I went ahead and installed a 2nd GB of DDR in my rig, primarily because I was down to 70-100mb of free physical RAM at the end of a fairly short flight in the lovely PMDG 747-400. I had not received the out of mem error, but even so, I figure I am keeping this 2y old machine a while longer as it works very well and has nice parts. In any case, I learned some things today doing some creative testing with the very potent Performance monitoring console in Windows (2K Pro in my case).I set up the console to show a couple of things of interest to me:1. Free Physical Memory2. Disk Write Bytes/Sec (I have 3 HDD, and I have OS on C:, FS9 on D:, and page file on E:) The counter was set to look only at E:, thereby assessing page file write to disk activity (I think!).Then, I set my Virtual Memory minimum at 2mb as recommended, and the max at 3070mb, also recommended by Win2K for my now 2gb RAM system. I found something interesting . . .I did a fly from KSFO > KLAX, maxed it all out, I mean everything, and used the perf console to see what would happen.Available Free Mem moved down to about 1.3GB fairly quickly, and stayed there until about 40 miles out from LAX. Write activity to drive E: was mild until being about 60 miles out, then jumped up to about 4-8 writes/sec, with about .4mb/sec average write activity steadily. Free Memory movded to about 1.2GB, and stayed there. Drive E: write activity increased from around 380K to 1MB/sec at times, though I couldn't hear it . . . Completed the landing with no troubles. The same held true when I set min/max to 3070mb, that is, page file write activity happened.Then I tried the same fly, this time limiting VM to max and min to 2mb. This time, no write activity whatsover to the E: drive. Free Memory = about 1.1GB upon arrival. No troubles.Then I limited visible RAM to 1024mb by modifying boot.ini with the MAXMEM=1024 parameter. Tried the same flight, with VM set down low again at 2mb min & max.All was well until final approach, when free RAM dropped down to about 40mb, where upon i had a CTD, presumably Windows got nervous and shut it down as all available mem was getting too close to 0% for comfort. Interestingly, with this 1GB ram/2mb VM setup, once again no write activity to drive E: where VM resides.So . . . if you are running maxed with these sorts of aircraft, and you want NO pagefile activity to speak of, it appears it's easily done by just setting min/max to 2mb, provided you have MORE than 1GB of RAM (at least with my setup, and this bird).Don't know if there really is any visible smoothness or performance advantage yet in have no paging of the HDD during flights, because there is a small amount of disk READING anyway (much more efficient I believe, from looking at drive performance tests) that goes on as a flight progresses. Pretty cool looking at the perf monitoring console, as you can set it up to watch all sorts of things. If you have multiple drives, you can easily see how much HDD read activity comes when, for example, a .wav file is called (as with ATC for example).Anyway, it was a revelation to me that A, you can shut off page file read/write by setting min/max to 2mb in the VM setup, and B, windows definitely uses the page file continuously even when there is mega amounts of unused physical ram available.I think I'll go ahead and keep the extra GB and leave VM at 2mb min/max. If I could figure out how to create a RAMDRIVE and access .wav files there I think it would be a good way of using that extra RAM sitting idly on the sidelines. Attached is a pic of perf mon console.Cheers,Noel

Thanks for sharing, Noel! Can't wait to try it out. Good luck with the .wav files, that would be awesome to get all of that on RAM. Keep us posted!-JeremyThe Ozark DogfighterHappy Flying!

Sorry Noel, but that's techno-babble of the highest order. Interesting reports of your activities in the sim, but you clearly don't understand how the Windows paging file/virtual memory works, nor the implications for the Operating System. I explained in a previous post that Windows Task Manager does NOT (NOT) measure paging file use, but `paging file in use + paging file reserved for use`. You didn't take the hint to look at some technical sites as this place has a past history of great `snake oil` debates (what a fantastic expression!) Suggest you take this to the Hardware forum where it belongs and where many more experts are on hand to advise you on exactly where your interpretations and conclusions are so wrong as to be almost dangerous. And now you're intending to extend your half-understanding to RAMdrives!Allcott

Mmm....:-roll Allcott a more polite tone would do better. I still consider the fora a place to discuss stuff and not to bash another.Best regards,Rob "Holland&Holland" de Vries http://www.emotipad.com/emoticons/Flying.gif"To go up, pull the stick back. To go down, pull the stick back harder"

You may be right, but this is not the appropriate forum, and if the poster can't even post in the correct forum, think about what taking his advice might do to YOUR system?So far he's resisted posting anything more dangerous than locking the VM settings, but the justifcation he's using might cause another member to take extended measures which could easily result in a failure to boot, reinstall and even a reformat. And the implications for creating RAMdrives need to be fully understood by anyone contemplating taking that step. I don't even use them with sound files, and I have a full home studio here with a dedicated computer with huge RAM. Most of the items he raises have ALREADY been discussed at length in the hardware forum. THAT is the place to put threads like this, not here where its very isolation implies that the advice is sound, which it isn't. Twaddle combined with guesswork is the kind of advice we don't need. Sure, post the results of twiddling, but don't suggest you even begin to understand what the results mean when it's clear you don't even understand the tools!I politely suggested he might like to view what more informative and technically aware authors had to say on the issue and don't get a `thankyou` or even a `that's interesting` before this rubbish appears - so don't talk to me about polite.Allcott

Allcott, first of all, I did take the hint, with all recognition that although I do not fully understand how the windows paging file yadda yadda works, I was NOT, hear me now, NOT, using Windows Task Manager. I was using Windows perfmon.msc, a HIGHLY user configurable tool to assess PHYSICAL DRIVES, processor, memory, and a myriad of and lots of other windows components. Try it sometime, before you pontificate. Everything I described was accurate, though may not have been interpreted correctly. Why don't you try duplicating what I discovered, and then tell me what you learned, if you are open to learning anything.Cheers

All your nasty insults aside, Allcott, what part of my technobabble was incorrectly INTERPRETED? I'm trying very hard to remain constructive with you Allcott. Go ahead and tell us exactly what is inaccurate about what I took the time to share. Was it that setting min/max to 2GB DOES NOT in fact stop writing to the disk that contains the poge file? Was it that despite how much unused RAM the machine has, when VM is set somewhere above 2mb that windows DOES NOT in fact write to the pagefile? Start talking, now this time try to be contructive. And, I'll remind you again, because you failed to take note, we aren't talking about usiing taskman. Good luck

I'd also like to hear what he has to say. Particularly about the implications to the OS of running a very small pagefile. I routinely run this system with no pagefile at all when using FS2004 with no adverse results. And, yes, I DO understand how WinXP allocates and uses pages so I hope that doesn't come up again :-) . But, Allcott seems to be in a fighting mood and I'm afraid a meaningful dialogue isn't going to be possible. Still, I'd like it hear a constructive critique of the downside risk of using a very small or zero size pagefile (Hint: with enough physical system memory there is no downside).DougP4 3.0C SL6WK @ 3.510 (1.5875 vCore - 234)Asus P4C800-E Deluxe (BIOS 1019)4 x 512MB Corsair TWINX CXM3700 (3-4-4-8)5 x 300GB WD HDD'sHercules 9800 Pro (128MB @ 410/360)Enermax 431W PSU

(sigh) Perfmon performs exactly the same - it measures not only the paging file in use but also the paging file RESERVED for use - you therefore CANNOT conclude I/O requirements from reading data produced via Perfmon.exe. Again, you don't understand the tool. Look, if I really can't get through to you, read this:http://www.aumha.org/win5/a/xpvm.htmwhich not only confirms what I say, it also provides links to tools that will enable you to measure what you want to measure, if you want to carry out your `experiment. Not only that, it tells you what you need to know about paging file use in Windows XP, most of which contradicts what you `discovered` (and if you search the forums you'll also find that your `discoveries are not new, not yours, and not correct.) I think that about summarises it, OK? After showing you that, I hope you will realise that EVERYTHING you claim as a result of such `measurement` is speculation and guesswork, or just plain wrong. Nothing wrong with being wrong, and if you choose to fool yourself then go for it, but please don't try and fool others with technobabble. If you don't understand the tools, stay out from under the hood. By all means report your findings, but leave the conclusions to others.FWIW, there have been many attempts to force FS to use more RAM, now that 1gig+ is commmonplace. None have been successful. It's an old engine, in need of tuning, but you and I are not the mechanics to do it. It's an MS issue that hopefully will be addressed in FS10. Good luck with your RAM drive, see you after the reinstall!Allcott

> I'd also like to hear what he has to say. Particularly>about the implications to the OS of running a very small>pagefile. I routinely run this system with no pagefile at all>when using FS2004 with no adverse results. And, yes, I DO>understand how WinXP allocates and uses pages so I hope that>doesn't come up again :-) . But, Allcott seems to be in a>fighting mood and I'm afraid a meaningful dialogue isn't going>to be possible. Still, I'd like it hear a constructive>critique of the downside risk of using a very small or zero>size pagefile (Hint: with enough physical system memory there>is no downside).>>DougYou need to read the same article. There's no such thing as using `no pagefile at all`. The same article also indicates likely trouble spots from trying to do so. You are NOT using your RAM efficiently if you try to work without a paging file. And troubleshooting Windows becomes problematic if you do. The fact that you can do so should not be assumed as being the reason to do it. You can throw yourself off a bridge, but would you want to?.Sound Engineers I work with have tried to achieve the same ends for similar reasons (`large file sizes, slow throughput, must be a better way - let's dump it all in RAM, oops, where's the system gone`) and have always come a cropper.Oh,and Noel, I did try your experiment. I/O stays the same, hard drive chatter (i.e. disk access) remains entirely dependant on the scenery in the sim, not the settings ion the OS. I tried without a pagefile too. After rebooting due to the `insufficient memory` error I conclude it doesn't work.Doug, it's not a question of being argumentative, it's simply a matter of stopping fellow simmers being suckered into making changes to their computers which at best have no effect and are a waste of time, and at worst will upset stability and possibly require serious remedial action. There's nothing here to argue. Noel is wrong, is all.Allcott

If you're referring to the Alex Nichol article, Allcott, I'm very familiar with the contents. It may be that there is a lot of confusion about that the meaning if "is" is. To quote Alex: "Strictly speaking Virtual Memory is always in operation and cannot be

> We do agree, however, that elimination of the pagefile is>not for the novice nor the faint-of-heart. As you point out,>many things can happen and many of them are bad. But for>anyone with 2GB, or more, of good system memory, eliminating>the I/O accesses associated with the pagefile can, indeed,>bring about sometimes significant performance improvements.>>Doug>>Edited because I can't spelBut only if you can force FS to use it, and I don't think we've ever seen any evidence that you can. Feel free to post it if you have it, though.Now I CAN think of exceptions where more than 1 gig of RAM can be useful to an FS9 user, but then you're into specific-user territory, and a user with the kind of addons that would benefit from such a setup, or multiple-threaded,interweaved systems (WideView springs to mind although I have little experience of multiple-screen, multiple CP flight sim setups) is likely to have the specific technical knowledge to know what will, and what won't, work on his rig(s).For the 99.7% of the rest of the simming community, you can paint it whatever colour you like, but you cant make a Lada into a Porsche. no matter what badge you stick on it. And adjusting the SIZE of the page file has NOTHING WHATSOEVER to do with the I/O. XP will always use RAM, will always use the paging file, even if only for debugging - and will protest if it doesn't have enough of either. We agree there.We've had multiple threads over a period of some years looking at how FS makes use of RAM, but no-one has ever been able to show that you can force FS to use more RAM, if more RAM is available. FS is a funny old program, from the point of view that it's heavy on CPU cycles, heavy on graphics, but when it comes to I/O calls it doesn't really matter how you set it up, how you adjust the O/S or how you present your `findings` to the sim community, at the end of the day the demand remains pretty much the same. Most of the I/O is at startup, when the bulk of the sim is loaded. After that, those irksome hard drive chatters have nothing to do with how the Operating System is set up, it's to do with how Flight Sim is designed. It's no more difficult than that, and if iot needs to be modified you'd do it by playing with the sim, NOT the operating system. I would like to ask our poster HOW he thinks that Virtual Memory size adjustment changes the nature of I/O calls in the sim? If it were possible to do so, FS would have shown an adaptability to new hardware the like of which we have never seen. But let's face the facts: Flight Sim was designed before Windows XP was even thought of. The latest incarnation, while refining the code and modifying it to suit the needs of the then-new OS, still failed to take advantage of the abilities of a Win 2000-based system, because it was designed to be used by all types of users, on all kinds of OS - Win 98 included. As the min specs on the box indicate `Win XP with 128meg of RAM` as a minimum, it follows that MS designed the code to fit with that profile. Ten times as much RAM was not even considered at the time, I'd be willing to bet.We've seen plenty of threads looking at the difference in performance between 512 and 1 gig of RAM, and 1 gig of RAM and more. The only time that more than a gig makes any sense is if you also have a raft of addons running at the same time. Otherwise the RAM call peaks at about 700-odd, and that's it. FS can't make use of any more. Thats FS, not XP.If Noel has made it do so, he is truly a genius, and I will bow down before him and kiss his little feet. But tell us how he's done that by modifying the VM settings in the O/S? It's meaningless, and the pseudo-technical jargon being used to justify something which hasn't even been correctly measured is likely to cause nothing but confusion. I remember the same approach - intellectual gibberish - was used by some mock-expert to justify changing the program settings for memory usage to `system cache` - and look how that ill-considered piece of `advice` caused chaos among ATI card users, who ended up reinstalling their operating systems to fix the disaster that particular `tip` created.At this point, FS is a mature product. Put simply, the real experts have devoured it, and there's nothing left on the plate. If there were any `miracle cures` floating around, don't you think all of the experts over many years would have found it?I don't doubt Noel is well-intentioned. But he could at least be honest enough to say: "This works for me, but I don't have a clue why" and then everyone else would at least be on warning if they attempted it themselves. Instead, he attempts to vindicate his findings in the most misleading garbage I've read in these forums in recent months, and that's not clever. Especially as the techno-vomit might encourage people to make changes that, some way down the line, long after they've forgotten doing it, lead to critical system errors or instability.Allcott

I thank all involved for clarifying. I was excited to see a rather large change in FS performance could be acheived, but alas! I was mistaken. Oh well :) I have a Question:I currently have 768MB of some of the cheapest PC2700 RAM out there. It's old. It smells funny. It's always gone when I wake up in the morning and never calls.Anyway, I have an Abit AI7 board, and I love this board. I can run 4GB of PC3200 RAM on it. I really want to. Better half says NO, haha. ANYWAY, I want to put at least 1GB, possibly 2GB, of Corsair unbuffered RAM in this system. Will FS take the hint and run like a raped ape? Or will it be only marginally different?Pretend I'm inherently clueless to techno-anything when answering please, I'm rather slow today.Thanks guys!-JeremyThe Ozark DogfighterHappy Flying!

My goodness Allcott. Once more, when I see zero write activity to drive E:, where my one pagefile resides, when VM is set to 2mb min/max, and plenty of disk write activity [on drive E:] when VM is set to 3070mb mon/max, what do you take from that? I'm certainly open to a better explanation, but you haven't provided one yet, and I doubt you can. There is no question drive E: is sitting idly in the former situation, and is being accessed in the latter. I also note the amount of free physical ram at the end of a long flight in the PMDG 747-400 with VM at 2mb, is much lower than when VM is set at 3070mb. Please help us all understand this rocket science.Also, Allcott, I did set up RamDiskNT over a year ago, when former versions would in fact allow you to put the pagefile to that ramdrive, and what do you know, I followed the directions carefully, and did not need to reinstall. Amazing. I'm sure many can follow directions besides you Allcott.And also Allcott, you are full of BS when you say Perfmon measures the same thing. Perfmon measures TWO variables for pagefile, at least for Win2k. Here they are: %Usage, and %Usage Peak. They are defined as "The amount of the Page File instance in use in percent." Notice how unambiguous that is. Page file 50mb, 20% in use. Hmm, wonder what that means . . .Also, go back and dig into perfmon some more so you know what the #### you are talking about. Perfmon also measures Physical Disk Writes and Write Bytes/sec. Here is the explanation from perfmon: "Disk Write Bytes is the rate bytes are transferred to the disk during write operations." Read it slowly so you get it Allcott. It's pretty easy once you focus on it rather than focusing on trashing other people. This parameter is actually about physical disk write operations Allcott. And, since in my situation my E: drive is essentially a data drive with no other ties to the OS or FS, except that it houses my only pagefile, I think the point I make is pretty much valid. Again, feel free to provide another interpretation, but you won't because you can't.Ken or whomever, please take note as to who began this nonsense.

Percentage-wise the need to relocate data can account for a good proportion of disk file read/write access. You have reduced your E: drive access? Then why did you INCREASE the Virtual Memory space? If you are trying to force RAM to be used then you should REDUCE it. You're contradicting yourself now. RAM offloads into VM. More VM, more offloading. Doesn't increase RAM usage in the sim. Just RAM offload in the system. You still haven't proved a thing.You truly do not understand what Perfmon does and how it defines `in use`. Go read up. I've explained what `in use` means and it doesn't mean what you say it does. If you're wrong on that, what are you right about?Physical disk writes bytes rate is exactly what it says it is. RATE, not volume. FS rarely writes back data. The virtual memory space is already allocated. The RAM offloads to the VM. It `sees` more VM it offloads more RAM. You have more RAM left at the end of the flight. So what's to explain? You haven't influenced the sim, only the operating system. A final analogy: I've got a little bucket with a small hole in the bottom. I buy a bigger bucket that holds lots more, but it has the same size hole in it. Which leaks faster? Neither. I'd need a bigger hole in the bigger bucket for that. Your VM/Paging File is your bucket, FS is your `hole`. Explain how you've made the FS hole bigger in your bucket as everything you describe seems to be the equivalent of painting pretty flowers on the outside of the bucket. Fine and dandy in its own way, but not likely to make the contents run out faster. You do not understand computers. That's OK. I'm glad your hot tip works for you. Just don't try and tell us why. And we do take careful note of who began this nonsense. Allcott