Possible to "dump" RAM?
 

I have no idea what the actual technical terms are, but is it possible to "dump" or clear the RAM so-to-speak without rebooting? After my Mac runs for some time, I noticed my RAM meter starts to look like this...

http://greg-loesch.com/pics/istat.png

That screen shot was taken with only iTunes, Mail, and Safari running. Why does it show so much activity? Is it a flaw in iStat or does my computer actually have that much activity going on? Any help would be great. Thanks!

Are you experiencing performance problems with your Mac?

Edit: I am learning to dislike these "system resouce meter" apps because they get people all obsessive-compulsive and anxious about what's going on under the hood, usually for no good reason. /end rant

What Mikey-San is saying is that your RAM meter is not showing anything that is a problem, and unless you are experiencing a problem (for example a slowdown of some kind), then there's no need to create a problem where none exists.

There's the old story (I'm not sure whether it's true or not) where one year, Jaguar changed one model's temperature gauges to show the water temperature in degrees instead of just having lines with no numeric markings. There were no other changes to the cooling system that year. Cars from that year kept showing up at the dealers with claims of overheating. The dealers would run tests and the cars were having no problems at all with overheating, but people were very sure that the cars were too hot. Complaints skyrocketed.

The next model year, Jaguar removed the temperature numbers on the gauge. That next year, complaints of overheating went down to normal again.

Trevor



Edit: You are showing pageouts, which could be fixed by adding more memory.

I can't specifically disagree with this but it's worth noting that there is another point of view.

If running a Mac were essentially problem free at all times, there would be no reason for regular users (i.e., non-geeks (not intended in the disparaging sense)) to worry about what's going on under the hood. Since it's often not problem-free, it is understandable that there is a desire to have some gauges and indicators (to borrow from Trevor's example) to monitor what's going on, so that one might get some indications about "non-standard" performance rather than all of a sudden be faced with a totally non-responsive machine.

What seems to be lacking (or at least I haven't found anything thus far, and I have searched!) is a fairly simple, layman-type guide of what's normal and what isn't. Things like "normal" processor usage and durations for Finder, for example, normalized to processor type and speed. When Finder starts to use 100% CPU along with 50% for WindowServer and 49% for kernel_task for minutes at a time, I have no idea if I should just wait until things settle down, start force-quitting applications, or run for a fire extinguisher.

Having various system resource meter apps does provide some sense of security (I know, it's probably false, but . . .) that things are working OK, within some definition of "normal" based on past experience with the apps.

I don't want to let this highjack the thread, but if Finder is taking 100% of one of your CPUs, then remove all of your system hacks, drivers, and kernel extensions that you installed (not ones included with OS X).

Trevor

I don't see much problem with that screen shot. Even the pageouts are a low proportion to the pageins. There is no immediate need for more RAM indicated by that screen shot.

Safari is known to do a significant amount of caching in RAM, possibly for performance reasons. Their programmers probably figure if you have all that RAM sitting around with no other programs asking for it, it should be used for performance caching instead of letting it go to waste, unused. I'll bet if you quit Safari the bar would get shorter.

The resource meters included with OS X are no better than those included with Windows or other OSs; in other words, they are built for geeks who know what the numbers mean. It would be nice if Apple designed a resource meter that interpreted all of the truly relevant factors for newbies without them having to understand them. For example, if the entire memory bar was 3/4 full but there are no pageouts, the "status readout for newbies" light should be green. If there are 50MB of "free" RAM, no "inactive" RAM, and 7 swap files, the same light should be red.

Yup, and when it comes to system resources I am kind of OCD myself and love to run a barebones type config, where I have the essentials running and nothing else.

However, OS X uses Unix memory management and it threads things. Even though your memory is being threaded and cached constantly it doesn't reflect what it does to performance. That is because Unix on the fly can reallocate memory to what task is being currently used. Therefore, you need not worry unless your machine is running slow.

I've been here before:). I've tried to figure out possible causes and solutions by reading your (and others') suggestions in other threads.

The only "hack" I've got is the Logitech Mouse Controller. To the best of my ability to discern, except for LittleSnitch.kext, I've installed no drivers or kernel extensions of any kind that didn't come via Apple S/W. The only stuff in my System Preferences panel that are non-Apple are DivX, Flip4Mac, MenuMeters, Perian, and Startup Sound.

I get these occasional spurts of high activity with Finder and other stuff whether these are installed or not. Fortunately, they're not too frequent, and are often (though not exclusively) associated with finding files via Spotlight on a large drive. In these instances, Finder will crash after a minute or two of thrashing. When I restart Finder, most of the time it will go back up to the nearly 100% usage, but will return to normal after the searching is done. And then things go on normally until the next time, which can be minutes or weeks later. (This hasn't happened yet using Easyfind, if that explains anything.)

This type of behavior has persisted through several "erase and install" procedures.

It's worth pointing out that Logitech's installer has been known to install Application Enhancer.

Logitech: Great Devices, Awful Software(tm)

I really didn't mean to start a derail. Heh.

The complete Mac OS X memory management model is complex. To describe it, you have to discuss low-level memory space partitioning, virtual memory handling, framework-level and language-level memory management policies, and how individual applications manage their memory. That's why I get annoyed when I see Yet Another Memory Meter App hit the streets. Unless you know exactly what you're looking at, a usage meter isn't going to tell you what you think you want to know.

Like a few of us have said, the bottom line is that unless you're experiencing performance problems, there's not much to care about. If you find that you are, a memory usage meter is only one part of the investigation.

Wow... thanks for all the quick responses! No, I haven't had really any performance issues, so I guess I should take your advice and not worry about it. I'm one of those that likes to get the most out of everything I have. If something's not running like it should be, regardless of whether or not it's really bogging things down, I want to make it better. Perhaps it's a perfectionistic tendency that I need to work on... :p

I do have some performance issues with Preview and Pixelmator (though I'm not sure whether or not System RAM is the issue there... since it relies heavily on GPU) when running certain photo-editing processes. Some actions are really laggy for lack of a better term. It doesn't seem like this should be a normality since I'm using a MacBook Pro that I bought this past summer...

I'm guessing that since nobody as brought anything of the sort up, there's no app/unix command that "resets" the RAM so-to-speak? Regardless, thanks again for the input.

Everyone has performance issues with Pixelmator. It's just a slow application, in many situations. Memory is not always the bottleneck.

It doesn't work the way you think. See what I posted above.

I'm not aware of any app that "resets" the RAM, but you can close applications to free up RAM. That's the easiest method. Besides, if you "reset" the RAM, all of your open applications would crash at best, or potentially do some really stupid things to your system depending on what gets written to the addresses that those applications are pointing to.

Also, I don't believe it was pointed out earlier in this thread, but CoverFlow can consume a lot of RAM. If you want to do a fun little experiment utilizing your RAM meter, do the following. Open only iTunes. Switch from list view to CoverFlow and watch your RAM usage go up significantly.

Given that the OP's question has been answered, perhaps the hijack is not that big a problem:), although the moderators may well decide the hijack-related posts deserve their own thread.

At a conceptual level, I can understand what you're saying and why.

Yet, I don't have to be an expert mechanic to know that, even if I am experiencing no specific problems, if my car's idle speed is below 675RPM or if the voltage is below 12.25V, etc, etc., I ought to be concerned enough to have someone who knows something take a look. Just the idea of not having insight until failure/damage/substandard performance actually occurs goes against my instincts. Obviously, a car isn't a computer, so the example may be completely inappropriate.

Is it your opinion that there is nothing along the lines of MenuMeters (displays memory usage, disk I/O, network traffic, and processor usage) that may be of value to layman users?

If there is something, can you recommend an alternative(s)?

I consider my question answered. Hijack away my friends :cool:

Since the hijack has been duly authorized.... :p

I can't speak for Mikey-San, but I will agree with him that having a meters for every aspect of their computer tends to make users obsess over the little things (or perhaps they would obsess anyways).

To respond with an analogy in kind: It's like putting medical guides in the hands of lay-people. Doctors are now inundated with people who have already "diagnosed" themselves before ever giving the physician a chance. My sister-in-law is terrible for this. I can't count the number of times she has misdiagnosed herself because she found something in a medical guide! She's a librarian, so she seems to think she is qualified to look anything up in a book or online!

While I think she has a right to keep herself informed, she should probably wait until there is a problem and then a diagnosis before looking it up in a medical guide.

Yes, it's inappropriate. Car analogies usually are. They're confusing and usually not directly applicable. They just cause misconceptions and force re-explanations. Instead of asking a question that someone can answer, they force knowledgeable people to dissect them, which just increases the complexity of the conversation.

I think that you first must define the scope of "value". What knowledge is desired? Only then can someone recommend a useful, appropriate way of obtaining that knowledge.

This seems to be a chicken and egg situation:o: being a layman, I have insufficient knowledge to know what it is I don't know about ongoing processes that would have value to me (didn't mean to sound Rumsfeldian).

It seems the best I can do is to provide some examples:

1. Unless I'm doing something like indexing or copying or similar, continual activity by MenuMeter's disk access indicator lets me know there is something wrong and I should check further for a misbehaving app or process or possibly a disk problem.

2. MenuMeter's network speed indicator, coupled with the spinning thingie indicating that .Mac sync is active, lets me know that I don't need to worry about long download times because there is something else that is legitimately taking up part of the bandwidth.

3. Only one arc in the airport signal strength indicator provides assurance that slow download speeds are probably not a computer or network malfunction, but excessive distance from my base station.

4. The charge indicator on the MBP lets me know that, even though current performance is completely unimpaired, I need to do something in the next n minutes to avoid a complete loss of function.

All of these require no more than layman's knowledge to be valuable, yet they give me a "feel" for what's going on under the hood, and provide a (most likely false:)) sense of security that I'm somehow in control.

My habit has been to try many tools recommended here and that I find at sites such as VersionTracker. I dump almost all of them within a couple of days, as they don't seem to provide added value to me. Trial and error has its place, but it is generally not the most effective manner of finding useful things.

I understand your point that tools that provide information that is not particularly useful to people with only a layman's understanding can be counterproductive. Do you have recommendations for tools that may be useful to a layman in a manner similar to the previous examples, rather than being counterproductive?

Honestly, it's a really difficult question to answer. It's not the tools that the layman needs; it's specific knowledge goals like those you mentioned and a basic but correct understanding of the information that the tools provide.

Many people, like the OP, run monitoring apps and freak out over the information it displays, when they've not actually experienced a problem. They had no information goal at all, and coupled with a lack of understanding of the information and their own situation, it just ends up with a bunch of anxious, freaked-out users posting to Discussions and the like. (I'm not knocking on the OP, but it's a good example that is right here. Sorry!)

That MenuMeters screenshot is essentially a pretty version of some of the info you get from the top utility. I can't hate on top, 'cause it's a pretty essential tool, so I can't dismiss MenuMeters as being completely useless. But it doesn't help anyone who doesn't understand what he or she is looking at. Neither does top. (Case in point: threads that involve someone complaining about the VSIZE column or the "free" PhysMem statistic.)

This is, of course, the crux of your question: What tool is there for someone who doesn't understand what other tools display?

I don't think there is one. There's always going to be the need to understand what you're looking at until someone figures out a way to display a "check engine light" in the menu bar. Oops, car analogy. ;)

Perhaps the layman's tool is a forum like this one where knowledgeable users post, combined with monitoring tools and a well-formed question or problem. :)

Imagine a person with extreme OCD organizing a closet of clothes and shoes and redoing it according to season and then by color or whatever over and over again, that is how OS X memory management kind of works in a sense. Its hard to explain out right with simple analogies. However, what it does is it just threads everything into active and passive processes, and passive it is like a ghost, there but not doing anything or using up any real amount of resources and OS X can switch on the fly how it allocates memory to what you are doing at that current time.

Sorry if that was confusing.

The app ifreemem does something like this.

No, it's not. I'm not trying to be mean, but this doesn't even make sense.

yeah i was trying to be really simple, and put it into a real world application that someone would do constantly...

i failed

And you didn't really read the thread. :(

And what part of this did I miss read?

The thread is more than a single question asked by the OP.

I did not see you or anyone else answer this part of compulsive's question. Is it not the goal to answer the questions? Yes I see there is more to the thread its pretty obvious.

I think modern memory management could be likened to a professional sports game. In a sport the goal is to score, so a sport newbie would see all the players taking various positions that don't seem to have anything to do with the goal and say "Why don't they just go to the goal? What is that guy doing way out there? How come that one basketball player stands still while his teammate heads straight for him?"

But the sports veteran would recognize all of the positioning and setup that has everything to do with anticipating upcoming situations before, when, and after the ball moves toward the goal, ready to take advantage of all the opportunities and also to head off all of the potential traps. The coach is like your OS. It really does know better than you do. If you try to do the obvious, the other team will cream you. If you instead execute plays that have been battle-tested to be fast, efficient, powerful, reliable, and secure, your team will do well. OS X has years of professional RAM "coaching" experience behind it thanks to Unix. I do not believe in apps like iFreeMem.

After reading that rather brief product description on the Activata iFreeMem product page, I am still at a loss to describe what the application does that the OS doesn't already do automatically? Isn't part of the system memory managers job to reduce "memory fragmentation" (as Activata puts it)? I don't see how this software can do what would otherwise require a "reboot or RAM upgrade". I'm just going to go out on a limb and guess that this is crapware!

I have nothing to do with this program and do not even use it but how can you just call it crapware? Do you know for a fact that it does nothing? It seems to me that if it releases all the memory in cache (like it says it does) then osx starts to re-assigning memory to the cache that this may be beneficial in some instances. For instance, you may have had 15 apps open all day and then want to start a recording session. It would make sense to release all the memory cache and then let osx rebuild them with the apps you are working with now.

I also think that osx would eventually do this on its own but the OP asked for an "instant" solution. Hence the suggestion and comment.

Now, does it work? I have no idea but then again I am not a osx programer. I would leave it up to other people that may need this service to decide if it works. I have to trust that the developer knows what he is doing and offers legitimate software.

I notice now that there is a note on the developers website that 10.5 users really will not benefit from the software. Again, would someone post that if they were just trying to get your 10 bucks or whatever it costs?

I am highly doubtful that this problem is useful in the sense of increasing the overall performance of your Mac. One string reason to be dubious is that otherwise you are believing that some application developer has a magic trick that fixes something so that it works better than what Apple's kernel engineers can do.

I note that my "VMTester" utility (free and open-source!) has an option to run a command to consolidate RAM. I am highly doubtful that such a command increases performance. Yes, it might result in your machine being faster later, but you need to factor in the time that you spend running that command. The system would have done the same thing (consolidation) in its own time when RAM is needed.

hayne, I agree with you completely on several of your points.

However, Pitting kernel engineers against an independent software developer is a bit over the top.

I think a clever developer could probably change or manipulate the memory management temporarily to give the results that this developer claims.

I really don't think its a magic trick. It is probably known by many people that are intimately familiar with the osx memory model.

Its the same thing as being inside a chemical lab that produces cool high end plastics or such. There are those that know exactly how the plastics are made. There are other chemical engineers that can figure it out and reproduce it. There are still other people that can figure out what you did and exploit it to do something a little different.

Sorry, but it sounds like crap to me, too. I, like Hayne, have a hard time believing that this magical shareware app is better at freeing memory than Mac OS X's kernel engineers. It just doesn't make any sense.

Even from a cost-benefit point of view, it's crap. This thing is 10 pounds. As of this post, that's about $20. That's half the price of a physical RAM upgrade, practically. If you're having memory problems, why not just buy more RAM at that point?

Calling "ripoff".

I don't think you understand how Mac OS X memory management works.

I'd say it's exactly the core of the matter.

Nope, I don't and I think I made that perfectly clear with this statement:


Agian as I said before:
I also think that osx would eventually do this on its own but the OP asked for an "instant" solution.

Is this justification for calling a program crapware?

I didn't call it crapware by the way. I said "I'm just going to go out on a limb and guess that this is crapware!" That's my guess from what I know of memory management. Mikey-San and hayne know more than me, to be sure, so I will leave it to them to correct me if I misrepresent something below.

The criticism I have of this app and its alleged usefulness is this:

1. iFreeMem developers claim to "recover memory for your applications to use". The built-in system memory manager in OS X also consolidates and organizes memory in order to keep the system running efficiently and provide running apps with the necessary memory. The developer does nothing on their site to differentiate their product or explain how they do this better than OS X - observe below:

"IFreeMem clears Inactive memory to help your applications avoid the considerable performance hit you get when running low on Free memory." (From the iFreeMem page on Apple Downloads)

vs

"Inactive memory
This information is no longer being used and has been cached to disk, but it will remain in RAM until another application needs the space. Leaving this information in RAM is to your advantage if you (or a client of your computer) come back to it later." (From Apple's document Mac OS X: Reading system memory usage in Activity Monitor)

2. The developer makes a grossly misleading statement on their website. "iFreeMem optimize feature is a quick and easy alternative to either a reboot or RAM upgrade to get defragmented free memory." It is NOT necessary to reboot in order to get "defragmented memory". OS X recovers memory from inactive applications and applications that have quit and shuffles memory around in RAM in order to maximize the available blocks of memory. If it didn't, the system would eventually crash because not enough blocks of memory (that are large enough to be usable) would be available for newly starting processes.

Yes. If someone who DOES understand OS X's memory management looks at those claims and says that it "doesn't make any sense", that is justification for that someone to call it crapware. I don't think Mikey-San is saying "I don't understand it... it just doesn't make any sense". I think he is saying "that's bull... it doesn't make any sense".

The OP asked for a solution he did not need; he asked the wrong question, so we gave him a different answer. This is what you would've learned had you really read the thread. Read this:

http://catb.org/~esr/faqs/smart-questions.html

Yes, as a matter of fact. They're trying to get people to pay them for an application that doesn't really make a whole lot of sense. So, crapware.

I was going to post something along the lines of what wdympcf just posted, but he beat me to it. I agree with his post; they're making wild statements about their product and not describing what it actually does, while misleading people into thinking that certain system stats suggest they need to run this magical program to "fix" their memory.

I am sorry wdymycf but you have just solidified all my points. Thank You.

Please enlighten me, so I can truly take credit for your thanks. What points have you made that I just re-enforced?

You're still missing my point though, where the OP asked the wrong question in the first place. Do you really understand the higher goal of the thread, or are you just trying to defend this app at this point?

I also agree with wdympcf and Mikey-San. A program like this was sorely needed in OS9 and prior, but Mac OS X does not have memory fragmentation problems.

I think you are missing the point of my initial comment. The OP had an unanswered question as I have said before (see the thread) and I tried to answer it.

Because people have suggested that the application is not possible is where we are now.
This is neither my intention or my fault.

Higher goal? Knowledge is the higher goal. Is that not the goal?

It's not that the application is not possible - it's not necessary!

And several of us are evaluating that answer and saying that it's probably not the best choice. I'm not faulting you personally for your answer - you were trying to help the OP. There is a difference between trying to help and actually helping. I'm suggesting that this application will not actually help.

I just realized there was another point that you made earlier that I forgot to address:

You don't HAVE to trust anything. That's an important point. You are giving the developer the benefit of the doubt without having used any critical thinking to assess whether this is warranted. We (Mikey-San, hayne, Las_Vegas and myself), however, are looking at this critically using our experience and knowledge (varied as it may be), and find it to be lacking substance.

hmm, sorry I over reacted. I like that last line though.

Don't take it personal its just the internet man. I am sorry you are feeling you are being attacked, but I think its all been taken out of context by everyone just a bit.

Wow! Umm... wow! I hope you come back and read this, because I think you've taken this way, way, way too personally. You were expressing your opinion and I was also expressing my opinion. I don't have any animosity towards you. I didn't intend to attack you and I don't think I said anything that should be taken as an attack. I also don't think that the others were attacking you, but I will let them defend themselves of that charge if they see fit.

Now that's an attack....