Looking for a free RAMdrive

I don't think RAM drive ( as featured in MS-DOS and/or Windows 95/98 ) will improve system performance by hosting pagefile.sys in free or unused Random Access Memory space.

Let's take a look at ubeateble Win NT expert's  Mark Rusinovich article :
"The Memory Optimisation Hoax" here --> http://www.windowsitpro.com/articles/index.cfm?articleid=41095

One more example that there's no solution for what you want to accomplish:
http://experts.about.com/q/1062/3219159.htm
or one more
http://www.hardwareanalysis.com/content/topic/49741/

Here's something you might wish to take your hands on and it totaly free Windows Registry hack:

http://www.cyberwizardpit.net/article3.htm

and some commerciall tryouts :
http://www.cenatek.com/product_ramdisk.cfm

good luck
nedvis

Let me see if i can explain to you why your arguement against is wrong in this case, so far your links have claimed that memory optimizers don't work.  This is true on modern operating systems, i.e Win 2000 and up and 128MB and up.  If you were running a Win 98SE system with 128MB of memory, a memory "defragger/cleaner" is beneficial, to a degree.  I already understand that at the size of of memory i have, and the OS i have, a memory optimizer is pointless.  But, the fact is, my computer is slower than it needs to be because i still have the pesky pagefile windows doesn't want to do without.  And since this is a laptop, i don't have the option to set the pagefile to a different mounted drive i.e a "non system drive".  Seeing as i only have 1 hard drive.  I would like to create a small RAM drive from real memory to substitute for my hard disk, i can set the environment variables to utilize that emulated drive for the page file.  But, yes, i'm basically looking for that "ramdisk" program you linked to, only free and NTFS.  Also, a key benefit is that paged data, is on the hard disk, a very power costly operation.  If i could access that paged data that windows says it can't do without, it would cost alot less power if i accessed it from memory than from hard drive.  And one more thing, the article about windows not functioning on non-mounted disks, this is 95%, it is possible to install the kernal on an external drive, but the performance loss is so great it's rediculous to try.  So yes, it's true you can't INSTALL the OS on a non-mounted disk.  The page file can be on any fixed disk, even a "pretend" one.

Umm, putting a page file in memory defeats the purpose of the page file.  The page file is only used when you don't have enough physical memory to hold what you need.

By putting the page file in memory you are using memory that would other wise be used to store programs, data, and files in file cache.  Which will cause more paging, but instead of paging to disk, you will page to memory.

If you have enough memory to where are you not paging, having a page file on disk is not slowing anything up.

Depending on what laptop you have, you may be able to add a second harddrive.  Not all laptops can, but some can.

What makes you think that your PC is being slowed by having a page file that is not being used?

It doesn't matter if you have the memory available or not, windows XP WILL use some of the pagefile regardless of the amount of memory available.  That is the disfunction of windows, is even when you tell windows to not use a pagefile, it still does.  And i understand what you're saying giltjr, but windows doesn't want to put these files into memory, it wants them to be paged(on a disk), these tend to be special OS files that it is hardcoded to put into the pagefile, they will never be put in memory unless fooled.  And the only way to force them into memory, even when you activate the special registry setting to "Use all available memory before paging" is to actually place the pagefile into memory.  And no, i can't add a second disk.  Plus, it's not so much is it REALLY being slowed down, but, it's one of those things, i want to do this to make sure i'm as blistering fast as i can be.

   "The page file can be on any fixed disk, even a "pretend" one. "
     AND THAT'S THE  PROBLEM  I think.
Microsoft has already published their ideas about adding/using.managing memory ( Vista only ,please):
http://www.vnunet.com/vnunet/news/2142265/usb-memory-keys-speed-windows
and
http://www.microsoft.com/whdc/device/storage/usb-boot.mspx

        Regarding power savings via utilising free RAM , I dont think we can prove SDRAM/DDRAM  (as pagefile host ) is drawing less power than spinning hard-disk as an "natural" media for pagefile.
As long as pagefile is located in RAM drive there must be permanent  currency of electricity in RAM memory circuits to keep RAM drive and pagefile in it alive meaning that RAM drive is draining laptop battery non-stop and hard-disk can stop spinning until OS has to swap memory contents from RAM to the hard-disk and vise versa.
     " But, the fact is, my computer is slower than it needs to be because i still have the pesky pagefile windows doesn't want to do without"
     How do you know the pagefile is slowing down your laptop ?
     Did you benchmark disk throughput  operations, read-write performance ?
     If your system hard-disk is trashing frequently and you think it's because of virtual memory file you would be better to review what services are running and shut down ones that you dont need.
    Besides that I dont think you can eliminate virtual memory file from C:\ drive. In Windows 2000 it has to be  at least 8 Mb I think ( other sources have different numbers though http://www.petri.co.il/pagefile_optimization.htm  )

  more information can be find here:
  http://members.shaw.ca/bsanders/WindowsGeneralWeb/RAMVirtualMemoryPageFileEtc.htm

   Anyway I understand your concerns about system performance and the idea  of RAM-Drive utilization. But now we have to see results in measureable parameters.
   
   I'm really wondering if there's anybody using RAM-drive for pagefile.sys hosting.


   good luck
   nedvis
     

   
   .

 

Well, power consumption is drastically different i'm afraid.  The power it takes to maintain data integrety in a RAM module is 10% or lower than that to operate the hard disk.  Think about a laptop running in standby, the only thing that is really running is the memory module, which every few moments the battery "flashes" to keep the data on the array alive.  While, a motor draws a significant constand power supply normally, and in desktop computers peaking upwards of 40+ watts(during upspin).  The math comes out to that for every 256MB it adds about 2-4W of power draw.  So for 1GB it's about 8-16W, much less than the normal 25-32W of a hard disk in standard operation.

If you ask is the pagefile slowing down your system, the answer is yes, because windows wouldn't store a file it didn't need at some point or another, that would be inefficient, like the good old Win98 days.  Also, if your bandwidth is being used to transfer page information, it can't be used for anything else, simply put, the CPU cycles it takes to transfer from disk to CPU (also a very slow operation) can be throughput simultaneously with other data if instead they are sent from memory to CPU(a faster operation), then the hard disk isn't "busy" with that other task.

As for whether virtual memory can be eliminated from c: ?...well, i know of numerous cases where the pagefile is located on another drive and/or partition, for the actual existance of VM on drive C: in that configuration, i do not know.  But the documentation i've read leads me to believe that this is indeed possible, and in many cases a very noticeable "reconfigure".

I mean, listen folkes, i'm just looking for a program to do it, i'm not contemplating cost to gain necessarily.

if you still want to mess around with a ram drive on xp..

http://www.cyberwizardpit.net/article3.htm
http://www.hardwareanalysis.com/content/topic/7663/
http://tinyurl.com/a3b72  (goes to a google translation french->english)
http://www.tweakxp.com/article37323.aspx
http://www.forumsforyou.com/p/alt.comp.freeware/Ramdrive_for_XP_35476.html  (read the replies)

you'll find some info & a couple free solutions, and a couple of "pay-for" solutions...

In response to nltech, i commend you on a well presented response.  At least i feel like i got pooned like a noob about this.  Honostly, i wasn't aware there was a nomenclature difference between pagefile and VM(though, i take that with a grain of salt).  To me, they have always been one in the same.  I understood that the task manager isn't that great.  But still, even when you start putting the numbers together, they don't add up.  I'll use my current listings as an example, and maybe you can explain it, and if you can, i'll give you the points and run away with my tail between my legs and surf the net some more.

My current PF usage is listed as 364MB, i have a total physical memory of 1024MB of which 530MB is available.  The difference is of course the system cache which is listed as 564MB and Kernal memory at 66MB, 35 of which is "paged" and 31 is "unpaged".  So then, why is it that i'm using roughly 564MB of physical memory, and 364MB of it is listed as PF, where did the other 200MB go?  Even if i add up the PF and the Kernal i don't make it to the full "used amount"  Naturally, memory doesn't disappear, so where did it go?  I notice that PF = Commit charge, but commit charge is a fancy name for something i have no idea.

remember, "PF Usage" <> physical space on the hard drive used up by a pagefile. you need to look at that vbs script i mentioned to find that out... so check it out: download it, unzip it to a folder on your desktop and run it. what does it say (or not say)?

Well, the VBS didn't work on my machine, but i'll take your word for it about how windows manages those settings.

IIRC, its been awhile, when Windows is showing you memory usage it does NOT include file caching.  So you case the other 200 MB is most likely cached files.  Memory used for file caching is not backed by the page file and if that physical memory location is needed for program use, the page is "stolen."

When windows pages something out, it does copy it to the page file, period.  However it may not mark that page unused at that time.  I will leave the page marked as used until it needs it.  If  it does not need that page for something else, and the original owner needs it, it will use it at that time, with no need to page in.

Commit charge is the amount the page file that  used to back up the currently allocated virtual memory space.  As I stated before, memory used for file caching is never paged out.

Having a page file will slow you down in two situations, that I am aware.  First is that you have over commited your physical memory.  That is you have 1 GB of memory but your programs need 1.1 GB.  You are now going to start paging in and out which will slow you down.  Second, is you are close to using the max amount of memory and the OS can't cache as many files, or as much of a file, as it may want to and so it can't pre-fetch a file into memory for programs use.  However even though it slow you down, it allows you to run with less physical memory than you may need.

I beleive that Windows 2000 and newer improved memory managment some.  Under Windows NT, the OS reserved about 50% of the physical RAM for the OS's use, it did this up to the first 1GB of RAM.  At least that is what it appeared to me.  I had PC's that had 512MB of RAM that never got above 300MB of memory used and yet were using 500MB of the page file.  Increasing the RAM and changing nothing else, I started using more and more of the memory.  Adding 128MB (total 640MB)  of RAM it starting using about 360MB, adding 256 (total of 768MB) it started using about 400MB.  Once we got to 2GB it was using 900MB of memory.  However it is possible that MS just changed how they reserved memory for file caching.  I think they reserve a percentage of memory for file caching no matter what.

 

you probably have something blocking the execution of vbscripts... here's an exe version of it courtesy of doug knox.
http://www.dougknox.com/xp/utils/xp_pagefilemon.htm