on and the Prefetch folder left alone.

[Bold_Fortune]

Thank you Bold! It's really awesome that you took the time to test this out.

You're very welcome, Zxian.

Matter-of-fact, I'm really glad about all this tit-for-tat that's been going on, because I'm so happy with my results, I've included my findings in PART 5 of my guide...and I am no longer using Prefetch.

Learning is what it's all about. Shoving it down the throats of others really isn't a very good way to go about things...at least in my opinion.

All you can really say is, "This is what works for me. These are the tests I've performed. ...But you do what you think is best for you." ...And that's just one man's opinion.