this one goes out to those of us who wrongly believe that putting your swap file on another partition on the same hard drive makes the computer faster these are my reasons 1) the first partition would always be the fastest so by putting your swap file in another partition on the same hard drive u have relegated the swap file to a slower area of the disk thereby decreasing speed minimally 2)some of us think by putting the swap file on another partition we reduce/eliminate fragmentation, i say try this right click my computer on the desktop,properties, go to performance, virtual memory, then put the swap file on your first partition then restart, go over to the partition where u just removed the swap file from and try to defragment whoa did u see how many %fragmented? Also do u know that when defragmenting, windows 9x/Me does not defragment the swap file for most of us who dont have software to defragment our swap file on our hard disks please (those who have only one partition (C:) try this u most have enough ram at least 16mb (but it works on my 8mb laptop) right click my computer on the desktop, properties, go to performance, virtual memory disable virtual memory, the computer would hala and tell you of the consequences, ignore it. REBOOT IMMEDIATELY!! then while rebooting press F8 until u get to a boot options screen, select safe mode which would load minimal drivers then go defragment C:\ it would defragment the whole hard drive swap file inclusive. after defragging go over and enable virtual memory settings like before then reboot normally your computer should be a little happier now!

DO NOT disable virtual memory if you have less than 64mb of RAM. You WILL be sorry.

you didnt read well i said DISABLE virtual memory, the computer would hala and tell you of the consequences, ignore it. REBOOT IMMEDIATELY!! then while rebooting press F8 until u get to a boot options screen, select safe mode which would load minimal drivers............... REBOOT IMMEDIATELY INTO SAFE MODE IS THE KEYWORD! u ENABLE it after you are FINISHED please spend time reading what i posted more carefully

u ENABLE it after you are FINISHED ( the "it") is virtual memory

If you want to have a defragged swapfile, specifiy a MIN (no Max) and this will create one that is contiguous.

The best way, in my experience, to create a fast and efficient swapfile is to make the swapfile large enough that it won't likely need dynamic re-sizing under normal use, then use a program like Norton's Speed Disk to move the entire swapfile to the outer edge of the primary partition on the fastest hard drive in the system. If the Min swapfile size is set large enough, it will never fragment, because it won't re-size. If, under heavy usage, the swapfile does need to enlarge itself, the new section of swapfile clusters will be separate from the main swapfile area (but not fragmented). When the extra swapfile space is no longer needed, Windows will eliminate the "extra" swap clusters it created, and the main swapfile will always remain unfragmented and on the fastest section of the platter. ONLY set the Minimum size. NEVER NEVER NEVER limit the swapfile by setting a Max size. IF Windows should hit the ceiling, it may crash! My swapfile (on a 10Gb drive) is 200MB. I have not seen it have to dynamically resize itself in over 2 years, even under very demanding usage. Hmmm, maybe my swapfile is too large...

The various points raised in this thread are of interest of their own. I find it enjoyable from reading the experience generously shared by the posters. I would not take 100% the face valued of what we said, though, pending further analysis. I only want to raise two points here for your consideration. [1] It is said : "the first partition would always be the fastest". I cannot agree without backing it up by evidence. Sure we have to compare Read with Read, Write with Write---buffered, sequential and random. Tests of the kind involve considerable fluctuation/randomness. My own small sampling of benchmarking tests on real working harddisk leads me to conclude momentarily that _no_, not so. Equal size partition on the same hard disk should have the same characteristic within statistical acceptable confidence level. If not, the only blame is on the manufacturer. They should have this behavior well documented on their product spec. I myself is satisfied from user's point of view of their uniformity. [2] If you monitor the dynamic sizing of non-fix-sized swap file, it increases and decreases normally in size of the order of 4+/-MB to 20+/-MB easily. And, in-process, memory never ceases to acquire and release. So I believe, in a short time-scale in comparison of power-up time for normal users, the physical locations of consecutive memory addresses in the linear address space are translated from a well-order physical address config to a near-to-choatic one. So, there is good reason statistically not to over-worried about the fragmentation of swap file. It is fragmented enough by its own working already, even for win3x. So sequential read & write in a major of time is a ludicrious ideal. Of course, it is all relative. Thank you for your attention.