I am trying to understand the nature of increased page file sizes on a 32 bit WinXP machine. I know the max RAM that can go on this machine is 4GB, but how does increasing the page file size affect this? Any detriment/benefit to increasing page file size?
migrated from serverfault.com Sep 25 '13 at 18:27
This question came from our site for professional system and network administrators.
Computer programs must be in main memory (RAM) to be executed. Main memory is the only large storage area that the processor can access directly. Ideally, we want the programs and data to reside in main memory permanently, but RAM is too small to store all needed programs and data permanently and loses its contents when power is turned off. So Windows (and other operating systems) use secondary storage (Hard Disk) to be able to hold large quantities of data permanently. Also there are many types of storages that we can use (Flash drives, external HDDs, which are also the same HDD as in your PC, memory cards etc. ), the main differences lie in speed, cost and size.
Windows use virtual memory, a technique that allows the execution of a process that is not in the RAM. Windows has a hidden file named pagefile.sys in the root of one of your drive. And when RAM is full or Windows detects inactive programs (such as applications minimized for a long time) Windows just move that applications to pagefile to free up more RAM. Actully pagefile acts like “secondary RAM”, but it much more slowly because of HDD speed. So switching to an open application that you haven't used in a while could take a little bit time. And the larger pagefile, the larger amount of applications can be there. Also you can disable pagefile, and Windows will be forced to keep every opened application in RAM(this could be not a good option if you usually have not enough space on RAM, and after filling it programs are going to start crashing).
But like the chain’s strength measured by weakest part in it, the speed you need to access (and use applications) measured by slowest memory type program use(in this case HDD). So RAM has too high cost/MB ratio, users have to deal with HDDs as a secondary RAM, or use ReadyBoost feature(that also creates file like pagefile) for using flash cards as secondary RAM(actually flash memories could be faster than HDDs.) Nowadays SSDs could get a place of HDDs. You can imagine SSDs like huge flash drives , they are faster than HDDs, and cost/MB ratio is in between RAM & HDD. As users we can’t always afford enough RAM, so you can use Windows features to use nearly any type of memory as a secondary RAM: HDDs, flash and SD cards, SSDs…
Short answer is that the swap is used by the system as an overflow if main memory is completely consumed using a slower media to keep the processes going. Performance will suffer as you swap to the disk. With a 32-bit system, you have correctly identified that you have 4GB of RAM but only 1.4GB are usable by a process unless special flags are set on the EXE to allow for using more of the memory available which has a limit of 3.5GB of RAM for the process.
I hope this helps with your question. There is no detriment to increasing your swap size but no real advantage either. Decreasing the swap size could impact the number of processes that can run if you run out of main memory. You could run without swap space at all but that is not recommended.