NBR Vista Tips and Tweaks Guide

I never said it will lead to malfunctions I said it CAN. There is no need to personally experience these malfunctions to understand that they can exist. All you have to do is understand the way virtual memory works to know the POSSIBILITY for data loss is real. I did not seek out sources that support my viewpoint, I sought sources who UNDERSTAND the underlying functionality of the virtual memory systems in Windows. Please Les, don't pretend you or anyone else who recommends removing the pagefile know more about this subject than Mark Russinovich. Your characterization of it as being some antiquated, arcane function is simply ridiculous. How easy do you think it would be for the OS to see that you have 8 gig of memory and just turn off the pagefile if that was the proper thing to do?

But lets get to the "end of the day" part of the discussion. At the end of the day you still have not produced a single scintilla of PROOF of the alleged performance increase. No measurements, no benchmarks, nothing. Nada. Zilch. Where's the proof? Where are the statistics to back up the claim? We have zhaden chiming in that " I have no pagefile and I must say overall response feels improved upon. Maybe just barely, but I'm pretty sure it's there." WOW! Not that is a real scientific testimonial if I ever saw one.

Sorry Les, I told you before, I will be here to debunk this myth every time it surfaces and I am being true to my word. We can agree to disagree, but I won't agree to stop asking for PROOF.

Gary

 

Gary...proof is in the mechanics of it all...

If u have to rely on a hard drive rather than ram to be a temporary storage location for information (as ram is), it is a given that the HD will be slower. I don't think there is any way to get around that in any for of thinking. A hard drive is slower than ram and I don't think there can be any dispute otherwise.

The simple mecanics of how information is picked up from a hard drive compared to that of ram is simple, not counting the fact that the hard drive has to consider multiple tasks in gathering the information. I won't bore you with how the disk spins at such and such a speed and how the arm must locate and pick up information from a certain address, after waiting for it to be at the exact point or how defragmentation slows this considerably on all hard drives.

With respect to the malfunctions comment, I wasn't quoting yourself, but rather, another poster. I, as well haven't ever seen a scintilla of proof of this given a system has pagefile turned off on the fresh install and ram is able to fulfill its function through adequate size.

So yes we can agree to disagree with one side having tried and true first hand experience and the other having.....articles of belief.

eheheh Have a good one my friend. We have both been at this for a long time now.

 

Your understanding of the mechanics of virtual memory is flawed. That is PRECISELY why I included Mark Russinovich's blog entry in my list of citations. But just for the sake of argument, let's assume your explanation of the mechanics was correct, where are the NUMBERS to back up the theory? Sure if you accept your mechanics, it does appear it should be faster. But the only way to know is to measure it.

The trouble is, no one has produced any numbers to support your theory. All we get is anecdotal "evidence" of the sort that zhaden provided with "I must say overall response feels improved upon. Maybe just barely, but I'm pretty sure it's there." And why is it that no one has ever provided any numbers to back up this explanation of the mechanics? Simple, the explanation is wrong.

Please read Mark's blog. He fully understands how virtual memory ACTUALLY works.

As for "one side having tried and true first hand experience" that "experience" has yet to produce any measurement of the alleged performance gains. Without such proof, I have to rely on the folks who are actual experts in this matter, namely Mark. Again, if you really think that the "first hand experience" trumps Mark's understanding that is just laughable.

Gary

 

So much anger >.< Sorry I spoke in the first place, to be honest. It isn't a matter of being laughable or making me look ignorant (which, frankly, is relatively silly on your part -- it's the internet), it's just a matter of preference. I don't need to waste any harddrive space on excess memory that is, what... well, vastly (~15ms at best vs 70ns at worst) slower at accessing data than my RAM as I have yet to even get close to capping it out.

edit below:

I am of the PERSONAL OPINION that not taking a large chunk out of the middle of your hdd for a pagefile and; thus, allowing Windows to more easily place files in sequential order at the beginning of the drive; allows for better hdd performance. Maybe my explanation is wrong, but the only thing you've done is say someone else said otherwise.

I have acquired my opinion over a long course of computer usage, and the pagefile in XP is definitely a drag-me-down with numbers to prove it all over the internet. I can only assume the same is true in Vista when you have enough memory to make up the difference the pagefile would've made.

 

In simplest terms, the page file is not an extension of RAM. It is a backing of RAM. Disabling the page file will result in higher RAM usage, pushing it closer to the system's limits, which is whatever amount you have installed. For example, the default user-mode stack allocation for threads is 1MB. This is an instant commitment to memory in the absence of the page file, which on my system, presently with 570 threads allocated, results in 570MB RAM usage on threads alone. Consider what else is occupying memory, too. This increase in RAM usage actually makes disk I/O more likely when less RAM is available for file caching, which is a performance degradation, not improvement. The space argument is moot as well, with disk space being so undeniably cheap; 8-10 cents on the gigabyte is hardly worth choking over.

Analyzing system behavior, which includes resource usages, on a case-by-case basis and making a sound decision as whether or not a page file should be present, even if it prevents the dumper from working correctly in the event of something going wrong on the system, is infinitely more responsible than Les' approach: blindly advising users to turn off their page files without any knowledge of their systems' application or behavior. I have a system with 12GB of memory and a page file is still present, with no performance degradation.

 

I figured simplest terms would've been what the operating system itself tells me...



With 42 tabs amongst 3 firefox processes, bittorrent, a chat client (logged in to 5 services), steam, and cs:s running, I use ~4200mb. I don't know how much of that is truly underutilized as a result of a minimum allocation per "thread", but it's more than I realistically use under a less streamlined OS than I will be using come Oct. 22.

 

You need to dig deeper than that, obviously.

 

I never said you were or what you said was laughable. What I said was laughable was anyone taking the anecdotal evidence of "it feels faster" over the opinions and understanding of Mark Russinovich.

As for not taking out a chunk of space for the pagefile making for, as you put it, "better HDD performance'. Sorry to tell you, but you are wrong. It has no effect. You need to read up on NTFS and fragmentation.

When you assume that anything about Vista's memory management by basing your assumptions on allegations about XP, again you are wrong. Vista does not handle memory in the same way a XP. Any assumption of similarity is flawed.

Finally, please don't attempt to play "a long course of computer usage" as some sort of trump card. I have been at this since I was in high school in 1968. I owned and operated the first computer store in 1976 when microcomputers were sold as a bag of parts. I have been a software and hardware consultant since 1980.

Look the whole point of this discussion is to try to prove or debunk a myth that having no pagefile makes a machine faster. If you can show real measurements that prove the myth, please do so. If not, it remains an unproven urban legend.

Gary

 

After reading through Mark's blog, I must admit he makes very valid points. And is much less aggressive about it. I didn't realize I would need to defend my personal worth with every response I got in this thread, so I think I'll stick with my initial feelings after this and just stop arguing as everybody else on this end of the debate has done. I will say that when 1024mb was an excessive amount of memory and my pc-133 was scorching in my system with no pagefile, I've never had one issue. However, after reading the very convincing evidence and thorough documentation done by Mark, I will retract my initial impression and enable a small pagefile to reduce excessive read/writes to RAM. We are, however, still lacking actual numbers in regards to performance. How do you feel this would best be benchmarked? HDTune with and without a pagefile? HDTune + ORTHOS at the same time?

I didn't realize stating I had a background in thorough computer usage for a long time running was "playing a trump card", but you win this one regardless. That's all fine and well, but don't attack me then turn around and pull the consultant card Same story here, barring the computer store and a few years.

 

So I did it myself. There is a conclusive difference in HDTune depending on your page file parameters. I ran each of these tests using the following methodology: 1) set page file parameters 2) restart 3) close steam and utorrent 4) run hdtune.

CPU usage and maximum/burst rate are minimally affected, but the dropoffs in the graph itself are much less substantial with the larger pagefile, resulting in a more consistant data transfer rate. It is possible to conclude that you will see a small decrease (good) in access time with a smaller pagefile (and the additional benefit of increased harddrive space - this is something the user must decide is important), with an overall lower transfer rate.

No pagefile:



1GB PF:



5GB PF:

 

Les,


Thanks for the effort !
I was able to gain 10 GB on my small 128 GB SSD and overall faster performance. ( I didn't touch the page file settings....yet )

 

I has the same issue...

 

I added this little program , it cleans out memory every 30 minutes. This one is different from the other ones...

http://www.softpedia.com/get/Tweak/Memory-Tweak/CleanMem.shtml

Before I had this program, Vista started with around 1 GB in Ram usage, after a full day of computing , I ended with around 2 GB ( I have 3 GB Ram) , ending in a sluggish way my day..

Now it starts out with 800 MB and is at the end of the day around 900 MB

 

that looks interesting... will it delete anything important? Never used one of these so not sure. And since Avira 64 bit doesn't protect it's process. As in when you try to terminate it it just does. Would it delete that too?

 

This is what it says :



"The program will install and set it to run every 30 min via the windows task scheduler

The program will install and set it to run every 30 min via the windows task scheduler. Install it and your done! The program doesn't run in the back ground, once it does its job it closes and doesn't run again until the task scheduler runs it.

The CleanMem application was designed to be a very simple to operate as it will run and clean the memory out of all processes it can, without any user input.

Some anti viruses and such protect their processes so of course cleanmem can't touch those.

This doesn't work like other memory cleaners that do nothing but force windows to free up memory by using up all the avail. memory. This old trick then causes your system to lag big time! "





I did run Avira over the different files and all was cleared.

Maybe some of the experts can comment if there could be any other risks...

 

So Avira was cleared?

 

I went to the Cleanmem folder and right-clicked on each file for a scan by Avira > Cleanmem was cleared,not Avira

Ok, found this program in the "Free Windows program list" sticky

"Tool that gives you back memory that has been used by programs. To elaborate, it does not force open programs to give back the memory it is using; the memory that closed programs have used are not totally given back to the computer to use, so Cleanmem can be used to do that"

 

@ Bassflow, Laptopaddict:

http://forum.notebookreview.com/showpost.php?p=5222561&postcount=619

There is another good utility, Instant Memory Cleaner (see above).

 

But so will Windows memory management. So what is the point of having this app? As soon as any app NEEDS this memory, the built in management functions will instantly make this memory available. I don't understand this penchant for running with as much free memory as possible. It makes no sense. Unused memory is wasted memory. The Vista internals work on this principle by leaving things in memory ready to be used or just as readily freed up if the memory is needed. It seems like this tool is superfluous, really.

Gary

 

I have to agree on this though.

Still, for those who want / prefer / would like to try an outside tool to do this, the two mentioned above are pretty good I guess.

 

Well, yeah, but when you want to have as much RAM as possible when you're video editing, wouldn't it be better to use this?

 

Not really. These applications can't release any memory that Vista's built in memory manager won't automatically release as soon as your video editor software requests more memory.

In reality they might negatively affect your editor when the cleanup app kicks in every so often and tries to recover unused memory.

Gary

 

So, basically, these things are useless.

 

Pretty much sums it up. These types of apps have been around for a very long time. And every technical review I have ever seen of them has said just about what you just said. There was a time where memory could get "orphaned", that is allocated by an application and not released when the app died or terminated prematurely. Memory cleaners did help with that. But that situation no longer exists. The memory management built into the OS keeps track of such things itself.

Gary

 

Gary,


Well, I start the day with 1 GB ram and end up with 2 GB at the end of the day without the application.
With this application, I start with 800 MB and rarely end with more than 900 MB

Does this application takes memory from applications that are still running, how do explain the huge difference ?

 

Just thinking Gary - what about applications with a memory leak?

Wouldn't in such a special case the RAM cleaning application be a good thing?

 

The difference is in the way Vista manages memory. It will leave things loaded but marked as available for immediate unload should the need arrise. Previous versions of Windows just unloaded stuff right away. Vista takes the approach that there is no need to do the unload UNTIL the memory is actually needed for use. That makes Vista APPEAR to be using more memory, when in reality it is using no more or no less than previous versions. It is just handling it differently.

Gary

 

I don't think so. First, my understanding is that any managed app that requests memory will have all that memory put back in the pool when the app expires for whatever reason. I readily admit my understanding of this is a bit sketchy with Vista, the last time I wrote any assembler code where I actually HAD to know how all that worked was under WinNT, ancient history. Now all my code is higher level where such memory mangement is all automatic.

But for the sake of argument, let's assume there was some leak that fell outside the bounds of that. How would these third party apps even know? Unless they are hooked into the memory allocation process in the first place and record their own info about each allocation request, they would have to rely on the underlying OS own set of tables about which apps have what memory. So they would not be able to know what is leaked vs unleaked memory anyway.

Gary

 

But what about an application that isn't closed??

KIS 2010 would be an example...

From what I've seen, I believe Photoshop can force Vista to give it all the memory it needs, and thn after closing Photoshop it seems to fall to normal levels - but KIS remains open.

Else KIS just seems to slowly eat RAM - especialy when idling though...

 

Forgive me if I am wrong, but isn't this the same as what Gary is talking about? When idling, KIS takes memory, BUT when an app like photoshop needs memory, Vista anyway gives it all that it needs. So, I am wondering if Gary was saying the same thing?!! Or am I missing something here?

 

Maybe you're right... but once when I had my laptop idling overnight, in the morning it was using 1,88GB of RAM (task manager) and was sluggish like I don't know what... result of the memory leak...

 

Hmm.. I understand. But then, lets say when you start PS in the morning, doesn't Vista give it what it needs and not remain sluggish? Or is the memory still being lacked for PS to start quick and fast?

 

I don't know - I'm not exactly sure what I wanted to do, but opening stuff was near imposible...

I suppose I could try if its important...

(anyway, heading off now ... )

 

I am not following you. What is KIS 2010?

Gary

 

Detlev meant Kapersky Security Suite - I think.

cheers ...

 

Same here, I start the day with 1 GB Ram in windows vista, I surf all day ( around 100 windows open at the end of the day) and a java application for charting of financial markets ( all day quotes coming in )
At the end of the day Ram is around 2 GB AND SLUGGISH

With CleanMem start Ram = end Ram AND not sluggish

How do you explain that if that application doesn't work ?!

 

Yes, KIS = Kaspersky Internet Security - thanks qhn.


Well, you may have a memory leak too.

Without the memory leak I end up with roughly the same RAM usage after and before I used applications.

 

Hey, new here.

What happens if I do all this stuff, but then upgrade over to Windows 7? Will it all be for nothing? I'm asking because my laptop I'm ordering is going to come with an upgrade to Windows 7 free deal and I just want to ask if it's a good idea.

 

Not recommended.

And honestly, its not needed.

Vista runs fine without tweaking.
(Shutting down junk from Software you added isn't tweaking - its cleaning up)

Because every OS is ever so slightly different, don't try to migrate is.

Yes, WIn7 is based on Vista, uses in some cases the same core components - but its not a carbon copy.

 

100% agreed with you.
I don't like the upgrade feature by Windows 7.
It sucks. Just get yourself a clean installation is much stable and better.

Example of Tweaking OS,
Turn on/off, disabling/enabling Windows Features and Windows Services as well as Regedit.

 

Just thought I'd mention a problem I noticed with a tweak mentioned at the start of this thread - disabling GUI boot really does nothing for performance BUT it does interfere with the ability of some applications to interact with you.

The two examples I've come across were checkdisc, which asked to run at next boot, so on attempting to run it could display no information about its progress (which simply made me think my laptop was stuck with a blank screen when it booted, and each time I rebooted it, same problem - blank screen.) The other time was using Acronis to do a disc clone - it requires the GUI boot option to inform you when the process is completed, otherwise you never see that message and are left guessing as to whether it has finished or not. My advice - do NOT disable GUI boot, you never know what application might need it in future.

 

My new laptop should be here tomorrow, ups....I cant wait to do some of these tweaks, it's the first thing i'm going to do after the prelimenary, register, get online....stuff.
If there is any of these you recommend I dont do because of any known probs that developed,,,pls let me know.
I'm going for 1,2,3,6,12, and 22...

thanks a trillion...
joe

 

I have a black screen for about 10 seconds everytime I start vista.... why could this be? I didnt even do all the tweaks here, just some. But I cant remember which ones exactly. Thats why I didnt do them all, because I knew I wouldnt remember them all

 

Does the screnn pyhsically switch off?
(then its linked to multi display support)

If its an actual black screen (with screnn on) then it can be any number of reasons.

 

some (if not all) of these tweaks can be applied to windows 7, right? I want to get rid of indexing and some windows components that i definitely dont need. thoughts?

 

Well, I think I said it above.
While they work for Vista, and Win 7 is built pon Vista it may possibly work.

However its recommended not to apply "tweaks" of an older OS to a newerone as you may break stuff.

And honestly - what do you expec to gain?
1s at startup?

Indexing will only use up resources once, once its done it will do so no more.
And Windows Components - you won't get a performance increase that way.

In most cases "tweaking" Vista/Win 7 is more akin to "Breaking the OS".

 

Are These Safe To Be Used On 64Bit Machines??

 

Yes. I believe Les has (had?) Vista Ultimate 64-bit on his computer that he used to test these tweaks out on.

 

I do use these Tweaks though on my Windows Vista Ultimate 64bit.
I do sometime more extreme though. LOL.
Some of these Tweaks even works on Windows 7 but not all.

 

thank you very much, i will try this out