Using all memory (Win-XP SP-2)

[CWBillow]

Back in the "early days" there was a way to configure your system so that it would use "all" your memory before using the swap file.

They seem to have gotten rid of this tweak in XP, probably in 2000 as well... but I miss it.

And a couple days ago, TechRepublic (?), Or O'reilly, -- somewhere I read/found the registry tweak that turned this feature back on in XP -- and it works great!

But -- I know, I know -- I didn't document my "find" right, and now it's a "lost".

Anybody with any idea, or maybe they saw the same article I did?

Regards,
Chuck Billow

[Jezza]

Dare I say , you are entering really dangerous ground here and warn you against even going down this road. I have seen whole operating systems hang just for the sake of a few micro-seconds

I refer you to another TechRepublic editorial

Avoid Tweaks of Dubious Value in Windows.

You have been warned

[Leif]

Just as a matter of interest, what does "it works great!" mean? I'm intrigued to know what benefits you have noticed!

[AlanMiller]

Chuck

It might be appropriate to read the thread I posted recently: The Truth about Windows Memory Optimizers
before embarking on tweaks of dubious benefit.

Alan

[CWBillow]

Edit:

Benefits: FAR fewer hangs, much quicker response.

CB
************************************************** ****************

Jerry / Leif / Alan:

As far as "Memory Optimizers" go, I would basically agree. I have tried a couple different ones in the past, with dubious results.

But this is not (IMPOV) an "optimizer" tweak. Rather it is putting back an "adjustment" available in earlier Windows additions: using "all" available physical memory before resorting to swap.

I have 2 gigs of memory on my machine, installed primarily to benefit the use of virtual machines. What I was seeing however, was that Windows was swapping out programs when I still had over half (1 gig) of my memory still available.

This registry entry ("somehow") causes this to be far different. I see an increase to as much as 85% of my memory... and that's without trying as yet to exploit it. And there's little doubt that my system response is far better than it was before relying more heavily on swapping.

Right now I'm safe, 'cause I had made the entry, and it's working fine. I just need to find where I stuck the darn notes on same in case I need to re-do my setup -- although we all know that that seldom happens to me!

And this didn't come from some unknown source, but rather one of the reviewed and "reputable" sources.

I'll post when I find same.

Regards,
Chuck Billow

[HansV]

> Sorry, the link you clicked on may be outdated or incorrect. Please contact customer support to report this. Thank you!

Apart from that, such recommendations are worthless, Chuck. Recent versions of Windows have excellent memory management, it's best to let it handle everything itself instead of trying to interfere.

[CWBillow]

Guys:

Here's a piece: Max Memory Use .

May be more... I'll see if I can find it...

Chuck

[CWBillow]

hans:

Sorry, I've never quite gotten he hang of that link tag... I had the 'URL' still there... 's okay now.

>>
Apart from that, such recommendations are worthless, Chuck. Recent versions of Windows have excellent memory management, it's best to let it handle everything itself instead of trying to interfere.
<<
Sorry to hear THAT, as it seems to help me quite...

Chuck

[bigaldoc]

In the hopes that I can say this without causing you insult Chuck, I'll add a comment about the conservatism you're hearing from many of us. One of the most difficult aspects of helping someone with a computer problem, whether in person or remotely, is the "habitual tweaker." I do believe this is one of the reasons that computer supplier's support people are so quick to resort to the "...get out your restoration CD and re-load the computer..." The tweaking syndrome is most significant when it comes to the core operation of the OS, rather than installed applications. When a computer user has made a number of modifications to the OS core, and suddenly starts having a problem, it's often impossible to tell the source of the problem. It often is the result of something that the user did four or five "tweaks" ago.

From a conservative like me (and many other Loungers) you'll usually hear that the few nanoseconds of speed improvement are just not worth the risk of the possible peripheral problems. Now, that's not 100% rock-solid - I've made many tweaks in my time at the keyboard, but I think long and hard about the ones that pronounce themselves to be "the best speed improvement" since sliced bread. OK, off my soapbox now...

[JoeP517]

I'll pile on a bit. See TweakHound's Super XP Tweaking Guide - Bad Tweaks. Unless you have definitive benchmarks for YOUR system of the effects I'd strongly suggest that you ignore everything on the Speed Up Your Windows PC page after the 'Make your menus load faster' tip. Also, see LangaList Std Edition 2004-12-16 for a short discussion of the ConservativeSwapFile setting and Virtual Memory in Windows XP for an excellent article discussing XP virtual memory.

Joe

[CWBillow]

Guys, you're all probably correct that I often (it's OK Al) "over-tweak". This can and has cost me a few terrifying moments, for sure. I can't help but admit though, that also on several occasions I have had a true "Oh, WOW!" Moment as a result of some of my tweaking. I suppose I've (almost always) concluded that the thrill of discovery was worth the jitters.

I can't help but constantly wonder why third-party developers are constantly coming up with solutions and improvements that could very easily have been addressed by the major players like IBM and HP from the very beginning. Part of the reason cold be the empirical reality in which we operate.

Why else would it have been Bill Gates himself who said that there was no reason why anyone would ever need more than 640 megs of memory?

I do now, however, back up my machine pretty darn often -- although that too has resulted in experimenting with different methods and programs.

I also now (try) to make such changes one at a time, so that I can, as suggested, get a better picture of their end result.

In this particular instance, I cannot help but think, as a result of assumed arithmetic and perceived gains, that when I have some 2 gigs of memory available, and the machine only chooses to avail itself of some 600 megs, that this is silly and wasteful; if there was (and is) a way to point operations in the direction of waiting a bit longer before dumping memory, this should only improve ops, as memory is a good deal more responsive than disk swapping.

What I should undoubtedly do is have a separate machine for my screwing around... that not currently being feasible, I work with what I've got.

The above all having been said, I'm going to take a closer look at the links you've provided, my own expected gains for worthiness assessment and any possible changes in my processes.

I know my questions can often be meandering, and for that I apologize. The (very few) times I have posted and received little or no response I assumed that part of that resulted from the worthiness of the question asked.

This is valid. This is what makes WOPR as valuable a resource as it is. I do and will try harder to be more thorough in my research and approach.

Thanks,
Chuck Billow

[HansV]

> why anyone would ever need more than 640 megs of memory

It was 640 KB (kilobytes)!

[HansV]

Microsoft certainly isn't perfect, but it has worked hard on improving memory management in recent versions of Windows (2000, XP, 2003 and Vista). It is now so efficient, that by far the best strategy is to let Windows itself handle it all. Virtually everything you add on to the built-in memory management will either have no discernible effect, or will in fact have a negative impact on performance.

[AlanMiller]

Chuck

I do think this issue is one worth "harping on". I can't help but feel that you're operating under the misimpression that memory management by the O/S is a simple matter of arithmetical bean counting and simple "rules of thumb" (an oversimplification even for Windows 3.1) and that the only issue involved is whether something is stored in physical RAM or the page file:

"... as a result of assumed arithmetic and perceived gains, that when I have some 2 gigs of memory available, and the machine only chooses to avail itself of some 600 megs, that this is silly and wasteful; if there was (and is) a way to point operations in the direction of waiting a bit longer before dumping memory, this should only improve ops, as memory is a good deal more responsive than disk swapping."

When studying for my Masters in IT, I opted to write an essay on Windows 32-bit memory management (ca. Win98/2000) and found the biggest challenge was trying to explain it all in under 4000 words! The best response to this misconception is to quote from the article Joe cited:

"Windows will always try to find some use for all of RAM — even a trivial one. If nothing else it will retain code of programs in RAM after they exit, in case they are needed again. Anything left over will be used to cache further files — just in case they are needed. But these uses will be dropped instantly should some other use come along. Thus there should rarely be any significant amount of RAM ‘free’. That term is a misnomer — it ought to be ‘RAM for which Windows can currently find no possible use’. The adage is: ‘Free RAM is wasted RAM’. Programs that purport to ‘manage’ or ‘free up’ RAM are pandering to a delusion that only such ‘Free’ RAM is available for fresh uses. That is not true, and these programs often result in reduced performance and may result in run-away growth of the page file."

You also ask, "... why third-party developers are constantly coming up with solutions and improvements that could very easily have been addressed by the major players like IBM and HP from the very beginning." In the context of "improved" memory management, especially for XP/Vista, I'd say that the answer is that they don't understand the complexities involved either.

In the context of some of these undocumented/ dubious tweaks, many of the settings people have stumbled upon may be there for reasons of backward compatibility etc. and are not intended to be fiddled with. I know you like to operate on what seems logical, so consider this one - if MS exposes a setting in the registry, don't you think that they understand its significance and don't you think they'd set it up for optimal (or safest) performance by default?

Alan

[CWBillow]

More the point.

CB