7 RTM vs. Vista SP1 vs. XP SP3 - Shootout, Part 1

7 RTM vs. Vista SP1 vs. XP SP3 - Shootout, Part 1

Summary: Over the past few months I've talked a lot about the performance side of things with regard to Windows 7. However, I've not been able to give you any actual numbers because the EULA for the beta and release candidate versions of Windows 7 prohibited me from doing so. However, now that Windows 7 has hit the RTM I can carry out benchmark tests and provide you with real numbers.


Over the past few months I've talked a lot about the performance side of things with regard to Windows 7. However, I've not been able to give you any actual numbers because the EULA for the beta and release candidate versions of Windows 7 prohibited me from doing so. However, now that Windows 7 has hit the RTM I can carry out benchmark tests and provide you with real numbers.

The tests

Rather than dump a whole series of benchmark results on you, I'm going to release them over a few posts. Here I'm going to look at four specific metrics:

  • Boot up speed
  • Shutdown speed
  • Compressing and then extracting a folder contains numerous files and folders (3500 files / 30 folders) totaling 5.15GB

Tests are carried out on 32-bit versions of Windows.

Test systems

I’ve used two desktop systems as the test machines:

  • An AMD Phenom 9700 2.4GHz system fitted with an ATI Radeon 3850 and 4GB of RAM
  • An Intel Pentium Dual Core E2200 2.2GHz fitted with an NVIDIA GeForce 8400 GS and 1GB of RAM

Results/Conclusion -->

The results


My test results here match up closely to the results I got when Windows 7 was both in beta and RC stage - that Windows 7 clearly, and conclusively, outperforms both Vista and XP. No ifs, no buts.

More benchmark results to follow shortly ...

[UPDATE: I've just noticed that CNET UK have published some benchmark results too, and the data for boot up seems at odds with some of my results here. In particular, boot up speeds seem to show Windows 7 as being slower than Vista and XP. I can't comment on the particular results, but my results from testing carried out yesterday seems to be consistant with results I saw with earlier builds. Also, in real world usage, I find that 7 boots up faster than Vista or XP. As I always like to say with benchmarking, your mileage can, and most probably will, vary ... sometimes dramatically!]

<< Home >>

Topics: Software, Microsoft, Operating Systems, Windows

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.


Log in or register to join the discussion
  • 64 Bit

    32 bit tests seem to be somewhat limiting given the advancement in 64 bit tech these days. Any plans for similar comparisons in 64 bit? Better yet, take a more real-world example. Since the majority of XP desktops are running 32 bit, and (hopefully) the majority of Win7 desktops will be 64 bit, how big is the performance delta there? I think it is relevant as a move from 32 bit XP to 64 bit 7 will represent a more real-world scenario and many of us are curious as to how beneficial moving to 64 bit will be.

    Thanks for the benchmarks.
  • Can we please stop with the bootup / shutdown focus?

    It just isn't that important a metric. Sleep / resume times are far more important if you are going to be measuring "starting" and "stopping" times, especially with the rise of laptop marketshare.

    One thing I would be interested in seeing is if the newer OSs can take advantage of newer computer hardware better than old OSs can. Because it could be used to flame Vista, I've seen a lot of focus on "minimum" specs but I haven't seen any tests that focus on "maximum" specs. In other words, take a task that is able to take advantage of multiple cores and GBs of RAM and compare the performance of XP, Vista, and W7 on a 4+ core, 8+GB machine. If W7 can take advantage of multi-cores and all available RAM better than XP can then W7 is the better OS for newer computers.

    Or how about multi-tasking? Can you continue to work on your Excel spreadsheet while it is transcoding something in the background? Just an example of course but if multi-tasking code has improved in W7 when compared to XP and Vista, then again, this is a good thing.
    • agreed

      [i]Can we please stop with the bootup / shutdown focus?[/i]

      I agree, as I rarely shut down my machines... I'm much more interested in application performance. For the once a month or so I do restart my machine, if it took 5 minutes I wouldn't care.
    • agree also.

      The start up and shutdown times are pretty pointless as you spend the majority of your time using applications on an OS. Why not do a comparison of startup times for certain applications, how many clicks to install a new printer, how quick it is to find a document or a configuration setting, how quick and easy to change between applications, etc
      Also the benchmarks on disk and network IO can be an indication of how performant the IO subsytems are but still not a true reflection on which OS is better - I would prefer slightly slower IO if it helped ensure data integrity under adverse conditions as well as allowing me to continue opening and using other application without the system grinding to a halt.
      Why not do some CPU and memory bound tasks on a multi-core processor e.g. encode a video file whilst formatting a large word document, whilst compressing numerous files, etc
      Finally do a combination of all the tests at the same time to really stress the OS.
      • RTP

        The rest will come over dayas as he knows that each and every test will generate lots of feedback.
    • Startup and shutdown...

      ...aren't these metrics a measure of the efficiency of the OS?

      The quicker you startup, the more efficient the OS is at loading drivers and startup programs etc. Similarly, the quicker the shutdown, the better the task management and memory release mechanisms built into the OS.
      • Not really.

        You can't really compare the operating systems unless they are loading the same drivers, services, startup programs etc. Also different operating systems do different checks on startup to make sure everything is functioning correctly - who cares if booting up takes an extra 20 seconds if you know you will have a stable OS to work with.
        • So, are you implying that XP is more stable because it takes longer?

          Hasn't been my experience.

          We are talking about basic setups here. If you want to start trying to level the playing field, there are a lot of tweaks that can be done to each to make them better, but very few people will go to that effort.

          I suppose the basic assumption is that an OS install will load up its defaults and most will run it as is.

          But, given they are installed on the SAME hardware, drivers appropriate to the OS will be installed.

          Each OS has certain services that are Automatic, Manual or Disabled. Some may change from one version to the next as lessons learnt form user feedback. However, I would not expect them to be great. The great majority of services have the same settings.

          It took a lot of tweaks to get XP running smooth, but Vista and now Win 7 seem to not need as many tweaks to just work. Audio recording (DAW) use on XP requires a lot of tweaks to be stable and avoid stutters, but a lot of those same services that have to be disabled in XP do not seem to create the same problems in Vista or Win 7.
          • I never said that!

            I said that you can't compare boot times between operating systems as they do different things on startup. I did say that I don't care if an OS takes longer to boot if it makes it more stable due to more checks. I completely agree that XP is not more stable - I hate using XP after having used Vista and now W7.
          • Sorry, I inferred it because XP IS the one that is slower

            I apologise if I misread your intent, but I was asking the question to clarify it.
          • Sorry also

            I can see how you might have inferred that. No apology needed.
    • Depends how fast they become

      I find Win 7, combined with SSDs, makes short work of complete power down and startup, so I have totaly disabled sleep and hibernation.

      As for the complex situations of which you write, I would suspect some of the standard app suite benchmarks would give some indications, but it depends upon the particular apps.

      However, some scenarios are so complex that the tests have to be very carefully designed so that they isolate the OS enough to draw valid conclusions about it.
    • You seem to not work in a corporate IT environment.

      Boot up and shutdown times can be exasperating when network resources "time out" overnight and you have to reboot every morning.
      IT departments load up so much (what is a polite word for cr_p) that it bogs down an engineering workstation with fast dual processors and 3 gigs of ram.
      If I could immediately be productive while Windows does it's primping and housekeeping it wouldn't be an issue, but having to wait 3 to 5 minutes is REDIKULOUS!
      It is like waiting for your wife while she gets ready.
  • Vista vs 7....

    The performance enhancement from Windows Vista SP1 (though most are probably using SP2) and Windows 7 is negligable at best. I dont see how that warrents going out and upgrading JUST the OS.
    • Because not everyone uses an OS just for speed

      If that were the case everyone would be using DOS, since it has no overhead and thus little performance hits.

      What's remarkable about 7 is not that it's faster than Vista but faster or equal to Windows XP, with all kinds of overhead like file indexing, background defraging, super fetch auto caching, etc.
      • Superfecth is one of the reasons it is faster (nt)

      • Even without those HDD compensating devices, W7 is still faster

      • The real story is Vista is faster than XP.

        That's in stark contrast to all the anti-Vista folks.
  • RE: 7 RTM vs. Vista SP1 vs. XP SP3 - Shootout, Part 1

    Why not use Vista SP2, that's shipping isn't it? In looking at these benchmarks, the only one where a user would notice the difference between Vista SP1 and Win 7 RTM is file compression. What this really highlights is, for these tests, how much faster Vista is than XP.
  • Is Shutdown\Bootup Times Really All That Important

    Given the minor differences in these results in modern systems, why benchmark these specs, I mean its just semantics at best, I remember you specifically mentioning in a post Adrian that boot-up times don't matter to u as you rather a fresh boot-up vs wake from sleep so u just prepare for the wait and have a cup of coffee. And I agree, I use my system sometimes 10hrs per day, whether my system takes 1 or 3 minutes to boot isn't gonna be a cry factor for me. I mean if it takes 10 minutes, then yeah, but I mean the few seconds or 1 minute difference is minuscule. So is shutdown time, specially if its a desktop, you turn off your monitor and leave, I mean really, how does a system with fast shutdown time help productivity. As Zealot said, what we want to know is solid performance, and as he said not just with low end stuff, I mean, I hear people crying cuz they want 1998-2000 hardware to run Vista/7 perfectly, that's like wanting your 1953 Cadillac El Dorado to run great on Ethane fuel or even Hydrogen. Plus quite frankly this is all business, so optimizing a modern OS to run old hardware is a waste of time and not very profitable.

    But in recapping, please, the shutdown/bootup benchmarks a quite useless for people who actually use their PCs for real work.