Find out max RAM usage

[jesterb84]

I tried Google and found software which shows RAM usage for each process, etc. etc. which the Windows Task Manager can very well do. I was wondering if anyone knows of a utility which I can leave running and shows the peak physical memory usage (not current memory usage but peak memory usage).

The idea is to have the utility in the background running for an entire day while I go about doing my tasks and at the end of the day, I look at the utility and see what was the max RAM usage.

 

[yinan]

Windows Performance monitor can do this I believe.

 

[WildW]

Isn't that the meaning of the "Commit Charge" in the performance tab of task manager? As in Total = Current, Limit = Theoretical Max, Peak = Biggest so far ?

I guess that's "memory used" , not "physical memory used" . . . since for some reason beyond the wisdom of mortal man Windows will use virtual memory when there's free ram. I'm going to stop now before I say too many wrong things in one post.

 

[jesterb84]

Thanks but not sure how to use it. I went into the Windows Performance monitor and added a counter for "Memory" (Available MBytes). I then shows me a scale from 0 - 100 not sure how to read the data. Isn't there an easier way to do this (i.e. HWMonitor can show me the peak temperature of the CPU in degrees)?

 

[jesterb84]

Originally posted by: WildW
Isn't that the meaning of the "Commit Charge" in the performance tab of task manager? As in Total = Current, Limit = Theoretical Max, Peak = Biggest so far ?

I guess that's "memory used" , not "physical memory used" . . . since for some reason beyond the wisdom of mortal man Windows will use virtual memory when there's free ram. I'm going to stop now before I say too many wrong things in one post.

Thanks, actually if you're right then that resolves what I am looking for. Basically, I have 4GB of RAM and decided to disable the Windows Paging file. So far, tasks have been noticeably a little snappier (most apparent during startup when programs are loaded and with game load times as well). That's why I wanted to see how close to the 4GB barrier I am reaching on everyday usage and whether I should upgrade to 8GB RAM now or wait a few months for prices to drop more.

 

[theAnimal]

Let Windows manage the page file. Even if you disable it, a page file is still created.

 

[hanspeter]

Originally posted by: theAnimal
Let Windows manage the page file. Even if you disable it, a page file is still created.

Why would it still be created if you disable it?

Anyways, if I disable mine, I don't see any page file being created anywhere. And I don't remember the manual saying otherwise.

 

[BlueAcolyte]

Even if you have 4GB of RAM, you need a pagefile.

 

[Nothinman]

Why would it still be created if you disable it?

I think there's more variables involved in it's decision to create one behind your back because it doesn't always seem to happen, but either way it's a terrible idea to run without one.

 

[hanspeter]

If Windows creates one anyway, why will applications and Windows itself begin to crash when you are out of memory?

There is a temppf.sys (I've never seen it in xp, and the KBs only mention it in conjunction with windows 2000 and earlier) that may be created when you boot Windows, if there is not enough ram to run stable.

 

[Nothinman]

If Windows creates one anyway, why will applications and Windows itself begin to crash when you are out of memory?

There are many different ways to run out of memory.

There is a temppf.sys (I've never seen it in xp, and the KBs only mention it in conjunction with windows 2000 and earlier) that may be created when you boot Windows, if there is not enough ram to run stable.

It's been a while since I booted a Windows machine without a pagefile so it probably was Win2K. And I vaguely remember NT4 just giving a BSOD on bootup when no pagefile was present. However removing the pagefile should have little to no affect on performance if you've got enough memory. It's just something that's pointless to mess with and try to optimize.

 

[GrumpyMan]

All right! Our monthly page file discussion. Still waiting for benches to prove that it benefits a system to have no page file.

 

[iCyborg]

Originally posted by: jesterb84
Thanks but not sure how to use it. I went into the Windows Performance monitor and added a counter for "Memory" (Available MBytes). I then shows me a scale from 0 - 100 not sure how to read the data. Isn't there an easier way to do this (i.e. HWMonitor can show me the peak temperature of the CPU in degrees)?

When you add counters, there's another tab "Log Files". There, I usually choose the logs to be written in a CSV file and read it with MS Excel.
You can also change the scale to be something other than 0 to 100 when you click on View Log Data.

 

[Yellowbeard]

Try Process Explorer. It's a neat little tool.

 

[jesterb84]

Thanks iCyborg and Yellowbeard. I ended up using Process Explorer and it's an AWESOME tool for sure.

As for the page-file debate...well I surely have no empirical evidence or benchmarks surrounding this and it could all be very well made up in my head but I'm quite sure I do feel some performance boost. At least, I've noticed less hard drive activity after a large program (such as a game) exits and I'm almost sure too there is a slight boost in performance when starting some apps. I don't believe I am hitting the 4GB wall as all I do is run ONE thing at a time on Vista 64 (if I game I close everything else, etc.). The only things left in the background would be Skype, Windows Live Messenger, AVG Free, Gmail checker, and a couple other things such as Daemon Tools.

Anyhow, time to do some monitoring!

 

[imported_wired247]

put pagefile on ramdisk?

 

[Yellowbeard]

Here was my reply to a Page File vs No Page File debate earlier today:

Purely playing Devil's Advocate here. Let's say that with sufficient system memory the user believes that he or she no longer needs the page file. (Rhetorical question) Why is that?

The user believes that the OS will not need to page to a hard disk because there is sufficient system memory that paging will not be necessary. If this is true, then why worry with turning the page file off? Turning the page file off only makes sense when the user has an abundance of physical memory. If that is the case, then turning the page file off seems to be a pointless move.

FWIW, I have recently done some testing with 12GB of memory vs 6GB vs 3GB. The OP here has 8GB so I think the same info applies. With 6 and 12, the system stopped most but not all page file activity. However, the miniscule amount of page file activity that persisted did not effect system performance at all. And, IIRC, the only data being paged was for the OS and not the game I was testing.

My point is, that the OS chose to place some data into virtual memory even when there was an abundance of physical memory available. So, until someone can explain to me in terms I understand why turning the page file off is a good idea, I'm leaving mine on.

 

[Phynaz]

Originally posted by: Yellowbeard
Here was my reply to a Page File vs No Page File debate earlier today:

Purely playing Devil's Advocate here. Let's say that with sufficient system memory the user believes that he or she no longer needs the page file. (Rhetorical question) Why is that?

The user believes that the OS will not need to page to a hard disk because there is sufficient system memory that paging will not be necessary. If this is true, then why worry with turning the page file off? Turning the page file off only makes sense when the user has an abundance of physical memory. If that is the case, then turning the page file off seems to be a pointless move.

FWIW, I have recently done some testing with 12GB of memory vs 6GB vs 3GB. The OP here has 8GB so I think the same info applies. With 6 and 12, the system stopped most but not all page file activity. However, the miniscule amount of page file activity that persisted did not effect system performance at all. And, IIRC, the only data being paged was for the OS and not the game I was testing.

My point is, that the OS chose to place some data into virtual memory even when there was an abundance of physical memory available. So, until someone can explain to me in terms I understand why turning the page file off is a good idea, I'm leaving mine on.

+1

It's really silly to think you know more than the guys that create operating systems for a living.

 

[jesterb84]

Originally posted by: Phynaz
+1

It's really silly to think you know more than the guys that create operating systems for a living.

That's a pretty odd statement. If we all thought it silly to think we knew more the guys at Intel/AMD who made processors for a living, then there wouldn't even be an overclocking subsection here and I wouldn't be enjoying a +1.2GHz/core speed boost right now. If everyone had this mentality, there would be no new innovations.

Secondly, nobody is saying that they "know more" than MS about Windows. Windows was configured for the largest compatibility with the largest user base. Just as Intel/AMD clocks their CPUs conservatively and car manufacturers detune their engines to a conservative level - I don't see the harm in trying to tweak Windows to run at a higher performance level.

 

[Nothinman]

That's a pretty odd statement. If we all thought it silly to think we knew more the guys at Intel/AMD who made processors for a living, then there wouldn't even be an overclocking subsection here and I wouldn't be enjoying a +1.2GHz/core speed boost right now. If everyone had this mentality, there would be no new innovations.

So you honestly thing you know more about a CPU than the people that designed and built it? That's a pretty bold statement.

Secondly, nobody is saying that they "know more" than MS about Windows. Windows was configured for the largest compatibility with the largest user base. Just as Intel/AMD clocks their CPUs conservatively and car manufacturers detune their engines to a conservative level - I don't see the harm in trying to tweak Windows to run at a higher performance level.

Because it opens you to a lot more potential problems and it doesn't get you a higher performance level, it's just a placebo. Sure, there are valid senarios where a pagefile isn't necessary but none of them include anyone's home gaming machine.

 

[jesterb84]

Originally posted by: Nothinman

That's a pretty odd statement. If we all thought it silly to think we knew more the guys at Intel/AMD who made processors for a living, then there wouldn't even be an overclocking subsection here and I wouldn't be enjoying a +1.2GHz/core speed boost right now. If everyone had this mentality, there would be no new innovations.

So you honestly thing you know more about a CPU than the people that designed and built it? That's a pretty bold statement.

That would be a very bold statement but only thing is, that's not what I said. Please, read my second paragraph again. I don't have to know more than Honda about my Honda Civic to tune it for better performance. Nor do I have to know more than Intel about my C2D to o/c it for better performance.

As I mentioned earlier in one of my posts, I am not debating on whether disabling paging has any advantages because I have no benchmarks or empirical evidence to show. The objective of this thread is also not for me to convince people that this is worth it either. I am more surprised at some of the "if you are not the person who created this then don't try new things out" type of mentality. I guess maybe it's just me but I'm more adventurous and like to find and confirm things out myself instead of blindly accepting what others tell me.

 

[Nothinman]

That would be a very bold statement but only thing is, that's not what I said. Please, read my second paragraph again. I don't have to know more than Honda about my Honda Civic to tune it for better performance. Nor do I have to know more than Intel about my C2D to o/c it for better performance.

But you also have a good idea about how each of those tweaks affects the rest of your car's systems, I would bet money that 99% of the people that are attempting to tweak their pagefile setup or run without one have no idea how virtual memory works.

 

[Yellowbeard]

I'll also add that at various times I have tested 4GB vs 8GB with page file on vs page file off. I could NOT find ANY difference within the testing structure we used at the time. I also put the page file on a Gigabyte iRam. Again, virtually no difference and any gains were likely within a 1-2% margin of testing variance. I've done the same with 3GB vs 6GB and again, no measurable differences that would lead me to believe that it was a good tweak. I did not repeat the page file off tests with 12GB as it seemed pointless and I had to move on to other things anyway.

All of this was with Vista 64 on 4GB vs 8GB and Vista 64 w/SP-1 for the 3GB vs 6GB.