Slashdot Mirror

← Back to Stories (view on slashdot.org)

XP SP2 Can Slow Down Business Apps

Posted by timothy on from the tradeoffs-are-everywhere dept.

An anonymous reader submits "Mobile PC magazine installed XP SP2 on a bunch of notebooks and benchmarked them, finding that SP2 caused a 9-percent performance reduction in business productivity apps. While a couple of notebooks performed better, the majority took a 3- to 22-percent performance hit." For now, the story is just at the top of the Mobile PC website, but they promise more details in an upcoming issue.

26 of 359 comments (clear)

  1. Buffer checks by JanusFury · · Score: 5, Interesting

    This is probably due to them recompiling a large number of libraries and system components with the buffer checking and other security features they added into the recent versions of Visual C++. If you ask me, it's worth it, just to know that my Windows box has a few less wide open holes to be exploited.

    It definitely has proven its worth so far - I may be wrong, but I'm pretty sure the reason SP2 isn't vulnerable to that GDI+ JPEG exploit is that they recompiled GDI+ with buffer checks.

    --
    using namespace slashdot;
    troll::post();
    1. Re:Buffer checks by metlin · · Score: 4, Interesting

      The reason I brought that up was because I was interview by Microsoft last summer at Seattle, and one of the groups that interviewed me was the systems group.

      (Funnily, systems wasn't even my area, but still they interviewed me, but that's another story...)

      They were of the opinion that since MS is a favourite target of hackers and the like, any MS programmer ought to go to extraordinary lengths to patch any and all buffer checks and foo bar. I was asked to write some piece of code for compiler design and memory management, and the guy kept harping on buffer checks.

      I would imagine that with ALL those checks, such things would not be common - but lo! and behold, there they are.

      Either they are not doing a good job of doing the whole buffer check thing that the guy harped to me about and it was all hogwash to impress upon you how "important" and "hard" coding in MS is, or there is something seriously wrong with the codebase that SO many exploits turn out everyday.

      I can only guess which one it is.

    2. Re:Buffer checks by Anonymous Coward · · Score: 1, Interesting

      So, isn't Linux "commercial grade"? How many Linux distros are compiled with these checks?

    3. Re:Buffer checks by IronChef · · Score: 4, Interesting

      I've heard so much about the programming practices at Microsoft and what not - and yet, ironically, these things keep cropping up so damn bloody often while some operating systems coded by a bunch of loosely connected hackers are way more robust and stable.
      Hmm, makes one wonder.

      the openbsd people are united by an ideology. Microsoft employees are largely, though not exclusively, united simply by the desire for a paycheck.

      I work in a Microsoft facility and let me tell ya, they aren't all smoking what Steve Ballmer is.

      Is it any wonder that quality suffers when compared to a project that is a labor of love?

      Or maybe my bad attitude is why I am a contractor and not full time there. :)

    4. Re:Buffer checks by AdamInParadise · · Score: 2, Interesting

      Hey, you! Wake up! This is the 21st century now, not 1988! Those kind of protections were build into the NT kernel since the very beggining (1993 or something), which means win 2k and xp have them.

      Regards,

      --
      Nobox: Only simple products.
    5. Re:Buffer checks by omicronish · · Score: 5, Interesting

      Either they are not doing a good job of doing the whole buffer check thing that the guy harped to me about and it was all hogwash to impress upon you how "important" and "hard" coding in MS is, or there is something seriously wrong with the codebase that SO many exploits turn out everyday.

      I was an intern at Microsoft this past summer, and I believe it's the shear quantity and perhaps complexity of software being written that's resulting in these bugs. They really do emphasize writing secure code now (I don't know how it was like before). I shared an office with two other interns, and during several code reviews another intern was involved with, there would be "did you check parameters here? potential buffer overflow? what if this is NULL?" And it wasn't even important code he was working on.

    6. Re:Buffer checks by Anonymous Coward · · Score: 1, Interesting


      First off, you seem ignorant of the point that many people did know better than Ritchie, which is why OpenVMS and OS/400 have infinitely better security records than UNIX does.

      This is a myth, or at least a very disingenous claim. OS/400 and VMS are completely different animals than UNIX meant to run in different environments.

      Historically, the vulnerabilities of these systems have not been subjected to the public Internet. Furthermore, only large businesses typically used these systems, and the security industry built around them is quite secretive. This is not necessarily just on purpose, but that specialized knowledge in many areas seems to be lacking on the Internet anyway. Anybody that thinks that the Internet is the be all end all of what information is out there is a complete idiot. Yeah, so OS/400 is not reported on Slashdot.

      And yes, I am completely familiar with the fundamental differences in OS/400 that make it nearly invulnerable to typical UNIX/Windows things like stack based buffer overflows. This does not mean it is the impenetrable fortress you seem to think it is. All software has bugs.

      Furthermore, the better security of OS/400 is in no small part helped by the hardware architecture. It is part of a complete integrated system. This is much different than the mini/micro world that UNIX and friends grew out of.

    7. Re:Buffer checks by Anonymous Coward · · Score: 2, Interesting

      Uh what are you talking about?

      Windows XP has users, groups and privileges, and not every program has the run of the system.

      Uh, what are you talking about?

      As a Windows NT programmer for 10 years I, and many others are fully aware that the Windows security model is more comprehensive and flexible than the POSIX model. This is due to the power of NTFS and the uniform interface of the NT object manager.

      However, while this is all dandy, the Windows environment is architected, and third party apps are delivered, that make this entire security system virtually useless. Furthermore, Windows has an aging capability based (policy) system that is now being bested by more modern and truly effective alternatives. It is very difficult to actually maintain Windows NT systems in server or desktop configurations that make use of sane file system and execution policy. To many services need extended system privileges. 99% of NTFS filesystems have uniform perms throughout.

      In short, the great security stuff in Windows is wasted on an overall lame implementation of the system.


      And UNIX is just as vulnerable to buffer overflows as Windows XP. They both are programmed in languages that are prone to such problems.

      It is possible to program in both these OSes with languages other than C. I mention this not only to be an asshole, but also because most security vulnerabilities are found in things other than the kernel.

    8. Re:Buffer checks by Anonymous Coward · · Score: 1, Interesting

      >usability is what made Microsoft a success.

      Yes, Microsoft's world-famous usability. They're the industry standard for usable software! Uh huh! Users loved DOS and Windows 3.1 so much that they threw out their less expensive, harder to use Macintoshes and flocked to the more expensive but easier to use DOS machines. Why use a GUI when you can just get a keyboard template to go over your function keys?

      Seriously. Where in the hell did you get that fucked up idea?

    9. Re:Buffer checks by sg_oneill · · Score: 2, Interesting

      actually the better record has to do with the fact fk all people use OS/400 / OpenVMS.

      Yeah unix had some silly bugs, but that partly cos it was written by a really small team in spare time and became uber-popular despite it never really being intended to , and in an age where hackers where guys who logged in and FIXED your shit.

      --
      Excuse the Unicode crap in my posts. That's an apostrophe, and slashdot is busted.
  2. This has to do with... by Anonymous Coward · · Score: 2, Interesting

    This has to do with a buggy CPU "driver" in SP2, rolling back that driver to the pre-SP1 version should correct the slowdown.

  3. Intentional bloat? by Anonymous Coward · · Score: 1, Interesting

    the more 'deflation' in performance for computer hardare, the more people think their computer is 'too slow', and that they need a 'new fast one' (which happens to be sold with a Windows licence).

    MS has done is unintentionally or otherwise since Win95 - each new revision is dramatically more bloated, encouraging new computer purchases. If WinXP was as light weight as Win98, there're be far fewer new computer sales. Since WinXP is having a long shelf-life, it'd make sense that a service pack would continue this trend.

  4. Why- by thewldisntenuff · · Score: 4, Interesting

    was this even posted at all?

    This wasn't even a readable story - just a small synopsis of a story that will be featured in Mobile PC mag next month. There could have been plenty more info, but instead we got two paragraphs.....

    OTOH, is an average 9% drop in performance even an issue? I mean, 9% in office apps is nothing....Who needs high performance when typing, making spreadsheets, or even a PowerPoint presentation?

    This (once again) illustrates the MS push towards security over performance/compatibility

    -thewldisntenuff

    --
    My MythTV HowTo
    1. Re:Why- by metlin · · Score: 2, Interesting

      Although I agree with most of you said, I have a bone to pick with this statement -

      OTOH, is an average 9% drop in performance even an issue? I mean, 9% in office apps is nothing....Who needs high performance when typing, making spreadsheets, or even a PowerPoint presentation?

      Hmmm, I guess you've never been in a corporate business office, where excel sheets running into hundreds of pages are opened. Or business plans and product specs that run into hundreds of pages are opened.

      Why go that far, you've apparently not tried writing a paper in MS Word - I've been trying to get a couple of papers done for a conference deadline coming up on Monday - and my system is so terribly slow that it's unbelievable.

      Sure, you can compromise on speed - but why? Why should I, when I don't need to? Had MS done this right from the beginning, the performance hit could have atleast reduced (to say 5%). That is still a significant gain when compared to 10%, which in my opinion is QUITE slow. If you had a latency of 10% in all your networks, you would know what am talking about.

    2. Re:Why- by mrowlands · · Score: 2, Interesting

      because "insert your favourite game here" runs 9% slower now

  5. Re:Of course. by savagedome · · Score: 4, Interesting

    my internet lagged so bad I had to reinstall Windows Xp. Worked better after that.

    Bollocks. Reinstall XP? Did you atleast try removing SP2 to begin with? You could atleast set a system restore point before you do any major upgrade that contains patches and/or including third party drivers.

    I am no Windows fan but just trying to make it sensational that you had to reinstall XP from scratch doesn't really do anything. I have installed SP2 pretty much after it was released and have had no problems. (Well, of course some people are going to see glitches considering the size of that damn thing).

    And maybe, just maybe, did you think of the possibility that your *P2P* app might be the bugger. Just a thought.

    --


    Free XBox, PS2
  6. Putty and SP2 (It's a business app, really! ;) ) by realdpk · · Score: 2, Interesting

    Has anyone noticed an increase in how long it takes Putty to start up post-SP2? I thought it was the firewall at first, but I disabled that. It still takes about 5 seconds to launch, where before it was instant.

  7. Welcome to Windows upgrades by coupland · · Score: 4, Interesting

    If you thought SP2 would be a speed upgrade then you also buy the previous lines that Win98, ME, NT4, W2K, XP would make Windows faster than previous versions. Of course these fallacies are based on the assumption that you would install the upgrade on a *newer* PC than their sample set. No Windows update has ever been faster than its predecessors.

    Period.

  8. Reality check by Card · · Score: 3, Interesting

    Correct me if I'm wrong, but given today's hardware, is 10-20% slowdown even noticeable to the average user running, say, Word? IIRC, the threshold for user to notice anything meaningful is around 30% in day-to-day operations.

    Games are a different beast, but does the user even care if loading a spreadsheet takes an extra second or two?

  9. hmm only 9% performance hit? by pavera · · Score: 2, Interesting

    I've attempted to install SP2 on three machines now and I'm not trying any more. After the 1st install, the system blue screened, and could not be recovered, had to reinstall from scratch.

    The second attempted install got about 2/3rds of the way done and then crashed resulting in an unstable system. The partial install could not be completely removed, and the machine would crash often, another reinstall from scratch.

    the third attempted install died in the early stages repeatedly (about 15 seconds after starting the install) and never got past that point.

    These were three completely different systems with different software installed, but all ended up with the same result, no SP2 without a complete clean installation of XP first. I'm so disgusted with MSs QA right now, I never plan to install SP2 again, because my time is too valuable to spend entire days rebuilding systems just because they can't write updates to their software.

    Hell in Gentoo and Debian I update the entire system with a single command and download hundreds of software packages equalling hundreds of MBs and it all goes smooth as silk, can't MS figure out how to copy files from an update package into the system without blowing it all to hell?

  10. Defragment C:! by prandal · · Score: 3, Interesting

    After installing SP2, defragment your hard drive - so many core files are replaced that the system's performance will be even more sub-optimal than usual until you do this.

  11. XP SP2 can give you a serious headache by Linker3000 · · Score: 3, Interesting

    Since installing SP2 on a laptop, the printouts from Treeview Pro (a directory listing program) have had every printed character flipped on its vertical axis - all the letters are in the right place but the wrong way round so - for example, all 'b's look like 'd' - it's readable but makes your brain hurt!!

    Does anyone have a weirder SP2 effect?

    --
    AT&ROFLMAO
  12. Re:It was to be expected... by Tim+C · · Score: 2, Interesting

    Oh, bull.

    with their exploit ridden code

    I've been running various versions of Windows for 7 years now, and have not been exploited once. Sure, there are exploits in there, but they only catch the incautious and foolish. Avoiding them is *easy*, with a little computer literacy.

    poor programming decisions

    I can't comment on that, and nor can you, unless you've been privvy to some of those decisions. I'll agree that some of the *design* decisions are curious to say the least, but that may well have been marketing-driven.

    and heavy handed activation tactics

    Install XP Pro. Activate - two, maybe three clicks. A year later, after futzing around with my hardware, I need to reactivate. Two, maybe three clicks later, it's reactivated. Total impact to me: 30 seconds? "Heavy handed"? Only if you're constantly swapping hardware around, or trying to use a pirated copy.

    inconsistent UI

    The *Windows* UI is consistent. *Office*, on the other hand...

    predatory business practices

    I'll give you that one, but temper it by pointing out that most (large) companies are as predatory as they think they can get away with being. At least MS isn't purposely draining pension funds, or flouting environmental laws, etc.

    FUD-filled marketing

    I'll give you that one too, but again, there's an awful lot of anti-MS FUD being pushed by sites such as this one. Two wrongs do not make a right.

    the push to DRM lock in

    I see that as accepting the inevitable. The *IAs are pushing hard for legally-mandated technological restrictions, and rather than wasting time and money on a fight MS might well lose, they're just doing what so many /.ers wish teh *IAs would do - working to embrace a technology-driven change. I dare say that lots of large meetings full of MS big-wigs have taken place, in which they've decided that it's best *for MS* to create the technology, rather than try to fight it. Sure, it'd be nice if they'd fight for our rights, but that's not what corporations are for, unfortunately.

    every user of Windows pays. Through the nose.

    Well, that's your opinion, but I can't say I see things the same way. All I see is a stable, largely hassle-free OS that gets out of my way and lets me get on with using my PC. Even after 5 years of using Linux, including 2 of using it as my primary OS at work, I couldn't say the same about that. OS X may well be the Second Coming of the perfect OS, but it doesn't run on my hardware.

    --
    It's official. Most of you are morons.
  13. Buggy uxtheme.dll causes a GDI handle leak! by fakeplastic · · Score: 2, Interesting

    XP2 also ships with a buggy xptheme.dll that causes any MFC application that creates windows with a caption (such as MDI child windows for example) to leak 6 GDI handles (HRGNs) every time a window is created. This bug is causing me some serious problems with one of my MFC apps that uses lots of CWnd-derived windows.

    To make matters worse, MS actually fixed this bug with SP1, but have gone and broken it again with SP2! After all that testing they did? Unbelievable.

    And, of course, I have no idea how I can officially report this bug to MS, let alone get them to fix it. There is a KB article for the original XP problem at http://support.microsoft.com/?kbid=319740.

    There is a workaround - use Classic mode instead of Theme mode - fine for me, but try telling my users they cannot run in Fisher Price mode!

    What a royal PITA. My users are screaming for a fix (as my app runs all day long, and as many users use Standby mode, it can stay loaded until the next reboot, leaking bloody handles at a frightening rate!).

    Anyone here have any experience reporting problems like this and getting them actioned? I don't know where to start (I have posted the problem to the MFC group).

  14. Defrag? by freakmaster · · Score: 2, Interesting

    just wondering. these service packs replace more or less every file in the operating system. the files/registry might have been fragmented somewhat, or perhaps the newer files weren't all grouped together nicely for quick access. Laptop disks are usually slower than desktops, i believe they only started shipping 7200 RPM disks in the high end laptops recently (and we all know how much difference 7200 rpm drives made when we started getting them in desktops 4-5 years ago). just a theory, but I would have fully defragged sp1 (including registry hives, pagefile etc...), benched, installed sp2, defragged again and benched. probably wouldn't account for all the slowness, i'm sure bufferchecking and sending personal information to mothership take up plenty of resources as well.

  15. snprintf is not ANSI C89 by tepples · · Score: 2, Interesting

    You wouldn't use gets to input a string, would you? I hope

    There's a difference between fgets, a safe alternative to gets that is in ANSI C89, and snprintf, a safe alternative to sprintf that is in ISO C99 but is not in ANSI C89. Not all useful platforms have a conforming C99 compiler. So how can one do the equivalent of snprintf portably?