I use my system for heavy gaming only, and just put in two 1GB ValueSelect sticks this afternoon. If possible, can anyone give me an estimate of the pagefile size I should use now?
What size page file w/2GB RAM?
- Thread starter xTYBALTx
- Start date
No matter how much RAM you have, you'll still need a pagefile, but it won't need to be as big most likely.
I'd say have it around 1GB, but make it 1GB min and max, that way it doesn't dynamically change sizes while doing heavy tasks - that could slow down performance.
I'd say have it around 1GB, but make it 1GB min and max, that way it doesn't dynamically change sizes while doing heavy tasks - that could slow down performance.
Although you are probably right about this belonging in the OS forum, I have to ask how you manage to make your second conclusion? Any references?
In the mean time, http://blogs.msdn.com/larryosterman/archive/2004/05/05/126532.aspx might be of interest to the original poster.
There is no reason to set a static pagefile at all. All it does is remove that safety net if it does come to a time where it does need to expand.
Because you gain nothing from messing with the pagefile.
Really? I was under the impression that changing the page file size does more to hinder performance than simply filling in the existing static page file. I thought it had something to do with pagefile fragmentation as well. I'll assume you know more about this than me, but can you explain all this at least so we know in the future?
A gradually growing pagefile would end up as a fragmented file. As I recall, it is quite an interesting exercise to defragment the pagefile... (easiest way I guess is to have more than one partition/disk and delete the pagefile from one of the drives and then reboot -- I don't know about you, but I don't reboot very often)
That isn't a reference, that's just you guessing.
In my case Windows' recommended size is 6GB. I recently bought a 140GB SCSI drive, so I do have some spare space at the moment, but I'd rather waste 6GB disk space on something else all the same. (and I'm always contemplating adding more memory, hence bumping up the recommended size again)
If you think messing with the pagefile is the best for performance, please provide benchmarks. Thanks. 
Dude, I've already provided a link to a guy working at MSFT who has atleast spoken to the guys who wrote the memory manager.
If you believe reducing the size of the pagefile (and making it static) has a negative performance impact, why don't you provide some benchmarks and get the glory of proving that guy from MSFT wrong?
Personally I don't believe it will matter much, even if you do factor in fragmentation. (it certainly won't matter for me, because I seldom experience paging -- and there's the rub: different systems doing different things will behave ...differently, and running the OS on full auto is seldom optimal) But bottom line: Why waste disk space if you don't have to?
If you believe reducing the size of the pagefile (and making it static) has a negative performance impact, why don't you provide some benchmarks and get the glory of proving that guy from MSFT wrong?
Personally I don't believe it will matter much, even if you do factor in fragmentation. (it certainly won't matter for me, because I seldom experience paging -- and there's the rub: different systems doing different things will behave ...differently, and running the OS on full auto is seldom optimal) But bottom line: Why waste disk space if you don't have to?
I'm not saying that messing with the defaults reduces performance. I'm wondering how much performance you are gaining.
I read the blog. The guy did not say that there would be a performance difference.
Disk space is cheap enough that a couple of gigs no longer matter.
RichUK
Lifer
- Feb 14, 2005
- 10,341
- 678
- 126
question if defraging the page file can improve performance which i believe it would do(as it will reduce the seek time), if you just defrag the whole system would that be sufficient ?? i would think yes .. just asking because i always like to tweak for extra performance
RichUK
Well to those who say that this is the wrong forum - I say "sorry, will do next time." I spend all my time here in the CPU, General Hardware, and Video forums so it never even occurred to me that I should use a different one.
And to everyone else, jeez I didn't know this was such a can of worms! I will try using a 1000mb static file and see how that works. And to anyone interested in such things, I have seen a very noticable increase in stuttering / lack thereof in Battlefield 2. You may not believe me, I won't bench it for you, but there is a definite difference no doubt about it at all.
And to everyone else, jeez I didn't know this was such a can of worms! I will try using a 1000mb static file and see how that works. And to anyone interested in such things, I have seen a very noticable increase in stuttering / lack thereof in Battlefield 2. You may not believe me, I won't bench it for you, but there is a definite difference no doubt about it at all.
You should have your page file size 1.5 times as large as the amount of system ram. Yours should be 3 gigs. You should set this as the minimum and maximum to prevent fragmentation. Ask anyone who has taken the Microsoft Certification for XP, this is the approved solution.
1.5 times is an overkill, try 1:1 ratio? 2gb = 2gb... very easy, even a kid will have no problems figuring this one out. Then again I could be wrong, different users need different swap size. Some 512mb some 1gb depending what application you run and how much is needed to swap out. If not work well for you, experiment what works best...
It will allocate more room, but then you end up with a fragmented page file, it has to add additional space from another area of the hard drive. If you set the min and max to the same, you reserve that space for the page file, whether its being used or not, it has that block of addressable space.
But the reason behind this advice is that you want to help the OS create a complete crash dump in case of a BSOD. If that isn't of any importance to you, then obviously you don't have to... (I take a different approach with my servers -- I'd like them to cough up a complete dump, but at home I'd rather only do this in case I get repeated BSODs and think the kernel dump isn't sufficient)
The pagefile is always in use. There is a window of opportunity while the system boots, but not all defraggers support that, and not everyone bothers booting all the time... (at work I only reboot when installing a security hotfix or add hardware -- could be as rare as every half year and even then would I be grumpy if I have to wait for a defragger to do its stuff)
As for seek time, as I said I have SCSI drives and plenty of memory. I don't expect it to page in very often and SCSI drives seek fast anyway... But certainly, this isn't always the situation, and I too would advocate a static pagefile. There's no reason to fragment if it can be avoided. I suspect though that if you let Windows manage the size, it will start at the size of physical memory, precisely for the reasons stated above (preserve a complete crash dump in case of BSOD).
Besides... If the system run out of virtual memory, then you'll get a "low on virtual memory" error message and you'll know you have to buy more memory, tweak pagefile settings or check if you run any apps that leak memory. IMO it is better to check these things properly rather than grow the pagefile automatically.
Someone said that disk space is cheap; Adding another drive adds more noise and heat. And SCSI drives aren't exactly what I'd call cheap. (but then I have to pay 25% VAT on top of whatever you guys pay...)
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter