Wednesday, February 17, 2010

(WCPI): 85% of Win 7 PCs are Underperforming

Updated: Thanks to an industrious reader, we finally have outside confirmation that our understanding of this issue was dead on. Yes, Virginia, this means we were right and you were wrong. So before you go posting yet another idiotic comment parroting yet another clueless blogger from yet another obscure corner of the Internet, please read our editorial titled “What took you so long?


Windows XP holdouts take heart: New data from the exo.repository shows that better than 8 in 10 Windows 7 systems monitored by the exo.performance.network are running alarmingly low on physical memory. And nearly the same number are demonstrating significant delays in I/O processing – ostensibly due to heavy virtual memory activity as Windows compensates for insufficient RAM.

Note: We’ve published a new blog entry that explains our methodology in depth. If you have questions about how we came up with the numbers in this post, this is the place to look.

This disturbing trend was identified by exo.performance.network researchers as they compiled data in support of the newly released Windows Composite Performance Index (WCPI). Computed values for the Peak CPU Saturation, Peak Memory Pressure and Peak I/O Contention indices – all specialized components of the more generalized WCPI metric – showed that Windows 7 systems had much higher memory pressure values than systems running Windows XP. This was true despite the fact that the Windows 7 systems in question had, on average, nearly twice as much (3.3GB vs. 1.7GB for Windows XP) RAM installed.

 

Win XP

Win Vista

Win 7

CPU Saturated

32%

45%

44%

Low on RAM

40%

80%

86%

I/O Backlogged

36%

83%

85%

Table 1 – WCPI Percentages by Platform

Such a discrepancy in recorded memory consumption is not surprising given Windows 7’s much higher RAM footprint. However, when you factor in the generally more advanced hardware that’s being deployed with Windows 7 (over a third of the legacy XP systems sampled have CPUs that pre-date the launch of Windows Vista), the realization that Microsoft’s new OS is quickly expanding to consume all available resources is still quite alarming. Add to this Windows 7’s higher CPU utilization (44% of sampled Windows 7 PCs are CPU bound, vs. 32% for XP) and it’s clear that Microsoft’s latest and greatest is well on its way to overtaking Moore’s Law.

Note: The above data points were compiled from the hundreds of millions of system and process metrics records that have been uploaded to the exo.repository by participating member sites. You can keep tabs on all of our research findings by visiting our web site: www.xpnet.com. There you’ll find a wide selection of interactive chart objects and free monitoring tools that you can use to compare your own systems to the WCPI and similar independent research metrics published by the exo.performance.network.

14 comments:

Dimitri said...

Have you not considered the different memory management and caching approach in Vista and Windows 7, as described here?

Random Alias said...

You do know that you are bloody idiots, right? You managed to get to the front page of slashdot, where thousands of people get to see how little you know about win7.

CACHING. Just because the ram is in use doesn't mean that it's overloaded.

Evil_Merlin said...

My goodness, and you claim to be a CIO? You don't even know how Windows 7 records memory in use and how it differs from the process used under Windows XP it seems. Do us all a favour. Read up on it. And get a new job. Maybe flipping burgers, because your method of recording memory usage in Windows 7 boxes is seriously flawed, flawed enough to be nothing more than a lie.

Unknown said...

You guys have generated alot of confusion because of this post and an article in Computer World.
Do you understand how Superfetch works?
Read here: http://www.osnews.com/story/21471/SuperFetch_How_it_Works_Myths

Please do not jump to conclusions about your data without understanding how Windows 7 use RAM.

Frank Bulk said...

I don't follow your assessment. So Windows 7 is more efficient in its use of memory, leaving less free. Isn't the point of memory to use it? And since when wasn't virtual memory a part of a modern OS?

I think this is worth further discussion.

Frank

fishyfool said...

Better check your facts.
The overwhelming consensus is that you are factually incorrect at a very minimum.

Unknown said...

There is so much wrong with this article I don't even know where to start.

Obviously it's not "Windows" using 3GB of RAM. That doesn't even make sense. It's applications that are using up this extra memory. And I can tell you EXACTLY why user's of Windows 7 have a higher memory consumption compared to other operating systems. It's because Windows 7's multitasking ability with the new task bar is so effective. I put Windows 7 on my computer illiterate roommate's laptop, and the one thing he said was "this is great, I can do so many more things at once" because it is all organized better in the task bar. He had multiple documents and browser windows open while checking his email and listening to music. So a better way to phrase "Most Windows 7 PC's Max Out Memory" would be to say "Most Windows 7 PC's Allow User's To Get The Most Out Of Their Hardware".

The whole purpose of a new operating system is to better utilize the new improved hardware inside computers. I mean, we could compare Windows 7 to Windows 3.5 and say that 99% of Windows 3.5 computers only use 1% of available memory in today's computers. But that's the whole point. Computer's today have upwards of 2GB of RAM, so obviously, let's actually put it to use. A "perfect" operating system would in theory use 99% of available memory and always have the perfect pieces of data stored in memory as to anticipate the user's actions and make the computer nearly instant since RAM is much faster than a hard drive.

And to add to all of that, I simply do not believe for a second the figure they give that "86% of Windows 7 computers use 90-95% of available RAM". That is a lie, or a seriously flawed test. At this very moment, I am using Windows 7. I have one Google Chrome browser window open with 4 tabs, one Windows Explorer window open, Winamp playing lossless FLAC, Microsoft Outlook, and a PDF open in Foxit Reader. I am using 1.00GB of RAM total right now, and I have 2GB total RAM installed. That is the entire Windows 7 OS, and over 5 applications running, some with multiple tabs, all within 1GB of RAM. If I look at memory consumption of just the applications, I'm looking at about 280MB of RAM, and that's rounding everything up. So I'm looking at about 720MB usage from Windows 7. And that includes all background running applications (Audio Driver, Graphics Driver, Dropbox, Antivirus, OneNote Clipboard, etc.). So let's say we double Windows 7 usage up to 1.5GB. They said the average Windows 7 computer has 3.3GB of RAM. Let's round that down to 3GB. So we now have 1.5GB of RAM available. Now let's round my very average application RAM usage up to 300MB. So that's 300MB to run 5 applications at once. Now I'm going to call my 4 Chrome tabs an extra application to be fair. So about 6 applications at 300MB. This would be considered a very average number of programs running at once, probably a fair bit above average actually. We would have to multiply this usage by 5 times in order to use all of the remaining 1.5GB of RAM. That's about 30 different browser tabs, music players, email clients, and documents open at once. I think we can all agree that this kind of usage is very, very far from average. The only other circumstance in which you would use this kind of memory consumption is if you're converting video, compiling source code, playing video games, or something else really computationally intense, which again, is obviously not average.

So I'm going to go as far as to call this article a downright lie. It is either very, very biased, or the science behind the experiment is very poor. Terrible article.

Unknown said...

You people are idiots. How can you know so little about how Windows 7 works?

There is a difference between cached and used memory. Check the "Available" memory, that will give the real amount of ram that your system has free to applications.

Either you are the most incompetent company ever, or you are trying to deceive people in order to sell them some bullshit "tuneup" software.

Unknown said...

This is an absolutely worthless article. Windows 7 utilizes memory that isn't actively being used for its Superfetch algorithms, that preload frequently used programs and such, so when you click on their icon to start them, startup is basically instantaneous. Windows 7 learns (like Vista did) what you do most commonly and predicts your actions. If you stray from those common occurrences, windows effortlessly frees memory it was using so it's available for the system to utilize.

Benjie said...

SuperFetch was discussed back in early 2005 when it was release with a Vista Beta.

In other news, quad core 2.66ghz cpu not as fast as 10ghz chip on single threaded apps. Intel and AMD are trying to screw us over with useless multi-core chips.

Rich said...

This post is terribly, terribly wrong. You have all just ousted your-selfs as idiots on a lot of tech blogs lol. Just saw this a second time on ars.

Unknown said...

A 10Ghrz chip would require too much power and produce too much heat to be a viable consumer chip. The theory with multicore computing is with mulithreaded apps, which is cheaper and faster than any single core solution.

just fyi

Unknown said...

whoever posted this article is a grade A+ idiot and needs to go back to school to keep up with the times.

Anonymous said...

Could you perform this test also
Windows Server 2008 (R2)