Re: [hlds] windows 2003 memory tweaking
Clayton Macleod wrote: I think perhaps a refresher is warranted here, since you're forgetting one basic fact, which is even mentioned in the article listed. How many kernels have you built? Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system. This statement bears no meaning on what I said. This rule is still for a dynamic environment. You should always have a pagefile, and it should be larger than your installed RAM, even if you have tons of RAM and your memory usage never nears your RAM limit. The people that wrote the memory manager in the OS are telling you this, and you're thinking they don't know what they're talking about? Actually, the article was written by technical documentation staff. There are plenty of videos from the Microsoft development labs regarding their memory architectures and dynamic caching routines. I was fortunate enough to study these in detail during the windows nt 4.0 days. Alot has changed, but many of the basics of the NT kernel are the same. Several MVP's have also published detailed documentation on this subject and none agree with the above statement. This was the principle of 9X paging, and is no longer required for a stable kernel (whereas it was required to keep 9X stable in high memory conditions). Besides, with HD space being well under $1/GB there's no legitimate reason to complain about the wasted HD space for a pagefile. Just set the minimum size No, you misunderstand the reason - what is the data transfer rate of your harddisk? And your ram? Now, next question, what is the speed of the bus that each of them is attatched to? Frequency develops latency and switching rate develops burst timings. Is the bus intelligent / managed? What about DMA? What's the DMA penalty? Blah Blah Blah. There is nothing wrong with the defaults - but the defaults are designed for systems that can sometimes be run by home users rarely loading more than internet explorer, who want blistering performance - in this case alot of the static data and unused strings are loaded into the pagefile - as they are rarely needed. When they are needed it is often the case that processing is in flow for other operations (such as building part of the gui) and the IDE latency rarely causes issue. As such the choice of data page here provided more free memory for new applications, whilst producing a minimal impact to system performance. Similarly a user running an arbritrarily large single application, such as say Lightwave, will find that paging can become an excess when close to the physical ram limits. This is no issue when handling that single application, but the paging nature is not idealistic - even loading a small app such as an explorer window can cause a massive re-page of the now unfocused app. In this case, because the memory balence between apps is poor, and the system takes no knowledge of each programs intentions - a bad decision is made. In this case (one that I have spent a good deal of time on) disabling the paging executive was the only way to get smooth non-paging performance out of the system. Finally, FYI - registry entries are rarely wasted by Micorosft - most configuration options will change something when the switch is flicked and they are put there for good reason! Why don't you try looking for some information on those keys and see what other Micorosft articles you find - if you read those the same way you read that last one, we'll have you screaming that they are contradictory by the end of the week. right. On 8/14/05, James Tucker [EMAIL PROTECTED] wrote: These rulings are intended for systems with dynamic environments and are done for coninual performance puproses, and built for dynamic run-time environments. This is not the scenario that a dedicated box will be running, which much more resembles a static application set. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
uh, yeah. Like I said, you need a refresher. It's quite apparent from your disabling the paging executive statement. Clearly you think you are disabling something called the paging executive which you think means you're causing the OS to hit the pagefile less. Sorry to tell you, but the setting you're talking about doesn't disable something called the paging executive. What that setting does is stop pages from the executive from ever going out of ram and into the pagefile. i.e. the kernel, drivers, etc. It doesn't affect any applications in any way, shape, or form. Any applications you have running will still get pages paged out just as much/little as before you disabled paging of the executive. side note: it's really not necessary to respond inline like that for something so short as this. It does nothing but make replying to such a short conversation take a lot longer than it needs to, since it takes forever to trim out all the stuff you've quoted. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: How many kernels have you built? This statement bears no meaning on what I said. This rule is still for a dynamic environment. Actually, the article was written by technical documentation staff. There are plenty of videos from the Microsoft development labs regarding their memory architectures and dynamic caching routines. I was fortunate enough to study these in detail during the windows nt 4.0 days. Alot has changed, but many of the basics of the NT kernel are the same. Several MVP's have also published detailed documentation on this subject and none agree with the above statement. This was the principle of 9X paging, and is no longer required for a stable kernel (whereas it was required to keep 9X stable in high memory conditions). No, you misunderstand the reason - what is the data transfer rate of your harddisk? And your ram? Now, next question, what is the speed of the bus that each of them is attatched to? Frequency develops latency and switching rate develops burst timings. Is the bus intelligent / managed? What about DMA? What's the DMA penalty? Blah Blah Blah. There is nothing wrong with the defaults - but the defaults are designed for systems that can sometimes be run by home users rarely loading more than internet explorer, who want blistering performance - in this case alot of the static data and unused strings are loaded into the pagefile - as they are rarely needed. When they are needed it is often the case that processing is in flow for other operations (such as building part of the gui) and the IDE latency rarely causes issue. As such the choice of data page here provided more free memory for new applications, whilst producing a minimal impact to system performance. Similarly a user running an arbritrarily large single application, such as say Lightwave, will find that paging can become an excess when close to the physical ram limits. This is no issue when handling that single application, but the paging nature is not idealistic - even loading a small app such as an explorer window can cause a massive re-page of the now unfocused app. In this case, because the memory balence between apps is poor, and the system takes no knowledge of each programs intentions - a bad decision is made. In this case (one that I have spent a good deal of time on) disabling the paging executive was the only way to get smooth non-paging performance out of the system. Finally, FYI - registry entries are rarely wasted by Micorosft - most configuration options will change something when the switch is flicked and they are put there for good reason! Why don't you try looking for some information on those keys and see what other Micorosft articles you find - if you read those the same way you read that last one, we'll have you screaming that they are contradictory by the end of the week. right. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Clayton - you have mostly understood the articles you have read. Whilst you understand the general principle of a paging system, clearly, you misunderstand that they are systematic, logical and non-optimal as with caching algorithms. Clayton Macleod wrote: MS themselves say that there is no way for us to know how much of any process is in RAM and how much is in the pagefile, Really? Well that individual needs to be fired. You couldn't even become an MCSE without some basic knowledge of where to find this stuff. HERES SOME TOOLS: TASKLIST, TASKMGR, PMON from Systeminternals.com. The first two SHIP WITH WINDOWS! There are some more tools (IIRC) in the support folders of the install CD too (although I cant remember how much sysint stuff is still being shipped with 2003). so, I don't know where you're getting this information from. Exactly, this should suggest another inference that should be made. How much do you really KNOW about this? Please don't waste time. Besides, *everything* is in virtual memory. Everything uses the kernels addressing system, this much is true. And there is no point in keeping pages in RAM that aren't being accessed when that RAM could be actively used for some other purpose. I'd agree with that statement, well I do, but don't forget this discussion is about excess FREE memory, not efficiency of use of ram vs. pagefile. If pages are active, they'll be in RAM, don't worry about it. Define an active page? How long before it should be moved out of physical RAM? 100ms? 2 years? If you could define this accurately, you could get rich. That's the whole point - these systems are not optimal yet - they are good, but not optimal. Currently MS build very very dynamic (read: adaptive loopback system) caching algorithms, in fact, there is no other OS currently available which is as sophisticated in this regard. I would recommend you never say the above statement again, as it's simply not held true by the systems that really exist, they do try though. My next questions to you are: (although this is really rhetorical as I know the answers) - If a page is inactive and therefore stored in the pagefile and becomes active what is the process by which the page is loaded? What is the decision for which pages to move out of ram? What data is used to make the decision? How long does it take? What is the impact on the process scheduler? Does this create a temporary starvation issue? ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Clayton Macleod wrote: uh, yeah. Like I said, you need a refresher. It's quite apparent from your disabling the paging executive statement. Clearly you think you are disabling something called the paging executive which you think means you're causing the OS to hit the pagefile less. Sorry to tell you, but the setting you're talking about doesn't disable something called the paging executive. What that setting does is stop pages from the executive from ever going out of ram and into the pagefile. i.e. the kernel, drivers, etc. It doesn't affect any applications in any way, shape, or form. Any applications you have running will still get pages paged out just as much/little as before you disabled paging of the executive. Go check explorer performance in this scenario and fuck off with your arrogance. If you think that dynamic systems can be optimised form the theoretical standpoint then you have little experience with doing so. The choice of settings was made using educated emperical changes in order to find the optimal, which is the only way this can be done in most scenarios. Sorry, but that response is totally defunct, even though most of it is correct. Besides, you made no effort to think what IS and ISN'T really getting paged in and out in the precense of a fat application like that. Most applications are no where near standalone, did you forget that? side note: it's really not necessary to respond inline like that for something so short as this. It does nothing but make replying to such a short conversation take a lot longer than it needs to, since it takes forever to trim out all the stuff you've quoted. Actually, that ensures I don't miss anything. Have a nice day. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: How many kernels have you built? This statement bears no meaning on what I said. This rule is still for a dynamic environment. Actually, the article was written by technical documentation staff. There are plenty of videos from the Microsoft development labs regarding their memory architectures and dynamic caching routines. I was fortunate enough to study these in detail during the windows nt 4.0 days. Alot has changed, but many of the basics of the NT kernel are the same. Several MVP's have also published detailed documentation on this subject and none agree with the above statement. This was the principle of 9X paging, and is no longer required for a stable kernel (whereas it was required to keep 9X stable in high memory conditions). No, you misunderstand the reason - what is the data transfer rate of your harddisk? And your ram? Now, next question, what is the speed of the bus that each of them is attatched to? Frequency develops latency and switching rate develops burst timings. Is the bus intelligent / managed? What about DMA? What's the DMA penalty? Blah Blah Blah. There is nothing wrong with the defaults - but the defaults are designed for systems that can sometimes be run by home users rarely loading more than internet explorer, who want blistering performance - in this case alot of the static data and unused strings are loaded into the pagefile - as they are rarely needed. When they are needed it is often the case that processing is in flow for other operations (such as building part of the gui) and the IDE latency rarely causes issue. As such the choice of data page here provided more free memory for new applications, whilst producing a minimal impact to system performance. Similarly a user running an arbritrarily large single application, such as say Lightwave, will find that paging can become an excess when close to the physical ram limits. This is no issue when handling that single application, but the paging nature is not idealistic - even loading a small app such as an explorer window can cause a massive re-page of the now unfocused app. In this case, because the memory balence between apps is poor, and the system takes no knowledge of each programs intentions - a bad decision is made. In this case (one that I have spent a good deal of time on) disabling the paging executive was the only way to get smooth non-paging performance out of the system. Finally, FYI - registry entries are rarely wasted by Micorosft - most configuration options will change something when the switch is flicked and they are put there for good reason! Why don't you try looking for some information on those keys and see what other Micorosft articles you find - if you read those the same way you read that last one, we'll have you screaming that they are contradictory by the end of the week. right. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please
Re: [hlds] windows 2003 memory tweaking
hahahaha, I'm just having a conversation, Mr. Angry Eyes. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: Go check explorer performance in this scenario and fuck off with your arrogance. If you think that dynamic systems can be optimised form the theoretical standpoint then you have little experience with doing so. The choice of settings was made using educated emperical changes in order to find the optimal, which is the only way this can be done in most scenarios. Sorry, but that response is totally defunct, even though most of it is correct. Besides, you made no effort to think what IS and ISN'T really getting paged in and out in the precense of a fat application like that. Most applications are no where near standalone, did you forget that? Actually, that ensures I don't miss anything. Have a nice day. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
actually, no, task manager doesn't tell you exactly what you think it is telling you. If MS didn't purge their beta newsgroups after the end of their betas I could find a quote for you. (dammit) I'm not sure if a similar explanation is in any of their public docs or not. I'll take a look in the morning, day off with nothing planned. (well, actually, putting off getting under the car and getting all greasy, haha) On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: Clayton Macleod wrote: MS themselves say that there is no way for us to know how much of any process is in RAM and how much is in the pagefile, Really? Well that individual needs to be fired. You couldn't even become an MCSE without some basic knowledge of where to find this stuff. HERES SOME TOOLS: TASKLIST, TASKMGR, PMON from Systeminternals.com. The first two SHIP WITH WINDOWS! There are some more tools (IIRC) in the support folders of the install CD too (although I cant remember how much sysint stuff is still being shipped with 2003). -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Clayton Macleod wrote: actually, no, task manager doesn't tell you exactly what you think it is telling you. If MS didn't purge their beta newsgroups after the end of their betas I could find a quote for you. (dammit) I'm not sure if a similar explanation is in any of their public docs or not. I'll take a look in the morning, day off with nothing planned. (well, actually, putting off getting under the car and getting all greasy, haha) The discussions on MSDN are not important here, as the process scheduler was made to report accurately for SP2(XP) and SP1(2k3). Amoung other things this is a requirement for DEP internals. Most of these docs can be found through a partner login. Sadly, the only public docs I've found so far are the exact same as you have just sent. The latter of which is an old bug we used to suffer frequently on our Citrix mainframes. (Ah, the reason he's been screwing with process and memory management!). -- http://www.microsoft.com/resources/documentation/Windows/2000/server/reskit/en-us/Default.asp?url=/resources/documentation/Windows/2000/server/reskit/en-us/regentry/29931.asp http://www.microsoft.com/technet/prodtechnol/windowsserver2003/library/DepKit/3d3b3c16-c901-46de-8485-166a819af3ad.mspx http://support.microsoft.com/default.aspx/kb/q184419/ N.B. Did ya miss MS's other two support sites? (just joking, I _love_ the way they make us trawl 5 different places for info ;-). -- None of these negate what I have said. In fact all they do is negate what you said about never changing from defaults, as here you will find MS suggesting that people look at these articles for their solutions (having been on the phone for hours to MS waiting for some id-10t (it's been a while) to point me to Q184419, despite being outdated and useless to the problems we ever encounter on the citrix platforms. The techs are never as good as the consultants. :-( Back to the point in hand, setting this option can reduce paging in certain instances where the data paged is stored in driver space - this is particularly more common in certian development scnearios. You CAN analyse these effects by careful observation of kernel and process running times, along with changes in memory deltas and other less important varaibles. Generally you will need a program in ring 0 to properly observe these things though, and never forget that everything you do on the system affects the system. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
re: memory reporting, even the working set value you get from the 'performance monitor' doesn't really necessarily tell you how much RAM the process is actually using just for itself, because this value includes any shared memory, not just private memory. I'm honestly sorry that I can't recall which MS guy said it, or exactly how he phrased it, but the gist of it was that you really cannot tell exactly how much RAM/pagefile is being used by processes since some of the data being reported to you includes memory that is shared between processes, and some of the values only relate to virtual memory space allocated. I just remember reading this during one of the windows betas in MS's beta newsgroups, possibly the initial release of XP. I imagine it was in the 'performance' group and came from one of the guys that deals with memory management. Yeah, yeah, anecdotal. But if you look through their various memory management/monitoring articles you get basically the same information. Most of the memory usage values reported are vritual memory stats, not real/physical memory, and even the value for the working set is 'dirtied' by shared memory. You can get a good enough idea from all the various reported values, yeah. But you can't get exact figures. Hell, they even changed task manager's title from mem usage to PF usage, and for good reason. Setting that option *does* reduce paging. I'm not disputing that. I only disputed that it is going to change the paging behaviour of applications. Clearly it will not, since applications and their memory don't fall under the umbrella of this setting, which only deals with the executive (NTExecutive I believe is the proper name) and the things which belong to it. i.e. kernel and drivers. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: The discussions on MSDN are not important here, as the process scheduler was made to report accurately for SP2(XP) and SP1(2k3). Amoung other things this is a requirement for DEP internals. Most of these docs can be found through a partner login. Sadly, the only public docs I've found so far are the exact same as you have just sent. The latter of which is an old bug we used to suffer frequently on our Citrix mainframes. (Ah, the reason he's been screwing with process and memory management!). N.B. Did ya miss MS's other two support sites? (just joking, I _love_ the way they make us trawl 5 different places for info ;-). -- None of these negate what I have said. In fact all they do is negate what you said about never changing from defaults, as here you will find MS suggesting that people look at these articles for their solutions (having been on the phone for hours to MS waiting for some id-10t (it's been a while) to point me to Q184419, despite being outdated and useless to the problems we ever encounter on the citrix platforms. The techs are never as good as the consultants. :-( Back to the point in hand, setting this option can reduce paging in certain instances where the data paged is stored in driver space - this is particularly more common in certian development scnearios. You CAN analyse these effects by careful observation of kernel and process running times, along with changes in memory deltas and other less important varaibles. Generally you will need a program in ring 0 to properly observe these things though, and never forget that everything you do on the system affects the system. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
well, I guess there would be some collateral damage done since there will be slightly less RAM available for the apps to use, since the executive isn't being paged anymore. So indirectly the amount of application paging would increase, though I don't think it would be enough of an impact to worry about. You may see some performance benefit from not paging the executive, but you may not either. All depends what you're doing. Probably the biggest thing you'd notice is if it is a machine you're actually using, and the UI would probably remain more responsive during high paging activity. On 8/15/05, Clayton Macleod [EMAIL PROTECTED] wrote: Setting that option *does* reduce paging. I'm not disputing that. I only disputed that it is going to change the paging behaviour of applications. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Clayton Macleod wrote: re: memory reporting, even the working set value you get from the 'performance monitor' doesn't really necessarily tell you how much RAM the process is actually using just for itself, because this value includes any shared memory, not just private memory. I'm honestly sorry that I can't recall which MS guy said it, or exactly how he phrased it, but the gist of it was that you really cannot tell exactly how much RAM/pagefile is being used by processes since some of the data being reported to you includes memory that is shared between processes, and some of the values only relate to virtual memory space allocated. The most common cause of misrepresentation is the fact that Windows pre-pages most data to prevent massive delays in freeing physical ram when necessary. Was it Wang that was talking about that? Might have been, and IIRC it was discussed during beta 2. I just remember reading this during one of the windows betas in MS's beta newsgroups, possibly the initial release of XP. I imagine it was in the 'performance' group and came from one of the guys that deals with memory management. Yeah, yeah, anecdotal. But if you look through their various memory management/monitoring articles you get basically the same information. Most of the memory usage values reported are vritual memory stats, not real/physical memory, and even the value for the working set is 'dirtied' by shared memory. You can get a good enough idea from all the various reported values, yeah. But you can't get exact figures. Hell, they even changed task manager's title from mem usage to PF usage, and for good reason. Just because taskmgr doesn't report accurately does not mean the system cannot account for all memory. Wang would be most upset (and probably out of a job) if this was the case. Complete enumeration of these values is costly however, which is why it's unecessary for taskmgr. It's almost more important to have the values provided anyway. Setting that option *does* reduce paging. I'm not disputing that. phew. I only disputed that it is going to change the paging behaviour of applications. It does, because all changes to paging rates will change paging rates of other applications too. Call it what you will, starvation, pre-tension, or any of the other terms that people have tried to use to coin the factor of side-effects within dynamic caching algorithms, but it's princliple is the same. If there is something that needs to be regularly accessed but is not regularly scheduled it can cause failures (well, fail is too strong a word, but you know.) in the algorithms which can be reduced by changing their run-time settings. This is simply what happens in this scenario. Never underestimate how active the kernel and driver pages are! Clearly it will not, since applications and their memory don't fall under the umbrella of this setting, which only deals with the executive (NTExecutive I believe is the proper name) and the things which belong to it. i.e. kernel and drivers. Please don't try to tell me that the executive memory space is unimportant, I know that you already know this isn't true. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: The discussions on MSDN are not important here, as the process scheduler was made to report accurately for SP2(XP) and SP1(2k3). Amoung other things this is a requirement for DEP internals. Most of these docs can be found through a partner login. Sadly, the only public docs I've found so far are the exact same as you have just sent. The latter of which is an old bug we used to suffer frequently on our Citrix mainframes. (Ah, the reason he's been screwing with process and memory management!). N.B. Did ya miss MS's other two support sites? (just joking, I _love_ the way they make us trawl 5 different places for info ;-). -- None of these negate what I have said. In fact all they do is negate what you said about never changing from defaults, as here you will find MS suggesting that people look at these articles for their solutions (having been on the phone for hours to MS waiting for some id-10t (it's been a while) to point me to Q184419, despite being outdated and useless to the problems we ever encounter on the citrix platforms. The techs are never as good as the consultants. :-( Back to the point in hand, setting this option can reduce paging in certain instances where the data paged is stored in driver space - this is particularly more common in certian development scnearios. You CAN analyse these effects by careful observation of kernel and process running times, along with changes in memory deltas and other less important varaibles. Generally you will need a program in ring 0 to properly observe these things though, and never forget that everything you do on the system affects the system. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit:
Re: [hlds] windows 2003 memory tweaking
Been too long I guess, I don't recall. For some reason my memory seems to do better with the meat than with the potatoes. Yeah, I addressed the collateral stuff in my next post there. I'm not saying the executive memory space is unimportant. I only said that applications' memory space isn't the executive's memory space. Because it's not, they're definitely seperate from each other, and handled seperately/differently. I simply stated that this setting doesn't affect applications and the paging of their memory. Perhaps it would've been more accurate to say that this setting doesn't *directly* affect applications and their paging activity. Since the only way it affects it is by the fact that if you disable the executive's paging you are left with a smaller amount of RAM that could be available to applications. And indirectly this could/would change the amount of paging the applications experience. But windows comes with a default setting to allow the executive's memory space to be paged out for a reason, because portions of it can be inactive enough to warrant paging it out. No reason for my scanner driver to be in RAM when the scanner hasn't been used or even looked at since bootup, for instance. And allowing the executive to be paged would likely mean that driver would indeed be paged out whenever the OS could use that RAM elsewhere. Even if it's only a few dozen k. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: The most common cause of misrepresentation is the fact that Windows pre-pages most data to prevent massive delays in freeing physical ram when necessary. Was it Wang that was talking about that? Might have been, and IIRC it was discussed during beta 2. Just because taskmgr doesn't report accurately does not mean the system cannot account for all memory. Wang would be most upset (and probably out of a job) if this was the case. Complete enumeration of these values is costly however, which is why it's unecessary for taskmgr. It's almost more important to have the values provided anyway. phew. It does, because all changes to paging rates will change paging rates of other applications too. Call it what you will, starvation, pre-tension, or any of the other terms that people have tried to use to coin the factor of side-effects within dynamic caching algorithms, but it's princliple is the same. If there is something that needs to be regularly accessed but is not regularly scheduled it can cause failures (well, fail is too strong a word, but you know.) in the algorithms which can be reduced by changing their run-time settings. This is simply what happens in this scenario. Never underestimate how active the kernel and driver pages are! Please don't try to tell me that the executive memory space is unimportant, I know that you already know this isn't true. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Minor interjection.. Did I miss understand the whole point of mixing Virtual and Physical ram? was it not the point of the OS telling the App where and how to get memory, and if I am not mistaken, Apps don't even directly touch ram, that little interface called HAL handles it, so as far as the App is concerned it's all physical ram. So back to the point, static setup for a static need, it's not like were going to have a melt down if you can't lunch solitaire on the server that's meant for serving... oh yes the MS KB, how I miss the internal one, if only I was allowed a burner while I was employed under the PSS outsource :'( From: Clayton Macleod [EMAIL PROTECTED] Reply-To: hlds@list.valvesoftware.com To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Date: Mon, 15 Aug 2005 05:25:53 -0700 Been too long I guess, I don't recall. For some reason my memory seems to do better with the meat than with the potatoes. Yeah, I addressed the collateral stuff in my next post there. I'm not saying the executive memory space is unimportant. I only said that applications' memory space isn't the executive's memory space. Because it's not, they're definitely seperate from each other, and handled seperately/differently. I simply stated that this setting doesn't affect applications and the paging of their memory. Perhaps it would've been more accurate to say that this setting doesn't *directly* affect applications and their paging activity. Since the only way it affects it is by the fact that if you disable the executive's paging you are left with a smaller amount of RAM that could be available to applications. And indirectly this could/would change the amount of paging the applications experience. But windows comes with a default setting to allow the executive's memory space to be paged out for a reason, because portions of it can be inactive enough to warrant paging it out. No reason for my scanner driver to be in RAM when the scanner hasn't been used or even looked at since bootup, for instance. And allowing the executive to be paged would likely mean that driver would indeed be paged out whenever the OS could use that RAM elsewhere. Even if it's only a few dozen k. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: The most common cause of misrepresentation is the fact that Windows pre-pages most data to prevent massive delays in freeing physical ram when necessary. Was it Wang that was talking about that? Might have been, and IIRC it was discussed during beta 2. Just because taskmgr doesn't report accurately does not mean the system cannot account for all memory. Wang would be most upset (and probably out of a job) if this was the case. Complete enumeration of these values is costly however, which is why it's unecessary for taskmgr. It's almost more important to have the values provided anyway. phew. It does, because all changes to paging rates will change paging rates of other applications too. Call it what you will, starvation, pre-tension, or any of the other terms that people have tried to use to coin the factor of side-effects within dynamic caching algorithms, but it's princliple is the same. If there is something that needs to be regularly accessed but is not regularly scheduled it can cause failures (well, fail is too strong a word, but you know.) in the algorithms which can be reduced by changing their run-time settings. This is simply what happens in this scenario. Never underestimate how active the kernel and driver pages are! Please don't try to tell me that the executive memory space is unimportant, I know that you already know this isn't true. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Clayton Macleod wrote: Been too long I guess, I don't recall. For some reason my memory seems to do better with the meat than with the potatoes. Yeah know that problem, last time I touched sendmail was nearly 5 years ago now, and I looked at the config the other day in shock. :) Yeah, I addressed the collateral stuff in my next post there. I'm not saying the executive memory space is unimportant. I only said that applications' memory space isn't the executive's memory space. Because it's not, they're definitely seperate from each other, and handled seperately/differently. I simply stated that this setting doesn't affect applications and the paging of their memory. Perhaps it would've been more accurate to say that this setting doesn't *directly* affect applications and their paging activity. Since the only way it affects it is by the fact that if you disable the executive's paging you are left with a smaller amount of RAM that could be available to applications. And indirectly this could/would change the amount of paging the applications experience. But windows comes with a default setting to allow the executive's memory space to be paged out for a reason, because portions of it can be inactive enough to warrant paging it out. No reason for my scanner driver to be in RAM when the scanner hasn't been used or even looked at since bootup, for instance. This is actually explicitly controllable by the driver, as you will see in other documentation. As you probably know from the discussions regarding reporting of memory usage, windows pages most of most applications to save time when clearing physical ram, if the applications are using most of the ram, this means that lesser scheduled apps will not recieve any priority in the time based components of paging controls - a problem common to most paging systems. This is also the reason that they recommend that your pagefile size is 1 to 1.5x your physical ram or greater - it prevents excess page dumps from ram to disk in order to clear physical space. This action is important, as I have stated before, for a dynamic environment where the running applications and average memory deltas are high. For systems where the application set will fit in physical ram paging is unnecessary and does use excess time on the processor and the bus. The important notice is page faults and the page fault deltas. It's not uncommon in the near extremes of these scenarios that excess page faults can be reduced by setting non defaults. The specific cases I have dealt with (typically high end machines and very large single process applications, or many many user environments) have benefited from this setting, by side effect or not. Other things, such as IOPageLockLimit can also be important, but are often mis-set by users trying to optimise in these areas. And allowing the executive to be paged would likely mean that driver would indeed be paged out whenever the OS could use that RAM elsewhere. Even if it's only a few dozen k. In general server machines should not suffer the problems associated with having excess hardware. Ideally, un-used hardware should be disabled. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
well, the point there wasn't that servers are going to have superfluous hardware. The point was that pretty much nothing is going to cause every single page used by the executive to be active enough to keep it all in ram at all times. Given that this is likely to be way under a hundred megs, way way under on a server, and that you're likely to have a gig or two of ram, keeping it all in ram probably isn't going to hurt much. And while I see your point about a single-application server only doing so much, I kinda doubt that having the executive eligible for paging is going to affect it that much either. Perhaps the biggest chance for any performance differences to arise here would be at map change, and that's actually a point when performance doesn't really matter much at all. Gameplay isn't going to be affected. Another freakin' all-nighter, whodathunkit, haha...6:35am here, gonna get some shuteye before I decide whether or not to put off working on the car after lunch. Which I guess will technically be breakfast, heh. Have a good one. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: In general server machines should not suffer the problems associated with having excess hardware. Ideally, un-used hardware should be disabled. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
yeah, basically apps just deal with virtual memory, and the OS is what deals between RAM and the pagefile. Each app gets its own 2gigs of address space to allocate memory for itself from, and the OS is what determines the actual RAM/pagefile usage. On 8/15/05, Dustin Tuft [EMAIL PROTECTED] wrote: Minor interjection.. Did I miss understand the whole point of mixing Virtual and Physical ram? was it not the point of the OS telling the App where and how to get memory, and if I am not mistaken, Apps don't even directly touch ram, that little interface called HAL handles it, so as far as the App is concerned it's all physical ram. So back to the point, static setup for a static need, it's not like were going to have a melt down if you can't lunch solitaire on the server that's meant for serving... oh yes the MS KB, how I miss the internal one, if only I was allowed a burner while I was employed under the PSS outsource :'( -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
RE: [hlds] windows 2003 memory tweaking
Your HLDS and SRCDS work don't they? STFU and go join a chatroom somewhere or shack up as roommates. Either way STFU. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Clayton Macleod Sent: Monday, August 15, 2005 9:42 AM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking yeah, basically apps just deal with virtual memory, and the OS is what deals between RAM and the pagefile. Each app gets its own 2gigs of address space to allocate memory for itself from, and the OS is what determines the actual RAM/pagefile usage. On 8/15/05, Dustin Tuft [EMAIL PROTECTED] wrote: Minor interjection.. Did I miss understand the whole point of mixing Virtual and Physical ram? was it not the point of the OS telling the App where and how to get memory, and if I am not mistaken, Apps don't even directly touch ram, that little interface called HAL handles it, so as far as the App is concerned it's all physical ram. So back to the point, static setup for a static need, it's not like were going to have a melt down if you can't lunch solitaire on the server that's meant for serving... oh yes the MS KB, how I miss the internal one, if only I was allowed a burner while I was employed under the PSS outsource :'( -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
On 8/15/05, Dustin Tuft [EMAIL PROTECTED] wrote: Minor interjection.. Did I miss understand the whole point of mixing Virtual and Physical ram? was it not the point of the OS telling the App where and how to get memory, and if I am not mistaken, Apps don't even directly touch ram, that little interface called HAL handles it, so as far as the App is concerned it's all physical ram. Yes, but changes in the environment affect the system in a particular way, you understand. The app controls merely the volume of memory it allocates (and most of the time de-allocs ;-p). Nothing contrary to this has been discussed by either of us. So back to the point, static setup for a static need, it's not like were going to have a melt down if you can't lunch solitaire on the server that's meant for serving... Well that's exactly it, also the memory manager is very creative in these situations. Don't you just love the yellow boxes. oh yes the MS KB, how I miss the internal one, if only I was allowed a burner while I was employed under the PSS outsource :'( I lost access to a gold partner and mvp login, not everything though. n'er mind. From: Clayton Macleod [EMAIL PROTECTED] Reply-To: hlds@list.valvesoftware.com To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Date: Mon, 15 Aug 2005 05:25:53 -0700 Been too long I guess, I don't recall. For some reason my memory seems to do better with the meat than with the potatoes. Yeah, I addressed the collateral stuff in my next post there. I'm not saying the executive memory space is unimportant. I only said that applications' memory space isn't the executive's memory space. Because it's not, they're definitely seperate from each other, and handled seperately/differently. I simply stated that this setting doesn't affect applications and the paging of their memory. Perhaps it would've been more accurate to say that this setting doesn't *directly* affect applications and their paging activity. Since the only way it affects it is by the fact that if you disable the executive's paging you are left with a smaller amount of RAM that could be available to applications. And indirectly this could/would change the amount of paging the applications experience. But windows comes with a default setting to allow the executive's memory space to be paged out for a reason, because portions of it can be inactive enough to warrant paging it out. No reason for my scanner driver to be in RAM when the scanner hasn't been used or even looked at since bootup, for instance. And allowing the executive to be paged would likely mean that driver would indeed be paged out whenever the OS could use that RAM elsewhere. Even if it's only a few dozen k. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: The most common cause of misrepresentation is the fact that Windows pre-pages most data to prevent massive delays in freeing physical ram when necessary. Was it Wang that was talking about that? Might have been, and IIRC it was discussed during beta 2. Just because taskmgr doesn't report accurately does not mean the system cannot account for all memory. Wang would be most upset (and probably out of a job) if this was the case. Complete enumeration of these values is costly however, which is why it's unecessary for taskmgr. It's almost more important to have the values provided anyway. phew. It does, because all changes to paging rates will change paging rates of other applications too. Call it what you will, starvation, pre-tension, or any of the other terms that people have tried to use to coin the factor of side-effects within dynamic caching algorithms, but it's princliple is the same. If there is something that needs to be regularly accessed but is not regularly scheduled it can cause failures (well, fail is too strong a word, but you know.) in the algorithms which can be reduced by changing their run-time settings. This is simply what happens in this scenario. Never underestimate how active the kernel and driver pages are! Please don't try to tell me that the executive memory space is unimportant, I know that you already know this isn't true. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
On 8/15/05, Clayton Macleod [EMAIL PROTECTED] wrote: well, the point there wasn't that servers are going to have superfluous hardware. The point was that pretty much nothing is going to cause every single page used by the executive to be active enough to keep it all in ram at all times. Ah so you agree with me there. What can we do to keep page fault latency down then? Given that this is likely to be way under a hundred megs, way way under on a server, and that you're likely to have a gig or two of ram, keeping it all in ram probably isn't going to hurt much. Indeed. :) And while I see your point about a single-application server only doing so much, I kinda doubt that having the executive eligible for paging is going to affect it that much either. In latency terms, we're talking about real-time importance. All latency added to the system for processing, particularly in IDE, are going to be significant to the next frame. Smoothing this out reduces jitter, or more definably reduces the standard deviation of server performance variables. Perhaps the biggest chance for any performance differences to arise here would be at map change, and that's actually a point when performance doesn't really matter much at all. Gameplay isn't going to be affected. Another freakin' all-nighter, whodathunkit, haha...6:35am here, gonna get some shuteye before I decide whether or not to put off working on the car after lunch. Which I guess will technically be breakfast, heh. Have a good one. Just out of interest with regard to this stuff, anyone started up an HLDS or SRCDS instance on QNX? ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
nah, that's ok. On 8/15/05, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: Your HLDS and SRCDS work don't they? STFU and go join a chatroom somewhere or shack up as roommates. Either way STFU. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
well, I'm not sure we're saying the exact same thing. The point I was trying to make was, that while a great portion of the executive's memory is going to be active all the time, and stay in RAM, a portion of it will also go unused and be candidates for the pagefile. I said I doubt that something is going to access every single page of the executive's memory. I didn't say that I thought paging activity of the executive's memory was going to be rampant, which I think is what you're saying. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: Ah so you agree with me there. What can we do to keep page fault latency down then? Indeed. :) In latency terms, we're talking about real-time importance. All latency added to the system for processing, particularly in IDE, are going to be significant to the next frame. Smoothing this out reduces jitter, or more definably reduces the standard deviation of server performance variables. Just out of interest with regard to this stuff, anyone started up an HLDS or SRCDS instance on QNX? -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Gold huh, I wonder what it was missing as far as internal comments. Did you ever get the red sections of text that were source code? I can't even count how many stupid times I got that talk on we do not release or copy red source out of KB articles. I mean realy, like who I am going to take it to and with out the complete code what real good was a few lines? Sorry walk down memory lane terminated (paged). From: James Tucker [EMAIL PROTECTED] Reply-To: hlds@list.valvesoftware.com To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Date: Mon, 15 Aug 2005 15:39:02 +0100 On 8/15/05, Dustin Tuft [EMAIL PROTECTED] wrote: Minor interjection.. Did I miss understand the whole point of mixing Virtual and Physical ram? was it not the point of the OS telling the App where and how to get memory, and if I am not mistaken, Apps don't even directly touch ram, that little interface called HAL handles it, so as far as the App is concerned it's all physical ram. Yes, but changes in the environment affect the system in a particular way, you understand. The app controls merely the volume of memory it allocates (and most of the time de-allocs ;-p). Nothing contrary to this has been discussed by either of us. So back to the point, static setup for a static need, it's not like were going to have a melt down if you can't lunch solitaire on the server that's meant for serving... Well that's exactly it, also the memory manager is very creative in these situations. Don't you just love the yellow boxes. oh yes the MS KB, how I miss the internal one, if only I was allowed a burner while I was employed under the PSS outsource :'( I lost access to a gold partner and mvp login, not everything though. n'er mind. From: Clayton Macleod [EMAIL PROTECTED] Reply-To: hlds@list.valvesoftware.com To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Date: Mon, 15 Aug 2005 05:25:53 -0700 Been too long I guess, I don't recall. For some reason my memory seems to do better with the meat than with the potatoes. Yeah, I addressed the collateral stuff in my next post there. I'm not saying the executive memory space is unimportant. I only said that applications' memory space isn't the executive's memory space. Because it's not, they're definitely seperate from each other, and handled seperately/differently. I simply stated that this setting doesn't affect applications and the paging of their memory. Perhaps it would've been more accurate to say that this setting doesn't *directly* affect applications and their paging activity. Since the only way it affects it is by the fact that if you disable the executive's paging you are left with a smaller amount of RAM that could be available to applications. And indirectly this could/would change the amount of paging the applications experience. But windows comes with a default setting to allow the executive's memory space to be paged out for a reason, because portions of it can be inactive enough to warrant paging it out. No reason for my scanner driver to be in RAM when the scanner hasn't been used or even looked at since bootup, for instance. And allowing the executive to be paged would likely mean that driver would indeed be paged out whenever the OS could use that RAM elsewhere. Even if it's only a few dozen k. On 8/15/05, James Tucker [EMAIL PROTECTED] wrote: The most common cause of misrepresentation is the fact that Windows pre-pages most data to prevent massive delays in freeing physical ram when necessary. Was it Wang that was talking about that? Might have been, and IIRC it was discussed during beta 2. Just because taskmgr doesn't report accurately does not mean the system cannot account for all memory. Wang would be most upset (and probably out of a job) if this was the case. Complete enumeration of these values is costly however, which is why it's unecessary for taskmgr. It's almost more important to have the values provided anyway. phew. It does, because all changes to paging rates will change paging rates of other applications too. Call it what you will, starvation, pre-tension, or any of the other terms that people have tried to use to coin the factor of side-effects within dynamic caching algorithms, but it's princliple is the same. If there is something that needs to be regularly accessed but is not regularly scheduled it can cause failures (well, fail is too strong a word, but you know.) in the algorithms which can be reduced by changing their run-time settings. This is simply what happens in this scenario. Never underestimate how active the kernel and driver pages are! Please don't try to tell me that the executive memory space is unimportant, I know that you already know this isn't true. -- Clayton Macleod get ye flask You cannot get ye flask
Re: [hlds] windows 2003 memory tweaking
Scott Tuttle wrote: One simple way of forcing Windows to use more ram and less hard drive for paging is to set you page file size and do not allow windows to manage it's own file size. But do keep in mind windows is designed to page data that is not needed out of RAM into the page file. Restricting your page file could set off a bit of thrashing. This is a bad idea. Windows is not designed to page out data that is not needed from RAM. It is designed to allocate space in the page file for running applications regardless of the amount of ram on the system. If you start a program it will automatically allocate space in the page file equivalent TO allow the storage of that application in the page file system only incase it needs to do so. You should always have a page file that is the equivilant size of the amount of memory you have even if you dont expect windows to use the page file for actually paging an application. You should read this http://support.microsoft.com/default.aspx?scid=kb;en-us;555223 These rulings are intended for systems with dynamic environments and are done for coninual performance puproses, and built for dynamic run-time environments. This is not the scenario that a dedicated box will be running, which much more resembles a static application set. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
MS themselves say that there is no way for us to know how much of any process is in RAM and how much is in the pagefile, so, I don't know where you're getting this information from. Besides, *everything* is in virtual memory. And there is no point in keeping pages in RAM that aren't being accessed when that RAM could be actively used for some other purpose. If pages are active, they'll be in RAM, don't worry about it. On 8/12/05, dexion [EMAIL PROTECTED] wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
There are servel paths to manage memory. Most of these paths are not published due to the prue fact the MS does not want to end up supporting a bunce of wet needles that are tring to over perform there OS so there solitare runs better. Please keep in mind that any thing you read from Microsoft is designed to lower impact on support, so they will never suggest you mess with a setting that is potential to cause the avagrage consumer to crash there system, espacialy when that can cause a problem loging in, forcing you into safe mode to correct the problem. And as we all know it's no fun talking Granma throught process of F8 on startup. Dustin - Original Message - From: Clayton Macleod [EMAIL PROTECTED] To: hlds@list.valvesoftware.com Sent: Sunday, August 14, 2005 7:19 AM Subject: Re: [hlds] windows 2003 memory tweaking I think perhaps a refresher is warranted here, since you're forgetting one basic fact, which is even mentioned in the article listed. Virtual Memory is always in use, even when the memory required by all running processes does not exceed the amount of RAM installed on the system. You should always have a pagefile, and it should be larger than your installed RAM, even if you have tons of RAM and your memory usage never nears your RAM limit. The people that wrote the memory manager in the OS are telling you this, and you're thinking they don't know what they're talking about? Besides, with HD space being well under $1/GB there's no legitimate reason to complain about the wasted HD space for a pagefile. Just set the minimum size On 8/14/05, James Tucker [EMAIL PROTECTED] wrote: These rulings are intended for systems with dynamic environments and are done for coninual performance puproses, and built for dynamic run-time environments. This is not the scenario that a dedicated box will be running, which much more resembles a static application set. -- Clayton Macleod get ye flask You cannot get ye flask. ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Thats up to you, it will do no permanent damage, and unless IOPageLockLimit is set to a completely arbritrary value it will also remain stable. Unfortunately there isn't at this point a completely definable optimum, it is dependant on your hardware and software configuration and you will need to test, essentially for, each new norhtbridge, as bus latencies are the killer here. dexion wrote: it seems to handle things very well, i might want to just leave well enough alone. Waste of ram though -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Friday, August 12, 2005 1:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Dunno, but it's easy enough to change the session manager options HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management In particular - the large system cache option will lift more of the kernel into ram - this may or may not be advantageous - look at page faults for your processes - might be a good idea to get pmon from systeminternals. disablepagingexecutive will page less data during general system run time, this is not always an advantage as windows dynamic paging algorithm is quite strong. there is another option you can add here to change your IOPageLockLimit (key not there by default), this allows you to specify the amount of ram windows will allow for IO paging operations. dexion wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
RE: [hlds] windows 2003 memory tweaking
Right, I do plan to have a test server to see any benefit if at all. Odd, no one commented on the -heapsize question. dex -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Saturday, August 13, 2005 9:03 AM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Thats up to you, it will do no permanent damage, and unless IOPageLockLimit is set to a completely arbritrary value it will also remain stable. Unfortunately there isn't at this point a completely definable optimum, it is dependant on your hardware and software configuration and you will need to test, essentially for, each new norhtbridge, as bus latencies are the killer here. dexion wrote: it seems to handle things very well, i might want to just leave well enough alone. Waste of ram though -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Friday, August 12, 2005 1:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Dunno, but it's easy enough to change the session manager options HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management In particular - the large system cache option will lift more of the kernel into ram - this may or may not be advantageous - look at page faults for your processes - might be a good idea to get pmon from systeminternals. disablepagingexecutive will page less data during general system run time, this is not always an advantage as windows dynamic paging algorithm is quite strong. there is another option you can add here to change your IOPageLockLimit (key not there by default), this allows you to specify the amount of ram windows will allow for IO paging operations. dexion wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
It does seem like a waste of RAM, but that's how Windows manages resources. One thing to consider, Windows might be levearging your hard drive more becasue of it's speed ablities. What type of hard drive(s) are you running on your server? A good S ATA (while not nearly as fast as RAM) drive with an 8 meg buffer could be the reason your not seeing a performance lag in your server. One simple way of forcing Windows to use more ram and less hard drive for paging is to set you page file size and do not allow windows to manage it's own file size. But do keep in mind windows is designed to page data that is not needed out of RAM into the page file. Restricting your page file could set off a bit of thrashing. If you want to realy out do your self you could setup a RAM drive and tell windows to use that drive to page to, but this was more of a prank pulled on co-workers :). It not that stabile and a great laugh to see the confused look on someone face when they terminat that program using a mass amount of ram that doesn't seem to be doing anything. Dustin Tuft - Original Message - From: dexion [EMAIL PROTECTED] To: hlds@list.valvesoftware.com Sent: Saturday, August 13, 2005 7:54 AM Subject: RE: [hlds] windows 2003 memory tweaking Right, I do plan to have a test server to see any benefit if at all. Odd, no one commented on the -heapsize question. dex -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Saturday, August 13, 2005 9:03 AM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Thats up to you, it will do no permanent damage, and unless IOPageLockLimit is set to a completely arbritrary value it will also remain stable. Unfortunately there isn't at this point a completely definable optimum, it is dependant on your hardware and software configuration and you will need to test, essentially for, each new norhtbridge, as bus latencies are the killer here. dexion wrote: it seems to handle things very well, i might want to just leave well enough alone. Waste of ram though -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Friday, August 12, 2005 1:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Dunno, but it's easy enough to change the session manager options HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management In particular - the large system cache option will lift more of the kernel into ram - this may or may not be advantageous - look at page faults for your processes - might be a good idea to get pmon from systeminternals. disablepagingexecutive will page less data during general system run time, this is not always an advantage as windows dynamic paging algorithm is quite strong. there is another option you can add here to change your IOPageLockLimit (key not there by default), this allows you to specify the amount of ram windows will allow for IO paging operations. dexion wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list
RE: [hlds] windows 2003 memory tweaking
they are all 15000 rpm scsi drives in a raid 1 array they are quite fast but no where near as fast as ram. I think ill not tinker with what workes fine, just equip the servers in the future with 1500megs of ram instead. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Dustin Sent: Saturday, August 13, 2005 12:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking It does seem like a waste of RAM, but that's how Windows manages resources. One thing to consider, Windows might be levearging your hard drive more becasue of it's speed ablities. What type of hard drive(s) are you running on your server? A good S ATA (while not nearly as fast as RAM) drive with an 8 meg buffer could be the reason your not seeing a performance lag in your server. One simple way of forcing Windows to use more ram and less hard drive for paging is to set you page file size and do not allow windows to manage it's own file size. But do keep in mind windows is designed to page data that is not needed out of RAM into the page file. Restricting your page file could set off a bit of thrashing. If you want to realy out do your self you could setup a RAM drive and tell windows to use that drive to page to, but this was more of a prank pulled on co-workers :). It not that stabile and a great laugh to see the confused look on someone face when they terminat that program using a mass amount of ram that doesn't seem to be doing anything. Dustin Tuft - Original Message - From: dexion [EMAIL PROTECTED] To: hlds@list.valvesoftware.com Sent: Saturday, August 13, 2005 7:54 AM Subject: RE: [hlds] windows 2003 memory tweaking Right, I do plan to have a test server to see any benefit if at all. Odd, no one commented on the -heapsize question. dex -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Saturday, August 13, 2005 9:03 AM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Thats up to you, it will do no permanent damage, and unless IOPageLockLimit is set to a completely arbritrary value it will also remain stable. Unfortunately there isn't at this point a completely definable optimum, it is dependant on your hardware and software configuration and you will need to test, essentially for, each new norhtbridge, as bus latencies are the killer here. dexion wrote: it seems to handle things very well, i might want to just leave well enough alone. Waste of ram though -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Friday, August 12, 2005 1:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Dunno, but it's easy enough to change the session manager options HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management In particular - the large system cache option will lift more of the kernel into ram - this may or may not be advantageous - look at page faults for your processes - might be a good idea to get pmon from systeminternals. disablepagingexecutive will page less data during general system run time, this is not always an advantage as windows dynamic paging algorithm is quite strong. there is another option you can add here to change your IOPageLockLimit (key not there by default), this allows you to specify the amount of ram windows will allow for IO paging operations. dexion wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman
Re: [hlds] windows 2003 memory tweaking
Nice, If you have two arrays, you can improve system performance by moving the Page file to the array that the OS is not on. But this only helps if the arrays are truely on diffrenet channels. Same would would go for a normal IDE drive - Original Message - From: dexion [EMAIL PROTECTED] To: hlds@list.valvesoftware.com Sent: Saturday, August 13, 2005 10:14 AM Subject: RE: [hlds] windows 2003 memory tweaking they are all 15000 rpm scsi drives in a raid 1 array they are quite fast but no where near as fast as ram. I think ill not tinker with what workes fine, just equip the servers in the future with 1500megs of ram instead. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of Dustin Sent: Saturday, August 13, 2005 12:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking It does seem like a waste of RAM, but that's how Windows manages resources. One thing to consider, Windows might be levearging your hard drive more becasue of it's speed ablities. What type of hard drive(s) are you running on your server? A good S ATA (while not nearly as fast as RAM) drive with an 8 meg buffer could be the reason your not seeing a performance lag in your server. One simple way of forcing Windows to use more ram and less hard drive for paging is to set you page file size and do not allow windows to manage it's own file size. But do keep in mind windows is designed to page data that is not needed out of RAM into the page file. Restricting your page file could set off a bit of thrashing. If you want to realy out do your self you could setup a RAM drive and tell windows to use that drive to page to, but this was more of a prank pulled on co-workers :). It not that stabile and a great laugh to see the confused look on someone face when they terminat that program using a mass amount of ram that doesn't seem to be doing anything. Dustin Tuft - Original Message - From: dexion [EMAIL PROTECTED] To: hlds@list.valvesoftware.com Sent: Saturday, August 13, 2005 7:54 AM Subject: RE: [hlds] windows 2003 memory tweaking Right, I do plan to have a test server to see any benefit if at all. Odd, no one commented on the -heapsize question. dex -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Saturday, August 13, 2005 9:03 AM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Thats up to you, it will do no permanent damage, and unless IOPageLockLimit is set to a completely arbritrary value it will also remain stable. Unfortunately there isn't at this point a completely definable optimum, it is dependant on your hardware and software configuration and you will need to test, essentially for, each new norhtbridge, as bus latencies are the killer here. dexion wrote: it seems to handle things very well, i might want to just leave well enough alone. Waste of ram though -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Friday, August 12, 2005 1:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Dunno, but it's easy enough to change the session manager options HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management In particular - the large system cache option will lift more of the kernel into ram - this may or may not be advantageous - look at page faults for your processes - might be a good idea to get pmon from systeminternals. disablepagingexecutive will page less data during general system run time, this is not always an advantage as windows dynamic paging algorithm is quite strong. there is another option you can add here to change your IOPageLockLimit (key not there by default), this allows you to specify the amount of ram windows will allow for IO paging operations. dexion wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com
Re: [hlds] windows 2003 memory tweaking
One simple way of forcing Windows to use more ram and less hard drive for paging is to set you page file size and do not allow windows to manage it's own file size. But do keep in mind windows is designed to page data that is not needed out of RAM into the page file. Restricting your page file could set off a bit of thrashing. This is a bad idea. Windows is not designed to page out data that is not needed from RAM. It is designed to allocate space in the page file for running applications regardless of the amount of ram on the system. If you start a program it will automatically allocate space in the page file equivalent TO allow the storage of that application in the page file system only incase it needs to do so. You should always have a page file that is the equivilant size of the amount of memory you have even if you dont expect windows to use the page file for actually paging an application. You should read this http://support.microsoft.com/default.aspx?scid=kb;en-us;555223 ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
[hlds] windows 2003 memory tweaking
Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
Re: [hlds] windows 2003 memory tweaking
Dunno, but it's easy enough to change the session manager options HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management In particular - the large system cache option will lift more of the kernel into ram - this may or may not be advantageous - look at page faults for your processes - might be a good idea to get pmon from systeminternals. disablepagingexecutive will page less data during general system run time, this is not always an advantage as windows dynamic paging algorithm is quite strong. there is another option you can add here to change your IOPageLockLimit (key not there by default), this allows you to specify the amount of ram windows will allow for IO paging operations. dexion wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
RE: [hlds] windows 2003 memory tweaking
it seems to handle things very well, i might want to just leave well enough alone. Waste of ram though -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] Behalf Of James Tucker Sent: Friday, August 12, 2005 1:05 PM To: hlds@list.valvesoftware.com Subject: Re: [hlds] windows 2003 memory tweaking Dunno, but it's easy enough to change the session manager options HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management In particular - the large system cache option will lift more of the kernel into ram - this may or may not be advantageous - look at page faults for your processes - might be a good idea to get pmon from systeminternals. disablepagingexecutive will page less data during general system run time, this is not always an advantage as windows dynamic paging algorithm is quite strong. there is another option you can add here to change your IOPageLockLimit (key not there by default), this allows you to specify the amount of ram windows will allow for IO paging operations. dexion wrote: Hi all, Just wondering about something. I have all windows2003 boxes and run multiple instances of hlds and srcds. I notice for instance that if I run x bumber of servers I am using 1.6 gigs of ram total, but only about 1000 of that would be in the actual physical ram. The rest is is virtual. I have 2 gigs of ram per box. My question would be, is there any performance gain in trying to push more of the hlds/srcds process into physical ram? I dont really experience any problems, but I would like to use the ram as opposed the virtual since I have it why not use it if there is a benefit. Would the -heapsize switch help since I do not use it? Anyone out there know how to get windows to be more sparing with the virtual and push more into physical ram? Why would the os want to use any virtual since im under the physical ram in the first place? Seems silly. tia dex ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds ___ To unsubscribe, edit your list preferences, or view the list archives, please visit: http://list.valvesoftware.com/mailman/listinfo/hlds
