Windows optimize

Understand how SuperFetch uses RAM to enhance system performance

Microsoft Windows Vista uses an applet called SuperFetch to manage memory. Greg Shultz explains how SuperFetch works and what it does to enhance the performance of your system.

I recently received a phone call from a business associate who had just got a new laptop with Windows Vista Home Premium with 2GB of RAM and an ATI Radeon video card with 512MB. He was happy with the system's overall performance, but he was concerned about the amount of available RAM on his system.

He told me that at any given time while using the system under normal conditions, with very few applications running, he would access the Performance tab in Windows Task Manager and discover that the amount of free memory in the Physical Memory panel would read 10MB or less. He then went on to say that his old Windows XP system with 1GB of RAM typically would, under similar circumstances, have at least 500MB to 700MB of free memory showing in the Available memory section of the Physical Memory panel.

He was surprised that there would be such a huge difference between the two operating systems' use of memory and came to the conclusion that Vista was just a memory hog. He then asked me what would cause Vista to use all of the system's available memory?

My answer to him was that what he was witnessing was a Vista performance enhancement called SuperFetch and the fact that he had very little free memory available was a good thing -- not a bad thing.

Since other Vista users may have the same reaction as my friend, I thought that a detailed explanation of how SuperFetch works would make a good topic for the Windows Vista Report.

This blog post is also available in the PDF format as a TechRepublic Download.

Comparing Vista to XP

Let's start by comparing the Performance tabs on similarly equipped XP and Vista systems just to see the difference. On the left of Figure A is Windows XP and on the right is Vista.

Figure A

Comparing performance

Both systems have 1GB of RAM, but this can be obscured by the fact that XP measures in kilobytes while Vista measures in megabytes. However, if you convert both figures to gigabytes, it looks more equal.

XP:     982,380KB = 959.355MB = 0.936GB
Vista: 	1,022MB = 0.998GB

Both systems are running only Internet Explorer and Paint. But, if you look at the available memory in XP, it comes out to be about 589MB. Yet, the Vista system has only5 MB of available memory.

The reason for this difference is that Windows XP treats RAM as a resource to be used only when needed while Vista treats RAM as a cache and uses almost all of it as soon as it becomes available.

The SuperFetch advantage

In order to effectively use the available RAM as a cache, Vista employs the SuperFetch technology to efficiently manage the use of that RAM. The SuperFetch memory management scheme is multipronged.

To begin with, SuperFetch tracks the applications that you use most often. As it does so, it builds a database and constantly updates it as you work. Using this database as a reference, SuperFetch instructs Disk Defragmenter to make sure that those applications are strategically positioned on the hard disk where they can be loaded into memory very quickly with the least amount of physical movement of the hard disk's read-write heads.

Then, each time you turn on your system, SuperFetch preloads those applications that you use most often into memory. As such, when you actually go to run your favorite application, it loads up much more quickly from the RAM cache than if it had to be read from the hard disk and loaded into RAM.

While the process as I've described so far sounds like a complete system, it's only part of the picture -- SuperFetch has a much bigger role to play in order to keep this system as effective as possible over time.

Because RAM is a volatile environment and many applications are all vying for space at the same time, SuperFetch employs an I/O prioritization technology, in which applications are marked as either a low- or high-priority I/O application.

With this system, SuperFetch will temporarily sideline a low-priority I/O application when a high-priority I/O application takes precedence. As it does so, SuperFetch keeps track of the previous priority levels in order to ensure the system remains responsive over time.

For example, let's suppose that you're writing a document in Word and have been working on it steadily for a while. All the time that you're using it, Word is recognized as the high-priority I/O application and takes precedence. When you go to lunch, Word becomes inactive and other background tasks, such as Disk Defragmenter, realize that the system is idle and begin to go to work.

When that happens, Disk Defragmenter becomes the high-priority I/O application and Word becomes a low-priority I/O application, so SuperFetch moves Word out of RAM to the paging file to make room in RAM for Disk Defragmenter to do its job. Once Disk Defragmenter finishes, SuperFetch will reload Word back into RAM so that when you get back from lunch, Word is as responsive as it was when you left -- no waiting for Word to be reloaded into RAM.

What do you think?

Now that you know that Vista treats RAM as a cache rather than as a resource, you can understand why the amount of free memory in the Physical Memory panel on the Performance tab in Windows Task Manager is very small figure. If you have comments or information to share about Windows Vista's SuperFetch, please take a moment to drop by the Discussion area and let us hear.

About

Greg Shultz is a freelance Technical Writer. Previously, he has worked as Documentation Specialist in the software industry, a Technical Support Specialist in educational industry, and a Technical Journalist in the computer publishing industry.

53 comments
jackhellmuth
jackhellmuth

I suppose SuperFetch checks to make sure that the application it has cached has not changed on the hard drive before it pulls it from cache, right? That seems like a reasonable supposition but I have seen SuperFetch bring up outdated application versions. What is going on?

radmanx10
radmanx10

While this sounds all neat and cool, I think that its performance is pathetic. I have an HP DV9000 with 2GB of RAM. While I completely understand how superfetch works(I designed processing code for Intel chips), there is no excuse for my system, at 1.25GB on the nice little meter, to not be able to load any more applications - not even notepad.exe. I bought a high performance system so that I would be able to run development environments (Visual Studio), databases (Oracle 10G, SQL 2005, MySQL) and other development tools without a significant decrease in performance. And no, all of these applications aren't loaded at the same time. I always disable the services for the DBs until I need to use one. Vista looks cool, but it isn't a fast performer. Forget the fancy memory management, get back to programming fast, efficient, tight programs and operating systems.

martink
martink

Thanks for the info. In theory it works fine with your every day applications. If you use VMWare once a day then Vista is not for you. You are squeezed for memory since most of the hardware comes with the memory maybe slightly up from XP but on the low side for Vista. With Superfetch you have less memory available for less frequently used applications which in case of VMWare causes the host to swap when the virtual machine is loaded. That takes ages. So instead of using it daily, you tend to avoid it because of the delay and start using it as rarely as you can. Not a good way for the OS to decide what the user wants to do. Martti K.

HBE
HBE

... I do wonder then why I do receive pop-up warnings from Vista about 'your system is running low on memory: close the following application: ' (mostly when working in Adobe Photoshop CS2) ?? HBE

invrabbit
invrabbit

Is it fair to say that since 3.1 that every new version needs more hardware and will be "slower"than the last?

angelo_elibz24
angelo_elibz24

Thanks for the info. Now I know why my computer still runs fast and smoothly even though I'm currently running 3 memory hogs at once and it already reads 0 in the Physical Memory panel lol.

byn99
byn99

Yes I wondered about that,thanks for setting me straight.bgh1109

bkrateku
bkrateku

I don't know if anyone else put this in their posts as I haven't read much, but I'm thinking with RAM full of stuff from SuperFetch, wouldn't this potentially make new programs pull up slower than XP since now Vista would have to move some stuff out of RAM to make room for the new program to run? I realize it isn't much of an issue, but a thought that crossed my mind. I would think a separate chunk of memory, like allocating a certain amount for SuperFetch instead of the rest of the available room, would work better. I know you can USB keys for that I think, but allocating actual memory would be quicker (although not as cheap). Thoughts?

Jaqui
Jaqui

the old TSR issue done deliberately. sorry, I don't want applications sucking up RESOURCES until I specifically start them. and I want them to go completely away when I shut them down.

OldER Mycroft
OldER Mycroft

If SuperFetch is supposedly sooo clever, why do I continually see these little underpowered laptops where the hard drive is thrashing constantly??

Mark W. Kaelin
Mark W. Kaelin

SuperFetch is an interesting technology -- one not available in Windows XP. But the question that really needs to be answered is if SuperFetch really increases performance? What is your experience with Vista performance? I am satisfied with my Vista systems, are you satisfied with yours? Explain why or why not?

Oz_Media
Oz_Media

I have read this before somewhere, a little better explained too, but this was a quick hit on my way out the door. In essence, it is supposed to use your RAM, that's what Superfetch does, manage RAM use, a lot like MEM Turbo and other old apps but without simply clearing RAM, Superfetch actually replaces old paged files in the RAM. I just mentioned the page file issue to Old Mycroft too, he was also aware of how few people actually adjust the page file to utilize RAM and reduce HD reads for faster access. anyway, here's another explanation for you: "If you have been using Windows Vista and were curious to check the CPU and RAM usage, then you have surely made your way to the Performance tab in the Windows Task Manager. Here, you may have come across readings similar to those in the screen shot on the left. Forget about all the data displayed and just focus on Physical Memory (MB). It reads Total: 2,045, Cached: 1,277 and Free: 6. 6 MB of free RAM? Why did my memory go? You can see that the CPU is at 3% so I'm really almost doing nothing at all. Then why does Windows Vista consume all the system memory? Windows XP did not eat up this much RAM. Checking the same Performance tab in Windows XP will always reveal a consistent amount of free physical memory. Is this another case of what XP did better than Vista? Well, to put it simply, no! This is in fact an example of the way Windows XP inefficiently manages the system resources available, wasting them. This is because the role of RAM is to be cached memory. If it's free, then it only is a piece of unused hardware. This is where Windows Vista and the SuperFetch feature come in. SuperFetch is a memory management technology introduced in Windows Vista to optimize the RAM usage. "Windows SuperFetch prioritizes the programs you're currently using over background tasks and adapts to the way you work by tracking the programs you use most often and preloading these into memory. With SuperFetch, background tasks still run when the computer is idle. However, when the background task is finished, SuperFetch repopulates system memory with the data you were working with before the background task ran,"

OldER Mycroft
OldER Mycroft

http://techrepublic.com.com/5208-12843-0.html?forumID=102&threadID=268459&messageID=2543939 As I see it, SuperFetch is an indication of how Vista was aimed squarely at the business sector of the market. The sector where the software load is more of less defined from the outset, once the 'system' is operational it keeps the same footprint. That way SuperFetch is in it's element. But M$ made the mistake of offering SuperFetch to the GAMING sector too, along with the self-inflated 'Vista-Only' DX10 (which you can now run on XP anyway), but gamers tend to change their software footprints on a steady basis, if for no other reason than downloading demos which are then deleted at some point and replaced with other demos or full games. SuperFetch doesn't have the time necessary to 'learn' in these environments. I'm with you 100%. :)

itpro_z
itpro_z

...that SuperFetch is controlled by the OS, not the application. Perhaps a better analogy would be the old disk cache software that we used to use, which would try to anticipate the files you would need from the hard drive and cache them into RAM for faster access. I can tell you that on my Vista machine, performance has improved over time as the system learns my usage patterns. Common applications load very fast, and the system doesn't bog down under load like XP.

Oz_Media
Oz_Media

Crappy retail reps. In a world where salesmen are few and far between, we are left with a bunch of underqualified people who can only sell based on lowest price and don't know how to upsell a product. notebooks are sold with the absolute MINIMUM spec to run it in order to keep cost down, again it is all about filling a market full of cheap consumers, they want the lowest price but still teh best machine, sorry it just doesn't work that way. A retailer will make more money selling 2 Gigs of RAM as an upgrade than selling a $1000.00 notebook at 2 points margin. But in the big box, low margin retail world of today, Future Shop, Best Buy etc., they hire kids (for low wages) that rotate bewtween departments so often that they know nothing nor do they care about learning it and helping customers. So the intent of teh manufacturer to offer a low priced, entry level machine, and the retailer to make money sellign the upgrades are completely lost. The manufacturer gets slammed for selling garbage, where it is simply underpowered and teh retailer doesn' tmake any money, thus perpetuating the low cost, low wages scene some more. PLUS I see many people with a heap of ram and small page file sizes. Thier comment is usually that they don't want to use up RAM for the page file. What I think they often don't realize is that teh page file is SUPPOSED to use a good portion of your RAM, so that it doesn't have to constantly read teh slower HD. With even average notebooks now having 4GB of RAM, page file sizes can generally be much larger I find. Sped mine right up too.

inertman
inertman

after reading all of these posts, at the time of this posting, none of your arguments hold any water, except in light of the fact that you're old, senile maybe. it's quite obvious from the article what point is being driven and how yet you are the one picking holes in the air. one of the questions here was 'why does the xp machine show 31 and vista show 84' or some such. the answer was given in the article, because superfetch loads the common code. if you, or others can't actually read an article, why are you posting comments?

ken.meyerkorth
ken.meyerkorth

Seeing that Old Mycroft has plenty of naysayers purporting to 'KNOW' Windows, let me just say that I run over 1000 machines of varying types and skillsets (yes I do mean that) and several different OS's (Windows, Vista, Mac, Linux, Unix) and thoroughly understand the issues here. Pretty plainly, the test machine for this article was very underpowered as the minimum requirement for anything above Vista Home Basic is: 1 GHz 32-bit (x86) or 64-bit (x64) processor 1 GB of system memory 40 GB hard drive with at least 15 GB of available space Support for DirectX 9 graphics with: WDDM Driver 128 MB of graphics memory (minimum) Pixel Shader 2.0 in hardware 32 bits per pixel Home Basic minimum is even worse at: 1 GHz 32-bit (x86) or 64-bit (x64) processor 512 MB of system memory 20 GB hard drive with at least 15 GB of available space Support for DirectX 9 graphics and 32 MB of graphics memory And remember, Minimum means enough resources to turn on the machine and get a minimal interface with no performance. Period. Having said that, now to shoot holes in all these inane arguments. XP consists of mindless stacking of Applications in RAM until RAM runs out period. There is no performance requirement for this process so naturally no relationship to performance exists except how much RAM in is the system. Vista and Superfetch is sort of an anomaly. The OS 'learns' from a users application usage and takes steps to cache common code and data into memory when the app is initially opened. This can be seen especially well in systems running simulations because the more times a simulation is run, the smoother it seems to get on systems running less than optimal (3Gb@32bit) RAM. The article having been written about how superfetch works versus the stack and burn XP method is not to put shiny star like auras around the word Vista in your mind but to educate you on a pretty cool Vista feature that if you are patient and understand the value, you will see that its pretty cool. Nuff said...

itpro_z
itpro_z

Those "little underpowered laptops" probably are short on RAM, making for the disk thrashing. Didn't you ever run XP on low end machines? I have had the misfortune to experience XP on machines with as little as 64 MB of RAM. Talk about disk thrashing! One of the first Vista machines that I set up was a low end Dell with 1 GB RAM and onboard video. Of course it came with a ton of crapware installed as well. After cleaning the machine, I was still not pleased with its performance, so I added another 2 GB of RAM. Wow, what a difference! That machine went from a slow poke to very decent performance. I have since found that 2 GB RAM works quite well with Vista, but as cheap as memory is I bump to 3 GB on many machines. Regarding laptops, since I work in the enterprise market my experience is mostly with higher end corporate class machines. On those, Vista run very well, and SuperFetch performs as advertised.

sderf
sderf

I bought a laptop with Vista on it and it was slower than snails. The laptop came with 1 gig of ram. I installed 1 more gig,still slow. I took Vista off and installed XP and it was like the difference of day and night. Now I can use my laptop with Photoshop, etc as that is what I needed it for. Sorry MS Vista just don't cut it a waste of money for folks that need a computer to do something besides write letters on. Derf

sderf
sderf

I bought an Acer laptop with Vista on it. 1 gig of ram Slow as snails.I added 1 more gig of ram very little help. I took Vista off and replaced it with XP a world of difference way,way faster. I use Photoshop CS3 it made it work way faster,so much so I can't use Vista It is a joke of an OS. I haven't found any of my friends that like it and some did the same thing I did they went back to XP Fred

Oz_Media
Oz_Media

I loke mine just fine. I also use a memory stick (4gig) a lot of the time just for that added speed, if you got it flaunt it, right?

tim uk
tim uk

Howcome the XP screenshot shows 31 running processes while the Vista one shows 84? Seems a bit odd to non-techy me. edit: Oops, this wasn't meant to be a reply to Mark's post.

rbmacdougall
rbmacdougall

My wife and I are two of the unfortunate "near casualties" of Vista ownership when it comes to gaming. As luck would have it, buying two Samsung BW226 gaming LCDs led to both computers dying three weeks later. Rather than try to find parts with the rest of the older equipment, we bought matching Toshiba laptops. Anything we installed ran horribly, from Crysis and Hellgate to WoW and Sims 2. After lots of searching for tweaks, adding a 4GB flash drive to each (2GB RAM out of the box), and shutting off Superfetch, they're finally able to take about anything we've thrown at them. After reading this article, I figured I'd give Superfetch another try. I didn't get far; the HD would thrash violently, games would stutter and lock up, and if they recovered, the framerates dropped to at least half. Maybe after we add more RAM I'll give it another shot, but for now, Superfetch isn't so super.

steveschwab
steveschwab

I certainly agree. It's been my experience that a Vista machine will feel significantly faster than an XP machine with both systems using 2 Gb RAM. Programs that I use often open almost instantly. I tend to open and close programs as I use them so the my user experience is significantly better on Vista.

mswift
mswift

They get their bonuses based on sales. I was recently in Circuit City and the machines they had in the best retailing locations (end of the aisle facing the main aisle, stand alone islands) were all top of the line. They were pushing $900 Gateway machines with quad cores and 5gb RAM running Vista 64. I listened in on several of them helping people looking at smaller machines. They pretty much pushed getting the more powerful machine. I don't think many retail store buyers do their own hardware upgrades.

OldER Mycroft
OldER Mycroft

I've always (at least I think I've always) known this, but it never fails to amaze me the number of users, and retailers, and even some techs, that don't. When you whip the panels off a tower PC, or open the memory box on a notebook PC, then slam in some extra RAM, then cover up the boxes again and power on and boot - a critical component of Windows is unaware that the amount of available RAM has changed, although System Properties reports the increased RAM. THAT is why so many users have undersized Pagefiles. After increasing the physical RAM, you have to adjust the Pagefile manually, even if you just want to leave it under System Control. I've seen users with 4GB laptops that STILL have 768MB Pagefiles and when I mention that their laptop used to have 512MB RAM they look at me like I've got psychic powers. Mind you, the local computer shop 5 miles away from me was responsible for me getting a lot of incoming work when these laptops were suddenly sluggish. The store owner was one of the unwashed - he installed the upgrades but didn't know either! :^0

Richard Turpin
Richard Turpin

My sentiments entirely.....read the bloody article properly then make constructive comment.

OldER Mycroft
OldER Mycroft

Because #1 It wasn't me that asked that question AND #2 It was a sensible question anyway. All day now I've been listening to how good this Vista system is compared to XP, how clever SuperFetch is (and that XP ain't got it of course), how important it is to have "all things equal" for 'benchmarking' purposes, da dah, da dah, da dah. Now you pipe in with your childish insult when in actual fact you could've maybe realised that by upping the number of processes running on the XP machine, a more balanced comparison could've been made. But of course I digress, that would never happen in a thread that should clearly have been entitled: The TR Vista-SuperFetch Appreciation Society / Fanclub. How's the inert brain man, oops sorry, - brain, inertman? ;)

OldER Mycroft
OldER Mycroft

That the user NEVER changes tack as it were, opting to load and run different software, that SuperFetch has hitherto never 'seen'. ;)

OldER Mycroft
OldER Mycroft

So, by definition, the article purports to advise that SuperFetch works its magic within a 1GB environment. THAT was my point - SuperFetch doesn't have anything to work on inside a 1GB machine. Therefore my question to the author is - why cite a 1GB-comparison between XP and Vista when you know that is not true to kind. By the same token, had the comparison been illustrated as XP and Vista machines both running 3GB, the author would've ruptured his own argument since a 3GB XP machine walks all over a 3GB Vista machine in terms of speed and XP doesn't HAVE SuperFetch. The comparisons drawn in the article are foolhardy since any fool knows that a 1GB Vista machine can't use SuperFetch, yet the article suggests that it does.

rbmacdougall
rbmacdougall

While I'm not positive, I'm thinking that it wasn't the new monitors themselves. The systems were around 8 years old and tons of gaming (no overclocking though) had them tired out. When I looked at the motherboards (Gigabyte 8IEXP with i845E chipset and 2.4GHz Northwood P4), they hadn't actually popped some of the capacitors, but had flux fizzing out of them. It almost looked like dried caramel. We've been pretty happy with our Toshiba X205-S7483's, but their video driver support is of the "bite-eth mightily" level. Laptopvideo2go has been a big help.

bkrateku
bkrateku

I had heard that Vista is not the OS of choice for games compared to XP. I guess that's why. SuperFetch makes sense for casual and business use, but not gaming or anything too intensive it sounds like. I had 1 GB of RAM in my work laptop and Vista ran good, but not great. I upped it to 2 GB and it runs much better, especially when working with several apps. I've heard there is a decent jump in performance if you go to 3 GB depending on what you run on Vista, but it wasn't cost effective to go to 3 GB at the time. I'm curious...what do you think caused the gaming LCDs to be the reason for the death of the old PCs? Just curious.

inertman
inertman

if you actually read the post, similar to my point that if you actually read the article, i didn't say you posted that question, it was just an example of the rediculous points you were making. and yes, you do digress, way off point. i don't recall anyone saying 'how good this vista system is' or 'how clever superfetch' is, only how it works and why it makes a vista system that's been in use work faster than when you first initialise it. but of course you haven't really read anything here it seems, not really, or else your comprehension is really limited. but i actually doubt either, it really just seems that you want to get on the vista bashing band wagon. oh, and as for 'a more balanced comparison', get a clue, there wasn't any attempt at benchmarking, only an attempt to show what is going on with superfetch and how the ram is getting used up more in the vista machine. get it? and you think my brain is inert, you really need to re-read everything in the article and every post here, including your rediculous posts! by the way, it wasn't a childish insult, just an observation and a possible explanation, senile is about the only way to explain what you've written so far.

Mark W. Kaelin
Mark W. Kaelin

You are reading much more into this than there is - I have made no claims about XP being worse, better, or the same as Vista -- You have! I can only report that I have had no trouble with Vista being slow. You claim XP is faster, I just thought you could show us. But then, by your own admission, you don't have Vista -- so, I remain confused as to how you know XP is faster. Instead of addressing the issue you call me obtuse and question my comprehension of reality.

OldER Mycroft
OldER Mycroft

"[i]I just wanted to see your evidence of that, because I have not had the same experience.[/i]" Yet you previously said: "[i]what I was driving at with the suggestion that you write it up for us, is that I have never seen Vista run slow compared to XP on my system.." "..When I was researching a new PC, I saw lots of discussion about how much slower Vista is than XP so I made my new gaming rig dual-boot. But I never have any reason to use the XP partition -- Vista runs all my games just fine.[/i]" Sorry, but I've simply GOT to ask - if you've NEVER run your games on the XP partition, how can you possibly know that Vista is doing better ?? That was my point when I said Crysis/DX10 running in XP, with everything turned on returns a count of 54fps. But you wouldn't know that since you've never run any games in your XP partition. So where is your benchmark for the Vista performance that you're so happy with? "[i]You seem so sure of your statement that XP is faster than Vista, I was hoping to see it for myself because I have not seen any evidence of that on my personal PC.[/i]" Might I suggest you take a walk on the wild side, and try booting up XP for playing your games. Odd that the capability is right under your nose, yet you expect me to run the comparison for you. ;)

Mark W. Kaelin
Mark W. Kaelin

I am referring to this statement you made in an earlier post is in this thread: "By the same token, had the comparison been illustrated as XP and Vista machines both running 3GB, the author would've ruptured his own argument since a 3GB XP machine walks all over a 3GB Vista machine in terms of speed and XP doesn't HAVE SuperFetch." I just wanted to see your evidence of that, because I have not had the same experience.

Oz_Media
Oz_Media

Practice what you preach

Oz_Media
Oz_Media

Old Mycroft, people like this need to take the entire bottle and do the rest if us a favour. Culling the herd and all. Mr.zimbabwe: RAM, if the user is not computer savvy they are at the mercy of the store sales person to give them the facts. Unfortunately, as noted in another post, retail staff are usually not at all trained on anything but how to use the POS system and an unsavvy buyer should find a different outlet. They SHOULD be recommending the right upgrades based on your computing needs. Someone buying a 1GB notebook and Office 2007 software should be told to get more RAM. Someone buying that same notebook that only types letters in notepad and checks email, surfs the web a bit etc. will do just fine with 1 GB. There are poeple available when you purchase a notebook that should help with these issues, no need to be computer savvy unless you are simply looking to buy cheap crap from a big box retailer and can configure your own hardware needs. Anyone without knowledge is best off to buy from a small local shop with good support and staff with experience, not from Best Buy, Target etc.

OldER Mycroft
OldER Mycroft

When M$ says that isn't possible. However, you appear to be slightly digressing now: "[i]As I have said in other posts, and what I was driving at with the suggestion that you write it up for us, is that [b][u]I have never seen Vista run slow compared to XP on my system.[/b][/u] People keep telling me that it is a fact, but I can find no evidence.[/i]" Uh huh! No evidence eh? So how come you're sitting with a 4GB rig then? [b]Try a package tour trip back to reality Mark.[/b] Whip out THREE of your 4 DDR RAM sticks (assuming you're running with 4 slots occupied) then see how your dream rig behaves when it's populated with 1GB running at 50% efficiency single data rate. Don't worry though - you'll be OK - Greg proved that in his article didn't he? SUPERFETCH will see you alright coz it can run in a 1GB system! ZZzzoomm !! What was that? That was XP...! THEN justify the comparison made in the original article.

Mark W. Kaelin
Mark W. Kaelin

No, it is my gaming rig - it has 4GB. When I was researching a new PC, I saw lots of discussion about how much slower Vista is than XP so I made my new gaming rig dual-boot. But I never have any reason to use the XP partition -- Vista runs all my games just fine. You seem so sure of your statement that XP is faster than Vista, I was hoping to see it for myself because I have not seen any evidence of that on my personal PC.

OldER Mycroft
OldER Mycroft

And running in 1GB you can see no difference in speed between Vista and XP? I think not! ;) Addendum: Greg's article states: "[i]Let?s start by comparing the Performance tabs on [u]similarly equipped XP and Vista systems[/u] just to see the difference.[/i]" That didn't suggest ONE machine, first running XP then running Vista. 'Similarly' is not 'Identical'.

OldER Mycroft
OldER Mycroft

You are now off down the Yellow Brick Road on a solo quest to pick holes in air. Your posting bears no relation to my previous entry. :D

OldER Mycroft
OldER Mycroft

Perhaps your listing as Technical/PC Support is a shade away from the actual fact? Perhaps that also explains your blinkered view of the present discussion: with every keystroke you slip further away from the subject. "[i]"For one thing I don't own any Vista-based systems, preferring my own three XP machines" Wow... 3 of them? All the same? How do they compare to each other? (Please note dripping sarcasm)[/i]" Sorry, but where the hell did that outburst come from? [b]I wasn't aware that I had published any article on TR featuring screenshots of the relative performances of my three XP-based machines!![/b] If you're gonna attempt to shoot me down in flames, at least check your ammunition status first. ;) "[i]Regardless, it was an equal load for each machine. This is what benchmarking is all about; [u]equal in every way except for the software/hardware that you are benchmarking[/u]. This is computing 101, dude.[/i]" Regardless? ...Regardless of what? You've just made out that my standpoint has no basis in fact! It is NOT an equal load - not when Vista's true potential needs at least 2GB to be unleashed! [u]As for your inept comment:[/u] what could possibly be EQUAL if it is "[i]except for the software/hardware that you are benchmarking[/i]" - s/w - h/w IS a computer! There isn't anything left to BE equal. At present I have 1,337,228 bytes available from this 2GB XP Pro system, I'd like to 'see' the comparison with a 2GB SuperFetched Vista system. I expect the vast majority of Greg's readership would have preferred that also. Wouldn't it be a humdinger if there was still only 5MB left, before attempting to open a program NOT in the cache!!

OldER Mycroft
OldER Mycroft

So what is the problem in running them both with 2 or 3 GB of RAM, huh? Unless it is dawning on you, what I already know. Don't forget - all that is under discussion here is [b]SuperFetch.[/b] [i]Don't get confused with how many other details you don't agree with.[/i]

Mark W. Kaelin
Mark W. Kaelin

Greg ran the "benchmark" like I would - same PC, running the same applications - once under XP and next under Vista. I have a dual boot machine and that's what I would have done. It may not be absolutely the best way to do it, but it is quick and illustrates the point. As I have said in other posts, and what I was driving at with the suggestion that you write it up for us, is that I have never seen Vista run slow compared to XP on my system. People keep telling me that it is a fact, but I can find no evidence. I was hoping you had some confirmation we could show.

Timbo Zimbabwe
Timbo Zimbabwe

"But an educated and experienced devil's advocate." You don't sound educated. "For one thing I don't own any Vista-based systems, preferring my own three XP machines" Wow... 3 of them? All the same? How do they compare to each other? (Please note dripping sarcasm) "He then strayed further into the quagmire by citing that each machine was only running with Internet Explorer and Paint - hardly a likely system load for anyone venturing onto the 'net" Regardless, it was an equal load for each machine. This is what benchmarking is all about; equal in every way except for the software/hardware that you are benchmarking. This is computing 101, dude. You call yourself an IT consultant? I would never hire you....

Timbo Zimbabwe
Timbo Zimbabwe

"M$ launched their new super-all-dancing system, supposedly 'enhanced' with SuperFetch. Your point that you "wouldn't think of ordering a system with less than 2 GB regardless of the OS" is not a stance based on marketing or product instructions, but borne out of computer knowledge and experience." Oh, you must be a basher (M$). That explains a lot. Complaining to complain? Listen, it would not be much of a comparison if machine A had a dual-core CPU and 2 GBs of RAM and machine B was a 133 with 512 MB of RAM, would it? In order to make a proper comparison, you need to create your benchmark with equal hardware, not best parctices. If you do not understand this concept, you are in the wrong business.

Timbo Zimbabwe
Timbo Zimbabwe

"Therefore my question to the author is - why cite a 1GB-comparison between XP and Vista when you know that is not true to kind." So, we should compare apples to oranges then? I mean, comparing 2 PCs with the same hardware but different OSes is just ridiculous, right? Yo, bubba... think!

OldER Mycroft
OldER Mycroft

But an educated and experienced devil's advocate. For one thing I don't own any Vista-based systems, preferring my own three XP machines, and realistically draw the line at purchasing a landline-telephone to pander to the likes of TechRepublic. I just think that Greg Shultz rendered his own proposition inadequate by illustrating his article with the wrong amount of RAM. He then strayed further into the quagmire by citing that each machine was only running with Internet Explorer and Paint - hardly a likely system load for anyone venturing onto the 'net (no antivirus, no antispyware, no 3rd party firewall, et al). The initial idea was full of promise but the eventual premise was fatally flawed. Maybe he DID run it with THREE megabytes but then found there would be no basis for the article! ;)

OldER Mycroft
OldER Mycroft

It is not a matter here, of whether Vista is faster than XP, or XP is smoother than Vista, or if new hardware is better serviced by Vista, or even if XP can be seen to be getting slower with the unstoppable passage of time. FORGET all this system bravado! The point here is SUPERFETCH. Period. M$ launched their new super-all-dancing system, supposedly 'enhanced' with SuperFetch. Your point that you "[i]wouldn't think of ordering a system with less than 2 GB regardless of the OS[/i]" is not a stance based on marketing or product instructions, but borne out of computer knowledge and experience. Bitter experience in most cases. However, even at this stage in the development of Vista, Micro$soft still maintain that 1GB is an acceptable level of RAM for systems pre-loaded with Vista and thses pitfully under-powered machines are still featured daily in computer advertising. Now to return to my original point - the original article favoured to illustrate SuperFetch being used supposedly in a 1GB Vista system. I still maintain that this conjecture is crooked, having little or no hard evidence that it would ever work in practice. As such, and standing by what was 'published', I maintain the author made a mistake using a screenshot comparison of Vista and XP both running with ONE gigabyte. As for XP walking all over Vista, I can only state what I know to be the case. Hell, my XP rig can even run DirectX 10 within certain games so I'm not missing out on Vista's 'special qualities' either. My comparison was based on system-to-system comparisons only. It was you that felt the need to introduce the use of 'new hardware' in order to bolster your argument - it is simply unrealistic to argue one system's advantage using hardware that the other system was not designed to use. Neither you nor I are running with systems that the average punter in the street is likely to be convinced into buying, only to get it home and boot into what is basically a top-heavy, under-powered, resource-hungry, SuperFetch-incapable, system.

itpro_z
itpro_z

We have both Vista Business and XP XP3 running on identical machines on our network, some with 2 GB and some with 3 GB, and I can tell you that XP doesn't "walk all over" Vista on modern hardware. Vista holds its own, and even outperforms XP at some tasks, especially under load. Vista performance improves over time, while XP seems to get slower. I personally use and support both, and much prefer Vista. I have ran Vista with 1 GB, and if properly set up it runs OK. I have had new XP machines with 2 GB that ran like pigs, so don't try to tell me that XP doesn't have to be tuned as well. SuperFetch works with whatever RAM it is given, up to the 32 bit limit (unless you are running 64 bit, of course). Considering the cost of RAM these days, I wouldn't think of ordering a system with less than 2 GB regardless of the OS.

Mark W. Kaelin
Mark W. Kaelin

Old Mycroft, would you like to perform the 3GB experiment for us and write it up in an article with screen shots. I would like to see what you describe.