Windows 7 memory usage: What's the best way to measure?

Ed Bott's Microsoft Report

Windows memory management is rocket science. And don’t believe anyone who tells you otherwise.
Since Windows 7 was released last October I’ve read lots of articles about the right and wrong way to measure and manage the physical memory on your system. Much of it is well-meaning but just wrong.
It doesn’t help that the topic is filled with jargon and technical terminology that you literally need a CS degree to understand. Even worse, web searches turn up mountains of misinformation, some of it on Microsoft’s own web sites. And then there’s the fact that Windows memory management has evolved, radically, over the past decade. Someone who became an expert on measuring memory usage using Windows 2000 might have been able to muddle through with Windows XP, but he would be completely flummoxed by the changes that began in Windows Vista (and its counterpart, Windows Server 2008) and have continued in Windows 7 (and its counterpart, Windows Server 2008 R2).

Click to expand...

More: Windows 7 memory usage: What's the best way to measure? | ZDNet

I would prob disagree with him that Windows tries to keep the blue part of the RAM graph as small as possible. It is usually quite large even when my green part is over half the graph length. I think Windows could do better in this area of keeping cached memory larger.

I thought this article was pretty good and easily worded for somebody with just a casual interest.

With regards to the size of the blue bar (both dark and light), I find that on my laptop with 4GB of RAM and 3.5 usable (32bit OS), that mine hovers around 1.7GB in Standby and about 260MB as flat out free. However, on my desktop with 8GB of RAM, and all 8 usuable (64bit OS), that my standby sits around 2.8GB free with 1.9GB as flat out free. In my case, my desktop mainly runs virtual machines and only fills up the RAM when I am running 2-3 concurrent VM's on it. Since I do so little else with the box, I can see it not filling up the RAM with superfetch stuff as it's only large VMDK files that I am opening when I launch my VM.

I also liked the fact that it was presented how superfetch works, and the fact that it is done at the rate of a few pages per second with VERY Low Priority I/O's....so it doesn't really impact performance. There is no denying that for performance reasons, people suggest turning off superfetch and the like and try to keep their RAM as free as possible...which I have maintained is not really the best route to go for performance.

What's interesting about SuperFetch is that Intel states to turn it off if you use one of there SSD's. I did as instructed and would love to know what all that has done to help or hinder my OS?

toughbook said:

What's interesting about SuperFetch is that Intel states to turn it off if you use one of there SSD's. I did as instructed and would love to know what all that has done to help or hinder my OS?

Click to expand...

Intel doesn't say why? They just tell you to do it? A bit odd.

Thank you Capt.Jack Sparrow, I learned a lot. I don't know weather the amount of Ram they show in each color is as it would be on my computer. Most likely not. It's the best way I have seen to monitor these things. I don't know if the percent for each group can be changed or if one would want to.

I'm not sure if I read that INTEL says to disable SuperFetch...but there certainly is discussion on that topic on the Intel boards.

The general rule of thumb seems to be that superfetch isn't needed with an SSD because an SSD is so fast anyway. Also, people are concerned with the limited number of writes to an SSD drive...so they often try to eliminate as much as they can.

My take on SuperFetch is that even with an SSD drive, it's not bad to have it on. #1) It preloads (or reads) from your SSD...that's not going to wear it out. #2). Regardless of how fast your SSD is, your RAM is tremendously faster.

So, I say...leave SuperFetch turned on.

I agree with pparks. I've found that leaving SuperFetch on, and disabling other services, like the homegroup services (not needed if you don't use it and they run automatically by default) saves more ram and boost performance more than the other way around.

Good read.

AS far as the SuperFetch question goes:

With my SSD, I have tried both ways.
SF Off, and then again ON.

I admit, after turning Superfetch back on, on the SSD there did not seem to be any difference.

However, after approximately 4days of use, there really was a performance difference.

After running SF a little over a week, I disabled it and could then tell the difference.

The reason I believe this is the case:
Although SSD's are still much faster than a traditional HD, they are still is no where near as fast as RAM.

So in my conclusion & opinion, SSD drives should run Super-Fetch.

I truly think all the hype about the writes SF does degrading the SSD is blown way out of poportion. At least for more recent SSDs.

The only tweaks I think matter for a SSD, are:
1) The initial install and setup (partitions,proper alignments)
2) Disable Windows defrag on the SSD
3) Turn off Drive indexing on the SSD drive

Beyond that, no further tweaks need be done on a Win7 system running on a SSD. Just my opinion.

Mine came with a disc, in that disc is an installation PDf. It cleary stated to turn SF off. I did. I will try it with it bavk on sand see if I noticed a difference after awhile.

Wishmaster said:

I truly think all the hype about the writes SF does degrading the SSD is blown way out of poportion. At least for more recent SSDs.

Click to expand...

Yeah, when people hear something like "limited number of writes", they often start jumping to conclusions and do almost everything it takes to eliminate whatever they can.

I leave Superfetch on and it seem to work well for me.