Windows 7 vs Windows Vista Performance Comparison

With the release of a new operating system, one of the first questions that pops into your head--besides how much it is going to cost you--is if it will increase the performance of your machine. Newer operating systems are usually more optimized, designed to start quicker, load programs faster, and increase performance with better drivers. With that question in mind, we took two identical Lenovo ThinkPad T400s loaded with Windows Vista or Windows 7 and compared every facet relating to performance.

To make this article relevant to consumers, we picked two identical systems with factory images (operating system installatons). One included the latest configuration of Windows Vista Business and the other Windows 7 Professional. They were both manufactured on 9/10/09 and are only four serial numbers apart. These systems match even down to the most basic internal hardware. To compare the system performance of both machines, we created a level playing field by removing the same bloatware from each machine including AV software and gave each laptop the same custom power profile in the Lenovo power manager.

The Lenovo ThinkPad T400s configurations included the following hardware:

• Windows Vista Business SP2 32-bit or Windows 7 Professional 32-bit
• Graphics: Intel X4500M Integrated
• Screen: 1440 x 900 WXGA+ LED Backlit (Matte finish)
• Processor: Intel Core 2 Duo SP9600 (2.53GHz, 1066MHz FSB, 6MB Cache)
• Memory: 4GB DDR3 RAM (2GB x 2)
• 128GB Toshiba SSD
• Optical Drive: DVD+/-RW
• Wireless: 802.11a/b/g/n, Bluetooth 2.0
• Battery: 6-cell 11.1 44Wh

General Performance
In the first section of the performance review, we compare daily activities of each system. This includes tasks such as boot times, shutdown time, sleep time, and wake from sleep time. These are activities the user has to sit through multiple times a day, meaning that a few seconds here or there really adds up over time. Both systems were set to automatically log in, without a password prompt to delay the process. The timer was started either when the power button was pressed, or the appropriate button inside the operating system was clicked. For the power-on sequence the timer was stopped when the system was at the desktop, idle and waiting.

Measuring the boot times side by side was pretty interesting, since our Windows 7 system turned on almost 30% faster than the Vista machine. We expected both machines were going to be quick with their relatively clean installs and SSDs, but coming in under 30 seconds was impressive for the Windows 7 system.

Startup Time

Shutdown time was equally impressive, with the Windows 7 configuration completely powered down in half the time over Vista. Timers in this case were stopped when the activity LEDs were off.

Shutdown Time

Sleep and wake times are important for users who want to quickly power down their machine--say at the end of a class or meeting--but might not want to lose what they were working on. In this test as well, the Windows 7 system had another clear advantage over the Vista configuration.

Sleep

Wake

For standard system activities, Windows 7 easily takes the lead. Clocking in with consistently faster times over and over again, there is no question in our minds that at least some areas have been optimized compared to Vista.

Battery Life
Our next test is battery life, which plays a major role in day-to-day use and also gives us an indication of how stressful background activities are on a notebook. In this test we installed FireFox on both systems with the ReloadEvery plug-in set to refresh the same webpage on each system every minute. This was to duplicate slow but steady usage which might be duplicated inside a classroom or reducing the boredom in a lengthy meeting. For this test both systems were configured with the same power profile inside the Lenovo Power Manager. The primary settings included the processor set to adaptive mode, screen brightness to 10/15, and all idle timers disabled. Inside BatteryMon, both systems reported a battery capacity within .6kWh of each other.

Out of all the tests we ran, I think this was the most interesting. I say that because of the sheer fact each system shutdown after 5 hours and 56 minutes, within 15 seconds of each other. Technically one system shut down with 5:56 on the counter while one showed 5:57, but for us it was close enough to be called equal. Both systems stayed at about 6.5 watts of power consumption during the test, without many background tasks coming up to stress the notebooks.

Software Performance
Application performance was a toss up with some software taking longer to load in the Vista environment and another taking longer inside Windows 7. We used AppTimer to measure the length of times GIMP and iTunes took to start up on each computer. After running the test multiple times on each system to find the average, we found that iTunes takes a bit less time to load in Vista and GIMP takes less time in Windows 7.

iTunes load time

GIMP load time

Video Playback Performance
Video playback and its stress to the processor was another area we checked, and in this scenario we used the latest CCCP codec pack with Media Player Classic HomeCinema, with Divx 7 installed x264 decoding. The test file was the 1080P trailer for the movie Up!. CPU utilization seemed to be better in Windows 7 (if only slightly), as shown by a more stable readout in the Task Manager. In Windows 7 if you go with Windows Media Player 12, which supports hardware decoding with the Intel X4500 chipset, the performance difference is huge. Instead of 15-25% CPU utilization, it is between 0-4% on average.

Vista

Windows 7

Windows 7 using WMP12

Synthetic Benchmark Performance
The last half of our performance comparison guide covers synthetic benchmarks, including wPrime, PCMark05, PCMark Vantage, and 3DMark06. Both systems had all available Windows Updates installed and we used a non-Aero basic theme during each test. The results we found were very surprising. After the pretty significant advantages Windows 7 showed in day-to-day performance in prior tests. The only area where Windows 7 performed equal to or better than Vista was in 3D performance ... which had marginal gains at best.

PCMark05 measures overall system performance (higher scores mean better performance):

PCMArk Vantage also measures overall system performance (higher scores mean better performance):

3DMark06 measures overall graphics performance for gaming (higher scores mean better performance):

Wprime processor comparison results (lower scores mean better performance):

HDTune Windows Vista:

HDTune Windows 7:

While most of the synthetic benchmarks might have the disadvantage of not being optimized or designed to run inside Windows 7, wPrime shouldn't have those problems. Even it ran slower inside Windows 7. To simulate gaming performance on each system, we used Real-time HDR IBL (rthdribl) in its default startup mode running at 640x480 on the desktop, with FRAPS recording the average framerate. In this test Windows 7 showed a small lead in framerate over Vista. The increase would probably be more significant with a better graphics card.

rthdribl FPS (higher scores mean better performance):

Conclusion
After testing all aspects relating to performance between two identical system equipped with Windows Vista and Windows 7, we found each version of Windows performed differently under certain circumstances. In day-to-day operation Windows 7 easily beats Vista with a quicker boot times, shutdown times, sleep times, and wake times. In some of those tests it was almost 50% faster, a huge lead when you consider how many of those cycles the computer will go through in its lifetime. Video performance is improved in Windows 7 if the user switches to Windows Media Player, where it now natively supports hardware-decoding with the Intel X4500 chipset. Battery life was unchanged with each system getting nearly the exact result in our side-by-side test.

The main area that seems to suggest a disadvantage for Windows 7 is overall system performance when tested with synthetic benchmarks. It is too early to tell if this is non-optimized benchmark software or driver related hiccups inside Windows 7. What we do know so far is 3D performance seems to be improved by a small margin; a plus for gamers. As more and more Windows 7 systems come in for review we will see a better picture of how it performs in our tests, but we can easily say for the average user, Windows 7 will feel much faster in day-to-day activities.

 

So this is what you were talking about "crunching day" on the G51j thread...lol

Thanks for the info, it is quite interesting, but I dont get what you meant here:

And isnt Vista faster? Is that a typo?

 

hmm... I'm going to have to install win 7 this weekend, I love free stuff from the University!!

 

Thanks for this detailed review. I won't be upgrading to Win 7 with my current computer. Might as well get Win 7 when I change computers in the future

 

What about a comparison to XP and 32 vs 64 bit?

 

I expanded your quote a bit for a better explanation. I think they typoed the period after the word tests, that should probably be a comma instead, thus: "After the pretty significant advantages Windows 7 showed in day-to day performance in prior tests, the only area where Windoes 7 performed equal to or better than Vista (in the synthetic benchmarks) was in 3D performance ... which had marginal gains at best." Which is why the results were so surprising.

 

can i ask if recent updates are installed installed in each os during your review?

since the rtm version came out, im using it in my laptop and im loving it. it feels more responsive than vista. windows is more fun to use if you ever switched from vista to win7. definitely win7 is the best software product of the year.

 

The PCMark scores were the most interest to me.

Interesting how Vista got better scores..........

I noticed how the graphs are confusing. On one graph, Vista is orange and Win7 is blue. On another graph they are switched.

 

The 3DMark06 benchmark graph is screwed up. Good review. Can you elaborate on your video playback tests? AFAIK, W7 comes with codecs that enable hardware acceleration by default. Not sure if the third-party media players would take advantage of this, but I'm sure Windows Media Player would...

 

Task Manager is suddenly a benchmark? This is a good quality review, except for that part. It tells the reader nothing.

 

The orange color represents the faster machine, sorry if that wasnt clear.

 

They are too close to show anything but it is a valid benchmark to show CPU utilization when doing video playback.

 

We are rerunning the video test currently, since WMP11 in W7 is allowing hardware decoding whereas nothing on the Vista side is allowing it in either MPC or WMP. The difference is huge though, 15-25% compared to 0-4%. MPC only seems to take advantage of ATI/NVIDIA cards, not the Intel X4500.

 

Basically what I asked why is it that Win7 was better on scores, if Vista had higher numbers, or did I miss something with the colors?

I agree with you, on both things.

Perhaps switching for Win7 blue and Vista orange through the tests? It was kinda confusing.

 

Day-to-day performance was better on the Windows 7 side (boot, shutdown, sleep, wake, some program launching) whereas Vista seems to have the upper-hand on synthetic benchmarks.

 

Oh! Thanks! That is what I was wondering.

I think the lower scores are due to driver support not being as good in Win7 yet...PCMarck05 Vista beats Win7, but Win7 beats PCMarkVantage.

 

This is the taskmanger readout while playing the trailer inside WMP11 in Windows 7. Significantly better, but Vista would be the same if something supported the X4500 for hardware decoding. Attached are the new and old graphs.

 

It's not a valid benchmark since it doesn't quantify anything. As you can see in the report, even the writer couldn't arrive at a definitive conclusion in the same manner as his other tests. "It looks more consistent" is hardly useful to the reader. To put it bluntly, it is patently pointless.

 

these numbers are better than I thought.

Anyway to slim down that fat taskbar in Win7?

 

Right-click TB > Properties > check the "use small icons" option

 

Ok, I understand your gripes, primarily since it was so close in that test it was hard to show how one was better than the other over a set period of time.

Using Performance monitor I reran the video test on both inside MPC, the Vista machine got an average of 22.3% CPU with a low of 16 and a high of 31, where Windows 7 was at an average of 21.7, a low of 14 and a peak of 28.

EDIT: Can to tie a CPU reading from one unique application instead of overall in perfmon?

 

Why don't you use Paint.NET or photoshop and overlap the two graphs? That could help.

I also think that it would have improved with ease of reading if you has assigned colors to each OS, but I understand why you did it how you did.

It's a very nice comparison to see. One think I would've done different if I was doing the test was to use a system with hybrid graphics - to test integrated and dedicated cards.

 

Add counter Process/% Processor time.

Or whatever the name is under Process. Just select the thread under Instance.

 

I really would love to see it compared to XP with SP3!!

A lot of people still run XP and are lost if they should upgrade or not!

Thanks for the nice review! (next time, I would make one color for one operating system in the graph - not one color for the faster one..confusing)

 

I am pretty sure the most awesome thing possible would be if some company sent me three identical notebooks, XP to Win 7, and each had some mid/high-range GPU.

 

Does the lower cpu utilization result in better battery performance for win 7 while playing video?

 

Good question. Yes it does depending on the content being played. On the Vista notebook when watching 1080P video, power consumption jumps to the 15-16w range during playback, where on the Windows 7 side it stays right around 11w. When playing 480p/SD content there is only a 1-2w difference between the two notebooks.

 

Thanks for the info. Seems like it makes a huge difference for hd movies. Might save more than an hour over the course of a full battery cycle.

 

Very interesting. I wonder if this is going to lead to poorer gaming performance.

 

So far it seems 3D has a small lead, but as we review more and more Win 7 gaming notebooks though we will be able to find out.

 

Dont think so. Perhaps the lower scores were due to the softwares not being fully/properly supported by win7, considering it is a new OS, and Vista has been around for a lot longer.

 

W7 service pack one will push it even further ahead in each category.

 

That proves my point that wprime runs faster in vista. I have a thread on this somewhere.

 

Excellent review. It's great that you were able to compare two systems that are for all intents and purposes identical - it makes it a much more valid comparison. There's a good number of tests, they are varied, and they are of practical items. The only flaw I can see is that the orange/blue on the graphs don't consistently mean the same operating system, but once I realized it and knew to check the label to be sure, it wasn't an issue. Certainly a small complaint given the overall quality.

This is also the first really interesting new news I've seen about Windows 7 since its launch.

In an ideal world they'd also send you a traditional hard drive system (what I suspect most people will have), and perhaps an XP system or x64 variants as well, but I'm impressed just that you were able to get two such machines. It's already a very good comparison set.

mew1838: The forum wPrime tests, most popular in 2008 IIRC, definitely showed Vista as having an advantage over XP in wPrime. The primary thread, Bye Bye Super Pi, was pre-Windows 7, but you could use it as a source for Vista and XP times, both of which there were many of.

 

Thank you very much for the excellent write up....but I am puzzled as to why the tests weren't done in 64 bit OS.

Really, just dumbfounded.

 

You and me both heheh but lest you think it was something nefarious, I could have saved the author a lot of time as, having run win7 rtm vs. vista in 64 bit versions AND 32 bit versions, on about 7 different lappies ranging from msi's u100 to a fully pushed and prodded Gateway FX, I found that heads up, vista almost always *won* when it got down to actually doing anything, and that 7 was faster in the 'show' categories, i.e. start up, shut down etc. If you were to employ even a moderately tuned vista install, it would nearly tie 7 in the "show" cat's and beat it pretty much in every other category....

Now "How can this BE!!!!' you exclaim...well, simple, Win 7 IS Vista, plain and simple. MS is trotting out the same old emperor in shiny new clothes trying to convince the world it has a totally new op system which is plain and simple bovine excrement. There's a reason they could shove this thing out the door in one year,,all they did was rearrange the menu's, throw on some new paint, and rearrange the load sequences. It's amazing it took them a whole year to do that. Now, that may b a bit harsh, as there are some tiny new bits BUT, bye and large, its a cleaned up Vista with some of the bugs taken out, and a bunch of code that almost no one uses, removed or otherwise delayed loading only when required thus the faster load times. Win 7 races to the start screen, then continues to load and churn in the background while all the fanboys and fan mags enlisted to push this thing "oooo' and 'ahhhhh' over the aero interface...making an appearance several seconds sooner...
hold me back...

The really interesting part will come with full implementation of OCL etc. And security seems to have been cleaned up, but peculiarly so. It can be argued that the default settings now render a machine LESS secure than Vista. All in all, as many reviewers have already observed, its "Vista as Vista should have been".

Further, i'm surprised no one is commenting on one of the other apparent new additions.

Win7 appears to record essentially everything you do on your computer, and everywhere you go, in files not easily found, identified, or read. It encompasses what appears to be the single largest data mining scheme I've seen, encoded in a non std. file format unknown to me (which doesn't mean a whole lot as cryptography/data skullduggery is not one of my areas of expertise). Still, if you don't believe me, install it, use it, and watch the installation magically grow and grow and grow. But as the recently departed black bearded pitch man used to say...
"But WAIT Folks, There's MORE!!!! If you call in the next two hours..." well anytime you hook a Win7 machine to a network in general, and the internet in particular, the computer lights up with activity, even if you're doing absolutely nothing. I'll have to defer to the expert packet sniffers, but I'm here to tell you, there are bits flying all over everywhere and you don't even have to touch the keyboard heheheh Come to think of it, if some enterprising young coder tied a sniffer feeding the unexplained net traffic into Media Player 11, you'd get a more colorful and explosive visualization than by playing Beethoven's 5th.

Maybe I'm just old fashioned, but I don't want my private data floating out there in the cloud, I don't want my every word 'parsed' for advertising, datamining, or political targeting, I don't want to be followed around the net by commissars or geek paparazzi. I want to read my news, I want to follow my intellectual curiosity, and I want to play my games. And I DON"T want to buy a new computer just to run this op system. LOL, makes me want to go back to Win95

All in all, as Sam Goldwyn used to say, "You guys have fun, but 'Include me Out'.

Seer

 

That's hardly evidence in support of Windows 7 spying schemes.

 

"Seer's Law No. 3".......

Bog, I get paid to render advice on a wide range of topics. I have one rule. I give it, the people who pay me either take it, or not. I am NOT a Burger King. You don't get it your way, you get it my way, or you don't get the sob.

In this case, I have offered up a freebie, embrace it, investigate it yourself, learn something, and profit therefrom or ignore it, as you wish. All the same to me.
Just so you know tho, I cut my teeth on fortran and cobol, back when punch cards and magnetic tape were kings..and THE premier cold war coder for the NSA (and a host of other alphabet soup type orgs) was/remains a good friend of mine. In any event, I've said what I logged on to say. Happy broadcasting...

Seer

 

If you're going to say that Windows spies on people, it seems logical to ask for good reasoning and evidence to support that, not some irrelevant information such as the fact that you're paid for advice.

 

Seer's Law also states that we could be in the Matrix and none of this is real.

Conspiracy theories are for the vain and egotistical. They can't be proven to be false for the simple reason why they can't be proven to be true.

It's like proving or disproving there is a parallel universe.

 

@Kevin O'Brien

I am shocked by your battery life numbers. I am using my T400s with W7 on battery with the lowest brightness setting and only getting 4.5 hours.

Was your test unit using W7's default power manager or Lenovo's Power Manager (Power Source Optimized?)?

 

I was using the "nbr" profile. I basically set it to match the balanced profile. CPU is set to adaptive, screen was at 10/15, and all idle timers were disabled.

 

Thanks for this very interesting review, and I appreciate the effort in making it happen...

I'd be curious to see a smiliar comparison on a lower end dual-core machine with integrated graphics, as one of the big complaints about Vista was always the hardware requirements needed to make it actually run well. I'd echo the 64-bit comparison asked for by others as well, although MS has still failed to use a small percentage of their gigantic marketing budget to provide developers with the incentives to write true 64-bit apps that would realize the promises of either Vista or 7. As an MS Partner, I still remember most of the advantages touted in the run-up for Vista being 64-bit, and then seeing almost no 64-bit software actually available. This review and the comments here make me think that one of the usual Windows axioms is going to apply, i.e. wait until SP1 to see much difference.

The other thing that would be interesting, and I think very doable, is to run this same testing and benchmarking set on an existing system with common applications, to see what differences exist before and after a conversion from Vista, as that is the question most people are faced with now...that is...is it worth converting my Vista machine to 7?

Based on what I'm reading here, and elsewhere, other than faster bootup and shut down times and a "feeling" things are running faster, I wonder if it will really make much real difference for most users, and be worth the money to convert.

 

For those not yet convinced that Win7 is just Vista sp(put whatever the next number is here), you might take a look at Anandtech's article testing DX11 on Win7 vs DX11 on Vista...<spoiler...>

same

duh

seer

 

Although your points could be valid seeratlas, please, when discussing, could you be a tad more polite?
It is a public place, and flaming all other people is not necessary.
Can you provide the link/source please? Thanks!

 

I'd really like to see a hardcore performance comparison between these variations as a desktop os when running demanding apps like rendering applications and 3D software, including gaming performance:

XP x64
Windows Server 2003 x64
Windows Server 2008 x64
Windows 7 x64

 

Sorry Serg,

I find as I grow older my tolerance level for marketing bs diminishes geometrically..heheheh.. anyway, here it is.

http://anandtech.com/weblog/showpost.aspx?i=652

Seer

Oh, and that wasn't a flame...trust me if I was to pull out the incinerator, you'd know it

 

There IS at least one addition in Win7 that might be of interest to gamers, particularly those with laptops. Win7 will permit two sets of video drivers running different cards at the same time. Vista will not (but you CAN work around that if you have an amd based system).

Seer

 

Why limit yourself to just one parallel universe? you DO know that current popular theory suggests the existence of multiple parallel universes don't you? or maybe not...) You think that's tough, try getting your head around string theory. Don't feel too bad if you can't grasp the concepts. I'm told it takes a certain type of trained brain.

In the meantime, regarding the data reserved by ms for snooping..
..here's just a snippet....


http://www.theregister.co.uk/2009/11/10/ms_forensics_tool_leak/

and this is just for joe law enforcement...you should see the one for the NSA and associates..course being in that 'parallel universe' u couldn't see it anyway.

sigh, oh, and how's that obama vote working for you?

seer

 

Surfasb's analogy served to point out that conspiracy theories are as pointless as attempting to prove or disprove something that can't be conclusively verified in any way. That present theories suggest there are parallel universes is irrelevant to the analogy. For now it merely serves surfasb's purposes that they are mere hypotheses, and no conclusive evidence to verify.

With that said, it is safe to say that no one fully understands string theory.

And another thing: the Register is practically a joke. Look at the home page, for God's sake. Half of the articles are satire.