Windows 7 |
Post Reply | Page <1234 6> |
Author | ||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
Interesting. Seems I'm not the only one to be suffering extended boot times as a result of SP2.
I see the same as a lot of others - a black screen for ages before the orb appears. Best regards, Vulcan. |
||||||||||
Matt N
Chief Pilot Joined: 02 Apr 2008 Location: Hertfordshire Points: 2287 |
Post Options
Thanks(0)
|
|||||||||
I've installed FSX, with SP1 and SP2. I've switched all traffic off, left the weather to the "Ultra High" setting, switched AA on, leaving lens flare and light bloom off, and set the Plane Detail to low. I then pushed all scenery sliders to the max. I get ~30 - ~50FPS below 3000 feet, and ~50 - ~100FPS above 3000Feet. Also no stutters or lagging. I'll give the same a go in Vista now and see what that comes out like.
Matt.
EDIT: Forgot to mention, that was just FSX alone, no add-ons. In the default A321.
|
||||||||||
Originally Posted by MartinW
I use mine for spare knickers when I'm traveling. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
Well if the program is constantly loading scenery files from disk it's only logical to assume that a slower disk is likely to casue and overall lower performance than a faster disk. As said, if you could get a suitably large floppy disk and run FSX from there, do you think it would possibly run as smoothly as from a fast hard drive...?
A system will always make use of a page file which is why it shouldn't be removed, you've disabled your page file under Vista which will only hinder performance by causing non backed changes to be retained in memory at the expense of backed stuff being dropped and having to be reloaded. If by processes you mean the standard Windows processes then they are miniscule and not worth the bother of shutting down.
Agreed. I use Avast which allows directory excemptions from scanning, i.e. don't scan any files loaded from the FSX directory tree. Don't think AVG (the free version) allows this so the on access scan is potentially at work when scenery is being loaded.
It's also an old engine, the same as Lock On afaik.
You can do that but I never saw any difference (neither have others on the forums) whereas under Vista and 7 it does seem to make some use of the dual cores. Not the sim itself, it's not coded to take advantage, rather the DX9 stuff if the background I think.
Haven't installed SP2 yet on my laptop, it's sitting in the update list. That's good to hear though if that's finally put right.
Something disabled or removed perhaps...
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
No - any new account I create doesn't exhibit this behavior. Only the first admin account.
Hmm. Interesting.
Lots of free memory so not an issue. If it does run out of memory it isn't pretty - the system crashes in ways you never thought possible. Quite dangerous actually as it becomes extremely unstable. I really had to go out of my way to achieve it though. I tested it deliberately to see what kind of margin I had. Lots is the short answer (2 Gb at least with FSX running). I was quite impressed actually. Best regards, Vulcan. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
So why disabled the page file in the first place...?. You are not gaining any advantage and only setting yourself up for problems as you've detailed and already experienced...
You can never have enough memory to be able to do without the page file unless you have more memory than all the programs you are going to use at once plus what the OS wants plus some extra to account for changed and unbacked store.
You've already encountered running out of memory so obviously your decision to turn it off has been proved erroneous - switch it back on, let Windows manage it (since it can manage its memory requirements far better than you can) and leave it alone...
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
No, if you read it correctly, you would see I tested it to see what margin I have. If it crashes because I disabled the swap file, that's my problem. It has never done so in *NORMAL OPERATION*. EVER. Therefore, by definition, there is no problem!!!
Now you're starting to understand (your comments suggest you don't, by a mile).
No, that's not what I said. Read it again please. Best regards, Vulcan. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
But there is no advantage to be gained from disabling it - only potential issues. So why disable it...? It's there for a reason, it's there to hold unbacked store, it's there because Windows can manage it's memory requirements better with it and can do so better than you.
There is no problem because it doesn't crash? The overall performance is potentially being handicapped by your decision to turn the page file off, it may never crash, doesn't mean there isn't still a problem. You are forcing Windows to drop potentially often used code and minimise its disk cache because it has no choice other than to keep unbacked changes and updates in RAM because it can't page them out.
My comments suggest I don't understand what? The absolute waste of time disabling the page file or what you are trying to achieve?
Either way, what is the reason for disabling it? It has no advantage and only brings possible problems...?
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
I disabled it precisely because it forces Windows to cache in memory.
To understand the behavior of the swap file in Windows you have to go back to Windows 3.1 or earlier. Windows does *NOT* use it like an emergency memory store when physical memory is low. It uses for operations it just doesn't need to. By disabling it, I'm forcing Windows to live without it, and it is quicker for it (memory will ALWAYS be faster). Please research the history of the swap file and how it is really used. No wonder there is so much mis-understanding. Even this guy hasn't got it right: http://www.aumha.org/win5/a/xpvm.php
Note that the problem is NOT with the swap file at all - it is in fact with the Norton software in question. MVP or not - this guy is wrong. This is completely bogus information too:
It should be the maximum you can fit on the disk. Even back in 1997 most systems had sufficient HD space for the full swap file size to be allocated and fixed. This is how I did it, and never suffered a single problem as a direct result. I read plenty about it back then though where people were running out of virtual memory. They ALL had dynamic swap files. Here is why a dynamic swap file is bad for your health (and you could read plenty of stories over at AVSIM about it too): Windows is coded to use it, if it exists. On 98 (and earlier), it was not possible to disable it completely. Even if it was 1 Mb in size, the system was happy. Disable it though, and it became unstable quickly and complained like crazy. I've tested this with every version of Windows since V3.1 (and 3.11) on the then current hardware. I spent hours analysing it. Where things get nasty is in the way it is resized. Our MVP friend doesn't seem to know his computer systems at the machine level. Assume we have a 1024 Mb swap file, that can dynamically resize to up to 4096 Mb. We're using 800 Mb of the 1024 Mb already, and Windows knows this. In quick succession, I come along and demand 400 Mb of memory, NOW. I then fill the memory with junk whilst performing complex computions that hammer the processor. I release 100 Mb. I then make a demand for 500 Mb more memory, but my process has hogged the processor, and the request for more memory got delayed. I try to make a write into the memory that is not yet allocated and...... "Insufficient Virtual Memory". I used to see this error a lot of systems with dynamic swap files. I'd check the swap file size, and it would be say 2500 Mb in size, and not the 4096 Mb it could have been (thus meaning there was actually enough space, if only Windows had allocated it in time which it had clearly failed to do). What just occurred was a RACE CONDITION. My app tried to use the requested memory ahead of Windows being able to supply it, because my app blocked the call that was supposed to resize the swap file. The memory wasn't there, and so it thought it was out of memory (technically, it was). By fixing the swap file to the largest size possible, this race condition is avoided, and so there is no problem. The other option (and the one I took) is to disable it, so Windows doesn't try and allocate to it, avoiding the associated problems. This is also BS (same site as above):
Any application that ever requests "hundreds of megabytes" more memory than they require are total junk and want to be avoided at all costs!! Real programmers NEVER allocate more memory than they require at any one time, and free it at the earliest opportunity. THIS IS BASIC GOOD CODING STANDARDS. Anyone who doesn't understand this is not fit to write commercial software. Given all the above, if you still think I'm wrong for disabling it, there is no hope. Even this MVP guy has it completely BASS ACKWARDS, and wants stripping of the qualification. Best regards, Vulcan. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
Windows will still 'cache in memory', that's what it does. You are just limiting it from managing its memory the best it can. It knows better than you.
No, it doesn't use it like an emergency memory store, it uses it for unbacked store, stuff that doesn't reside in physical files on the disk. "It uses [it] for operations it just doesn't need to" - what's that supposed to mean...? It uses it for unbacked pages, stuff that doesn't exist anywhere other than in RAM, it doesn't need to use it for that...?
That is just plain wrong. Yes, you are forcing Windows to live without it. No, it's not quicker for it and memory will certainly not be faster at all, let alone ALWAYS because you are forcing it to retain stuff that it otherwise wouldn't have to resulting in progressively less and less available RAM because that stuff that isn't backed and can't be backed because you've cut it off from the page file designed for that purpose has to remain in memory...
Sorry matey, I think it is someone else who needs to do the researching.
Disabling it forces Windows to retain all non-backed stuff. I.e. stuff that has changed, stuff that you've edited or made changes to, stuff that games have made changes to etc. Stuff that doesn't exists as a physical file on disk.
Because there is now no page file Windows cannot move any of this out to disk as part of its memory management, something that it is perfectly capable of doing and something that is can do far better than you, so instead it's forced to drop stuff that is backed on disk, i.e. normal program code in the .exe's you are running, the .dll code that is running, the application and OS code etc. The stuff that is often needed and required and would be far better sitting cached in RAM ready for use, Windows cannot make the best decision on what to drop and what to dump out to the page file to make space because you have forced its hand so it has no choice other than to drop such stuff and then reload it from the physical files later when its needed again.
If you are referring to the part bolded, then that is perfectly valid. It is best to leave Windows manage the size of the page file.
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
You don't write software do you? Best regards, Vulcan. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
You've edited the post but I'm not going to really add anything, it's not worth it. However...
Why the hell would you want a fixed page file the maximum you can fit on the disk. Sorry, that's just crackers.
I have best part of 400GB left on my system drive, are you suggesting that should be a fixed page file? Why? Let Windows manage it, it will set an initial size and then adjust it if you ever need more than that space.
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
I just read your response. No wonder the IT world so is so stuffed up. That is complete rubbish.If that is really how you think it works.....
You obviously don't know how this works AT ALL. I can't be bothered dragging out all the MS Technical Articles on updating system files, and how copies are held in memory, thus releasing the on-disk copy, and how the disk copy can be updated without affecting the in-memory copy, etc... It is possible to delete system files and not crash the system. God help you on the restart. Best regards, Vulcan. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
"You don't write software do you?" - Plenty I could say in response to that but I'll just leave the roll eyes to say it all...
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
Unless I'm sat here marshalling every byte in RAM, I *never* manage the memory. I simply say "Hey, Windows! Give me 4 bytes for this LONG will ya?". It tells me where to find it. If I don't have 4 bytes to allocate - OOME strikes, or a system crash ensues, or something. Who knows. That's the nature of no free memory. Best regards, Vulcan. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
You've turned the page file off, you've made a decision that says "I know better than Windows"...
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
No I haven't. I've told Windows it isn't going to swap out memory from physical memory to the disk. That's all it means in reality.
Best regards, Vulcan. |
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
See anything wrong with this code?
|
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
"Whoo, impressive..." What's this, an e-peni' competition or something - blowing my own trumpet is a sad pastime I don't really like doing I'm afraid.
I don't really see anything in the code, means nothing to me, not my field. I could equally drag out something of mine but I can't be bothered.
You've written a few things above suggesting what I've posted is rubbish and I don't know what I'm on about etc. I'm afraid it is you who is wrong matey, far wrong...
Edit:
|
||||||||||
VulcanB2
Chief Pilot Joined: 02 Apr 2008 Points: 13365 |
Post Options
Thanks(0)
|
|||||||||
Nope - like I said, removed the code.
Simple experiment that anyone can try: disable the swap file and tell me if your computer implodes as a result. If I'm wrong you'll know pretty quickly. Answer me this, too: why so many complaints of virtual memory errors, etc.., if dynamic swap files are the way to go? Surely you've seen this, working in IT????? To answer the question, there is nothing wrong with the code, except to say that the line that frees up the handle when it goes down the FALSE code path is redundant. The system is shutting down, so why care about the memory? See, I'm considerate when I write code, and wanted to make a RL example of that idiot in the link I posted. I could leave the handle hanging - it won't matter. We're only looking at 4 kB or so of memory. He's talking of apps allocating "hundreds of megabytes they might not use", then perports to be an expert on the ways of the swap file, and worse, you seem to agree with him? Programmers care nothing more than how much memory their apps use - me included. I worry about 4 kB whilst he thinks we just arbitrarily allocate "hundreds of megabytes"?? He's on another planet entirely, and why I don't ever want to be an MVP anything. I'm already avoiding MCSE like the plague. You aren't helping by actually agreeing with his obviously uninformed comments, whilst he is using his MVP status to say that everything he is saying must be correct. Worse, you actually seem to agree with him, but based on what? You can't beat studying a system to determine how it really works. Read all the text books you like. RW Example: The Airbus FWC (Fault Warning Computer) calls "RETARD" passing 20 ft RA on landing. It's the only time it calls it out (or is it..?). http://www.youtube.com/watch?v=oNk-JuXCGsM They don't mention that anywhere in the manuals. The real thing does it too... My point is that the way a system ACTUALLY works is not always what the designers TELL YOU. Windows is hardly the best piece of software ever written anyway. Best regards, Vulcan. |
||||||||||
Magic Man
Chief Pilot Joined: 02 Apr 2008 Location: South Wales Points: 5336 |
Post Options
Thanks(0)
|
|||||||||
Nobody said anything about it imploding. You yourself said what happens when it runs out of memory. You purposely did that but the end result is the same - it can run out of memory and you lose stuff.
Simple experiment anyone can try: disable the page file (correct name) and tell me if you notice any performance improvement. No? So why bother. Now load up Photoshop and edit some large raw files with plenty of undos, layers etc. or open up Premiere and do some video editing but don't yet save any of your work and now see what happens. Now imagine all those things you were doing had taken some hours of work and now the system is sitting there hung or has crashed and you've lost it because you thought that you were better at managing system memory than the OS. (Actually, think PS uses it's own memory manager anyway and uses scratch disks so may be okay, but the example stands, just subsitute your favourite large memory hog app').
Never seen any complaints of virtual memory errors. Actually, the only problems with virtual memory I've come across is where someone has disabled their page file or fixed it at a small size resulting in the same thing, oom errors.
I doubt very much that the systems or code you've written are the dogs danglies neither but I'm sure if some guy comes to you and says that he's just turned off a big chunk of stuff that your code uses to make informed decisions on optimal performance because he thinks he knows better then you'd laugh at him as well...
Some similar opinions to you and their corrective, better informed replies on the matter...
Yep, all from the same place but I don't want to waste any more time looking around for the correct facts. You may disagree with them, if so join up there and post your opposite views... Go on, I'd love to see it...
Best regards,
MM
|
||||||||||
Post Reply | Page <1234 6> |
Tweet
|
Forum Jump | Forum Permissions You cannot post new topics in this forum You cannot reply to topics in this forum You cannot delete your posts in this forum You cannot edit your posts in this forum You cannot create polls in this forum You cannot vote in polls in this forum |