page file config and imaging

[AsusUser]

There is a great thread with links, and a very nice pissing contest on here about page files, whta they do, what they don't, etc etc. I'm not looking to create the same pissing contest or arguments, my question is pretty basic.

My systems hard drives consist of this:

A seagate 160g Sata drive C: and D:

This has 2 partitions, a 30gig C: partition that I use pretty much exclusively for windows and installed applications and the remainder of the drive D: I use primarily for game installations, and general usage. (ie: stuff I do not need backed up in a ghost image)

A seagate 120g IDE drive E:
This is pretty much only used for storing more commonly used MP3's, and my filesharing downloads go here as well.

An external 500g drive
This obviously has no merit in my question.

If I go into the section for page file usage, by default the main partition with windows, has a paging file of a relevant size. The logical drive D: and the secondary drive E: have the page file disabled. This is the default for windows.

What I wonder is, since my games are installed on D: should I turn on a page file for that drive? Am I being silly for installing games onto a logical drive? Should I make the page file on C: something real small like 100MB, and turn the pagefile for D: on?

Have I drank too many rum and cokes tonight? WIll I laugh at myself for posting this in the morning?

I have two motivations for asking. I have a pretty strict regimene of imaging my install, and when possible I like to keep the image size down to where it will fit onto a 4.7GB DVD. This is why my games are on a logic drive. I've never received a virtual memory resource error in the last 3-4 years, largely because RAM amounts have gone up so much. (I currently have 1G of ram which is not stellar but way more than adequate, but I'm sure we all rememebr what it was like running with 64megs of ram on windows 98)

 

[Langford]

having a pagefile in two separate partitions of the same physical disk will slow your system down. Having a pagefile in one of the partions on each physical disk will speed your system up.

 

[AsusUser]

Makes sense to me.

What do you think about this scenario.

No page file on C:

The same size page file that I have on C: now, but instead placed on D:

Add a small page file to E:

My quandary is, what would happen if I imaged C: and had to boot to it? Would there be any......oddities? The image recovered would obviously not have a page file of any kind, I'd assume windows would make one, I could just make a point to change that option once I would recover off it.

Cheers.

 

[Langford]

Well, I suppose it would expect a pagefile to be wherever it was at the time of the image was made. If it gives you grief, you should be able to boot to safe-mode and fix it in jiffy.

 

[bob]

I wouldnt mind having some more rum and cookies.

In my opinion, I think it would be best to put the pagefile on the fastest hard drive out of the bunch, and possibly the one that is least used.

If you have 2 IDE channels, and channel A is for the windows hard-drive, and channel B has the drive for all of your commonly accessed files, I think it would make sense to stick the pagefile seperate on the fastest, least used drive.

 

[AsusUser]

I'm not sure how your reply pertains to my setup, but thanks for reading.

 

[djnes]

I'm not sure how your reply pertains to my setup, but thanks for reading.

You should, if you are creating a thread asking about pagefiles. What he's saying is this, since you have two physical hard drives, theoretically, you will get the best performance by moving your pagefile to the second physical drive...the one without XP installed. This assumes both drives are relatively even in terms of performance. Do not split the pagefile up over two drives as suggested. I've never seen anything positive happen with a setup like this.

 

[Langford]

You should, if you are creating a thread asking about pagefiles. What he's saying is this, since you have two physical hard drives, theoretically, you will get the best performance by moving your pagefile to the second physical drive...the one without XP installed. This assumes both drives are relatively even in terms of performance. Do not split the pagefile up over two drives as suggested. I've never seen anything positive happen with a setup like this.

When the pagefile is split between two phisical disk, Windows will use the disk that it detects is not in use. I use this setup on my media recording system to great advantage.

 

[djnes]

When the pagefile is split between two phisical disk, Windows will use the disk that it detects is not in use. I use this setup on my media recording system to great advantage.

Show some links to prove that's better. Pagefile fregmentation can lead to reduced efficiency...and by splitting the file in two, you are in a sense, creating fragmentation.

EDIT: I found one link suggesting it may help, but it gave no description of how or why. I'm disgregarding that source, because a description should be in order for tweaking tips.

 

[AsusUser]

Sigh, this thread is quickly turning into the trollfest just like the one I mentioned.

I'm sorry.

Why I even bothered, is beyond me.

(noone actually reads posts anyways)

 

[djnes]

I'm saying this as a question, not to be cocky, but have you read the sticky on this very subject?

 

[Langford]

Show some links to prove that's better. Pagefile fregmentation can lead to reduced efficiency...and by splitting the file in two, you are in a sense, creating fragmentation.

EDIT: I found one link suggesting it may help, but it gave no description of how or why. I'm disgregarding that source, because a description should be in order for tweaking tips.

Here you go: http://support.microsoft.com/?kbid=308417

 

[djnes]

That doesn't specifically say divide the pagefile. It says to divide virtual memory, more than a few people on here are quick to point out is not just the same thing as the pagefile. There's no description about it either, as in when it may be good, when it might not be, nor is it mentioned as to what size or percentages should be split. I'm questioning this concept because it's always been an accepted believe NOT to split up the pagefile, along with keeping it a static size. Neither point is mentioned in that article.

 

[ThreeDee]

well ..as already stated ..go with creating your pagefile on the best performing least used drive

if both drives are similar in performance , then go with the least used drive.(noticed that I said "drive' and not 'partition')

I run a 75gb Raptor (10,000rpm) and a 250gb secondary (7200rpm) ..tho my secondary is used less , my raptor out performs it significantly , so my pagefile is on my raptor and not the secondary drive.


[F]old|[H]ard

 

[AsusUser]

I'm saying this as a question, not to be cocky, but have you read the sticky on this very subject?

Yeah. The sticky is actually what made me ponder it in the first place, I just happened to read it because the last time I imaged my drive I made the file smaller so an image would fit onto a 4.7g DVD.

However, what I have always been unclear of, is the use of logical drives and pagefile resources, which is why I posted the thread. It is quite obvious that a person would be rather foolish to put a page file on a logical drive, but what if the application that was in need of UTILIZING the page file was indeed on a logical drive? My first thought would be that it would not be a preferred scenario, for exactly the same reason you wouldn't put the pagefile on the logical drive in the first place.

I probably answered my own question with that statement, in that I should alter how I use my setup in the first place. If I'm so worried about slowdown from a game or an application being on the logical drive and having to access the pagefile that is located on the other partition, I should really consider either moving the application to the main partition or at the very least move the page file to the IDE hard disk.

 

[Vette5885]

My quandary is, what would happen if I imaged C: and had to boot to it?

I did this in my system - it just boots without a PF. No problem really, as long as you don't run anything truly memory intensive. If you have 1GB or more, I don't think it would hurt at all (I know some people who actually turn off the PF if the system has over 1GB of memory - but then again, thats for office computers that typically run outlook as the most memory intesive app).

I do think that running the PF over two HDD's is a good idea, especially if you're a chronic multitasker. It isn't always faster, but it has the potential. I see it like this, in a "default configuration" with 2 drives and 2-disk PF:

with 1 app: Load app from, C; windows sees that C is in use and uses D for PF info. App runs with PF on D, which is good - speed boost.

with 2+ apps:
Load 2 apps, one then the other. the first app does the same as above, then the second app will follow suit - windows can then migrate the first app's PF info onto the first HDD as the app is used, and the second app uses the second HDD's PF (yes, this is speculation of a perfect environment, but it could and probably does happen).

The only thing that can go wrong is if you run say 4 apps and they all use the same PF. But then again, you'd be in the same situation as if you had the PF on one HDD. So I only see good with splitting the PF.

The only time that the dual HDD PF would be bad is if the drives run a different speeds. For example, my Maxtor is faster than my WD drive - so I'm testing out the PF only on the Maxtor (with the WD being the boot/app drive). Yes, i know that it would've been better to have the Maxtor as the primary, but time and HDD space limited me from clearing the Maxtor for a clean install.

 

[AsusUser]

Yeah, both of my internal hard drives are 7200 rpm's with 8meg cache, fairly standard drives. In the speed test under IDE/ATA/ATAPI drives, the speed test section the IDE drive does report that it is actually faster. 74 vs 81 megs per second. I'm not intimately familiar with that test, but it does lead me to believe that for whatever reason, my IDE drive is faster. Might have something to do with the configuration of the drive since it's a little bit smaller. I'm not worried about it however.

 

[Vette5885]

Yeah, both of my internal hard drives are 7200 rpm's with 8meg cache, fairly standard drives. In the speed test under IDE/ATA/ATAPI drives, the speed test section the IDE drive does report that it is actually faster. 74 vs 81 megs per second. I'm not intimately familiar with that test, but it does lead me to believe that for whatever reason, my IDE drive is faster. Might have something to do with the configuration of the drive since it's a little bit smaller. I'm not worried about it however.

I ran both NVidia's test as well as SiSoftware Sandra and found my WD drive to be faster when I had windows installed on my Maxtor. Now that the WD has windows on it, the speeds have swapped; Maxtor is faster. I'm not entirely sure as to why though.

 

[Phoenix86]

I have an idea... Why don't you TEST the various setups and see what works best for you.

That's the best advice I can give you w/o turning this into a discussion, or bitch-fest.

Langford, that link, at best, is a pie in the sky article. It doesn't describe when or why you should divide the PF. I can guarantee there are scenarios (common) where you should NOT divide the PF.

djnes, even MS uses the terms interchangeably (and they shouldn't), this is one. They are talking about the PF when they say to divide the VM space between drives. The only place VM is stored on drives is the PF.

 

[Langford]

I have an idea... Why don't you TEST the various setups and see what works best for you.

That's the best advice I can give you w/o turning this into a discussion, or bitch-fest.

Langford, that link, at best, is a pie in the sky article. It doesn't describe when or why you should divide the PF. I can guarantee there are scenarios (common) where you should NOT divide the PF.

djnes, even MS uses the terms interchangeably (and they shouldn't), this is one. They are talking about the PF when they say to divide the VM space between drives. The only place VM is stored on drives is the PF.

Mine is divided onto thier own partitions at the end of my two hard drives. It runs great. The article is only linked there to appese dingbats who request outside articles after ignoring hands on testimony.

 

[djnes]

Mine is divided onto thier own partitions at the end of my two hard drives. It runs great. The article is only linked there to appese dingbats who request outside articles after ignoring hands on testimony.

Wow...nice way to gain credibility, by namecalling. Furthermore, who are you, that everyone on the board should take your word as truth? So, if I posted comments that I run my computer underwater, that means it's true and everyone should do it? You really haven't been reading the internet very long, have you? The reason why people ask for links is for proof. Ever been in high school or college and been asked for sources on a paper? I can't wait to see the response Phoenix86 is preparing for you on the burden of proof.

 

[Fark_Maniac]

Mine is divided onto thier own partitions at the end of my two hard drives. It runs great. The article is only linked there to appese dingbats who request outside articles after ignoring hands on testimony.

Langford, I think the reason we "dingbats" ask for outside articles with hard numbers / reasons / facts is because some wahoos are out there that may splurge a bit on their results.

"Hands on experience" does not counter any placebo effects. I used to be a huge proponent to disable the page file. It felt like it flew! Then I ran some tests adjusting the size of the pagefiles from 0 to 1GB...the results were negligible. Later, I went to Barnes & Noble and read some of the Windows Reference books and all strongly discouraged the action.

I think this is what the dingbats are trying to prevent. Straight facts now prevent wahoos of the future.

 

[djnes]

"Hands on experience" does not counter any placebo effects. I used to be a huge proponent to disable the page file. It felt like it flew! Then I ran some tests adjusting the size of the pagefiles from 0 to 1GB...the results were negligible. Later, I went to Barnes & Noble and read some of the Windows Reference books and all strongly discouraged the action.

I think this is what the dingbats are trying to prevent. Straight facts now prevent wahoos of the future.

I used to swear by running with no pagefile, and on a RAID0 array. Now I wouldn't dare do either. I guess I'm just a dingbat for wanting to see proof, because I know better than to believe everything I see/read/hear.

 

[Langford]

Wow...nice way to gain credibility, by namecalling. Furthermore, who are you, that everyone on the board should take your word as truth? So, if I posted comments that I run my computer underwater, that means it's true and everyone should do it? You really haven't been reading the internet very long, have you? The reason why people ask for links is for proof. Ever been in high school or college and been asked for sources on a paper? I can't wait to see the response Phoenix86 is preparing for you on the burden of proof.

I appoligize at getting miffed over your attacks. You want my creditials, I've delt with computers since my old Vic20 in the 80s. Ive used DOS 2.0 to 6.22, Windows 1.01 to Windows XP and all versions in-between. I've delt with pagefile/swapfile since Windows 3.0. I've served 4 years in the Air Force providing all levels of technology support within out isolated squadron, which could not recieve support from the outside. I left to pursue college, and am currently enrolled as I work as the Technology Specialist for the local Title V Grant.

You probably don't belive it, but I ain't giving it to you in text. You woldn't belive then either.

 

[djnes]

I appoligize at getting miffed over your attacks. You want my creditials, I've delt with computers since my old Vic20 in the 80s. Ive used DOS 2.0 to 6.22, Windows 1.01 to Windows XP and all versions in-between. I've delt with pagefile/swapfile since Windows 3.0. I've served 4 years in the Air Force providing all levels of technology support within out isolated squadron, which could not recieve support from the outside. I left to pursue college, and am currently enrolled as I work as the Technology Specialist for the local Title V Grant.

You probably don't belive it, but I ain't giving it to you in text. You woldn't belive then either.

I don't care about your credentials, and I don't mean that in an attacking way. What you can type in on an internet forum can be 100% true, 100 BS, or somewhere in between. That doesn't mean a damn thing in the end, and I would think you'd be smart enough to grasp that....that links to respected sites are considered proof...not personal experience. I could post my background as well, but that doesn't mean my "personal experiences" should be taken as fact. I'm not even sure how to describe this any clearer, but you are missing the point myself, Phoenix86, and Fark Maniac are trying to make. It's not that any of us were calling you and idiot or a moron, it was that the way these forums work is, if you make a claim, back it up. That goes the same way as in college, and the work world as well. If you take a venture into the video card forum, you'll find a ton of threads and posts saying "Nvidia/ATI is better because that's what I use". You'll start to see why links are always requested.

 

[Langford]

I don't care about your credentials, and I don't mean that in an attacking way. What you can type in on an internet forum can be 100% true, 100 BS, or somewhere in between. That doesn't mean a damn thing in the end, and I would think you'd be smart enough to grasp that....that links to respected sites are considered proof...not personal experience. I could post my background as well, but that doesn't mean my "personal experiences" should be taken as fact. I'm not even sure how to describe this any clearer, but you are missing the point myself, Phoenix86, and Fark Maniac are trying to make. It's not that any of us were calling you and idiot or a moron, it was that the way these forums work is, if you make a claim, back it up. That goes the same way as in college, and the work world as well. If you take a venture into the video card forum, you'll find a ton of threads and posts saying "Nvidia/ATI is better because that's what I use". You'll start to see why links are always requested.

I knew you wouldn't care, thats why I gave it to you. I provided an anwser you were unable to give. You questioned it, so I gave written proof of that anwser from the manufactuer. You decide not to read it, or you couldn't understand it. I don't care what you believe kid.

 

[djnes]

I actually did read it, and I wasn't the only one to take issue with it...if we are going to discuss who read didn't read what. You're also losing any credibility based on your attitude and ASSumptions. By your logic, since I have a Bachelor's in IS, my word should be believed over yours....if personal qualifications are all that matters. You're insecure, thinking anyone that asks for links is disbelieving you, and now you're making ASSumptions into thinking I'm a kid. Nice.

 

[Langford]

I actually did read it, and I wasn't the only one to take issue with it...if we are going to discuss who read didn't read what. You're also losing any credibility based on your attitude and ASSumptions. By your logic, since I have a Bachelor's in IS, my word should be believed over yours....if personal qualifications are all that matters. You're insecure, thinking anyone that asks for links is disbelieving you, and now you're making ASSumptions into thinking I'm a kid. Nice.

You know what, I appoligize. You can have your opinion, thats fine. It's clear that the issue here is that we are both very opinionated. It bothered me to see the poster mislead by what I perceved as someone who seemed confused. I am sure you didn't not see yourself as having mislead anyone, and are likely well intentioned.

 

[Phoenix86]

Mine is divided onto thier own partitions at the end of my two hard drives. It runs great. The article is only linked there to appese dingbats who request outside articles after ignoring hands on testimony.

Oh, another one of these... types.

You know, people like you are the reason I post less here now (generally available by AIM/PMs for anyone).

If you have been in the industry this long, you shouldn't scoff at the idea of being asked for a link. Heck, I'm sure you probably went through college, so you should know that citing your sources is IMPORTANT. You should also know that sources, even experts, can be worng. One only need to read the section of the document you are sourcing.

You can optimize virtual memory use by dividing the space between multiple drives and by removing space from slow or heavily accessed drives. To best optimize your virtual memory space, divide it among as many physical hard drives as possible.

Umm, virtual memory is VIRTUAL, the HDD is PHYSICAL. You don't divide the virtual memory, you divide the page file. Should I link you to a MS doc that says set the PF at 2.5 times your RAM ammount? Cause I think we can all agree that's not "optimal".

None of that addresses the point you are making, but I think it should help put some perspective in your attitude.

That's great that you span the PF across multiple drives and it works for you. I don't think anyone's going to argue that it won't work. However...

Did you ever consider it's not the "best" setup? Did you test? How? What were the results? Why did you decide, based on those results to span the PF?

There is a time and place for spanning the PF, but I'm going to guess you don't fit the bill. Again, that's a guess. You running RAID? Do you have multiple controllers? How much RAM/commit charge peak (total memory used) do you have; are you even paging?

 

[djnes]

No one was misleading anyone. You made a claim that it's best to split up the pagefile. I've never heard of that, so I asked for a link. You offended because I simply didn't bow down and say, wow...Langford said it, so it must be true. I asked for a link, and you copped an attitude. Several other people either asked for a link, or disputed claims in the link...and yet you have a problem with me. Also notice, you're the only one who stooped to name calling, and then referred to me as the kid. The fact that we may or may not be opinionated has nothing to do with any of this. I asked for a link, and so did others, and you took that to be a personal attack. A small amount of logic and common sense would dictate that simple typing in something doesn't make it a fact. You won't survive in the business world if you think everyone's out to attack you, so I hope, for your sake, you aren't like that in real life. Chill out, and calm down. Not everyone is out to get you. We asked for links to describe your points, that's all.

 

[Langford]

Oh, another one of these... types.

You know, people like you are the reason I post less here now (generally available by AIM/PMs for anyone).

If you have been in the industry this long, you shouldn't scoff at the idea of being asked for a link. Heck, I'm sure you probably went through college, so you should know that citing your sources is IMPORTANT. You should also know that sources, even experts, can be worng. One only need to read the section of the document you are sourcing.

You can optimize virtual memory use by dividing the space between multiple drives and by removing space from slow or heavily accessed drives. To best optimize your virtual memory space, divide it among as many physical hard drives as possible.

Umm, virtual memory is VIRTUAL, the HDD is PHYSICAL. You don't divide the virtual memory, you divide the page file. Should I link you to a MS doc that says set the PF at 2.5 times your RAM ammount? Cause I think we can all agree that's not "optimal".

None of that addresses the point you are making, but I think it should help put some perspective in your attitude.

That's great that you span the PF across multiple drives and it works for you. I don't think anyone's going to argue that it won't work. However...

Did you ever consider it's not the "best" setup? Did you test? How? What were the results? Why did you decide, based on those results to span the PF?

There is a time and place for spanning the PF, but I'm going to guess you don't fit the bill. Again, that's a guess. You running RAID? Do you have multiple controllers? How much RAM/commit charge peak (total memory used) do you have; are you even paging?

Say what you want about giving opinions. I'm still the only one in th this thread so far who has linked an outside source.

 

[Phoenix86]

Say what you want about giving opinions. I'm still the only one in th this thread so far who has linked an outside source.

I'm questioning your opinion. I asked:

Did you ever consider it's not the "best" setup? Did you test? How? What were the results? Why did you decide, based on those results to span the PF?

There is a time and place for spanning the PF, but I'm going to guess you don't fit the bill. Again, that's a guess. You running RAID? Do you have multiple controllers? How much RAM/commit charge peak (total memory used) do you have; are you even paging?

Those were not rhetorical questions.

You want a link? Here's a better one.

A frequently asked question is how big should I make the pagefile? There is no single answer to this question, because it depends on the amount of installed RAM and how much virtual memory that workload requires. If there is no other information available, the normal recommendation of 1.5 times the amount of RAM in the computer is a good place to start. On server systems, a common objective is to have enough RAM so that there is never a shortage and the pagefile is essentially, not used. On these systems, having a really large pagefile may serve no useful purpose. On the other hand, disk space is usually plentiful, so having a large pagefile (e.g. 1.5 times the installed RAM) does not cause a problem and eliminates the need to fuss over how large to make it.

edit:

AsusUser, this is the BEST recommendation ANYONE can make. Have more RAM than commit charge peak (total memory usage, found in task manager/performance). Then you can pretty much forget the page file's setting 'cause it's not being used. End the debate by making the file useless.

 

[Langford]

I'm questioning your opinion. I asked:

Did you ever consider it's not the "best" setup? Did you test? How? What were the results? Why did you decide, based on those results to span the PF?

There is a time and place for spanning the PF, but I'm going to guess you don't fit the bill. Again, that's a guess. You running RAID? Do you have multiple controllers? How much RAM/commit charge peak (total memory used) do you have; are you even paging?

Those were not rhetorical questions.

You want a link? Here's a better one.

A frequently asked question is how big should I make the pagefile? There is no single answer to this question, because it depends on the amount of installed RAM and how much virtual memory that workload requires. If there is no other information available, the normal recommendation of 1.5 times the amount of RAM in the computer is a good place to start. On server systems, a common objective is to have enough RAM so that there is never a shortage and the pagefile is essentially, not used. On these systems, having a really large pagefile may serve no useful purpose. On the other hand, disk space is usually plentiful, so having a large pagefile (e.g. 1.5 times the installed RAM) does not cause a problem and eliminates the need to fuss over how large to make it.

Your source also agrees with what I said in post #2 and #12. When the help in #2 was questioned in post #9, I gave a source in post #12 that explained why. When insulted in #19, I gave a personal account in #20. I shouldn't have poked at djnes in #20, but felt personally insulted that the both of you were attacking over an article.

 

[AsusUser]

Thanks for the replies everyone.

Having said that, I would lock this thread if I could.

 

[Phoenix86]

Langford, thanks for answering my questions... Tells me everything I need to know about you and your experience.

AsusUser, PM a mod to lock the thread.

edit: LOL, you expect me to read what you typed when I can't bother you to respond to questions posted twice. Good luck with that and all.

 

[djnes]

Thanks for the replies everyone.

Having said that, I would lock this thread if I could.

You can, but going under the thread tools menu when viewing this thread. You are the one who created the thread, so you can do it. I would agree with your choice of locking the thread, because it was derailed, unfortunately by people simply asking for a link.