XP Service Pack Slows Programs

AEton writes "Vnunet and others are reporting that Windows XP's Service Pack 1 has introduced a flaw into the operating system. Changes to memory handling code result in programs which often allocate memory (which is many of them) can take up to ten times longer than normal to start. Microsoft has acknowledged the problem in Q815411, and while a patch is available by request from Microsoft Product Services, it will not be widely released until Service Pack 2."

Comments

w2k is effected as well

[Billly Gates]

Both Openoffice and Mozilla have slowed down quite alot with the latest security updates from Microsoft.

When I open openoffice is just sits there doing nothing for like 20 seconds and then launched. No excess cpu overhead or anything. It just stalls and then runs. Its just annoying and I wonder if its a conspiracy theory.

Has anyone else noticed this?

So that's why Mozilla's been slow...

[Anonymous Coward]

Once it's using 250+ megs of memory after a week of continuous use, it becomes a real beast. Apparently, I'm not supposed to use Mozilla for a week at a time without closing it though...

More /. FUD.

[tshak]

My first experience with XP (no pun intended) was just a few weeks ago with my new laptop. Everything about it is faster than my workstation. True, my workstation has a slower CPU, but it also has a significantly faster harddrive, which is usually the bottleneck for loadtimes. XP boots way faster, loads programs noticeably faster, and has a more responsive GUI even considering the more resource intensive graphics. So, if in fact there is a bug in XP that slows the loading of some programs down, then that just means that SP2 will make things even faster.

Oh, and let's not talk about load times for X window managers. Even with XP's "bug" KDE nor Gnome stand a chance*.

* Disclaimer: this is based on my experience RH8. I'm currently downloading the Mandrake 9.1 ISO's (slowly... mirrors are hammered) so my observation is not taking into account potential performance improvements made recently.

I'm such a whore

[alexburke]

Why bother asking Microsoft for the patch? Here you go. :)

Re:Well, I am not installing XP SP1 anyway...

[mabinogi]

It's ok, it just refuses to install SP1, it doesn't invalidate your product key.

XP SP1 Slowdown

[aSiTiC]

I don't know about anyone else but I noticed a fairly noticable slowdown in XP when I installed SP1. I've since uninstalled SP1.

Personally I prefer speed over security.

Re:The fix will cost you

[mentin]

It is Oracle which uses the "Captive Audience" pricing plan - their user do have sign for service agreement (and pay for it) to download service packs and even security fixes.

Microsoft at least releases the fixes free.

Has anyone actually witness this slow down?

[antdude]

I haven't on my machines and test machines at home and office with various softwares. Are there any known programs or games that does show this?

Release candidates

[Tomster]

It is plain that Microsoft's internal testing is insufficient. I don't really fault them for this -- it's simply impossible to have enough configurations, testcases, and procedures to cover more than a small percentage of the actual ways the product is used.

IMO, Microsoft would benefit by issuing public release candidates for new OS versions and patches. It would greatly reduce the impact of problems with patches and new releases.

-Thomas

As a programmer

[SynKKnyS]

As a programmer, I have yet to run into this bug. I am using Windows XP SP1 with all patches from Microsoft Baseline Security Analyzer applied. My game engine contains a few routines that rapidly allocate and deallocate memory for linked lists. There might be more than 400 allocation and deallocations every second from the time the engine starts. I have used GlobalAlloc (or the synonym LocalAlloc), HeapAlloc, VirtualAlloc, and CoTaskMemAlloc. Not one has proven to be slow at all, however, I settled on HeapAlloc since it seemed to be the most flexible. But, this may change if my plug-in system needs direct access to allocated memory (which only CoTaskMemAlloc provides). Does anyone know how to reproduce the bug in code?

Re:As a programmer

[NeuroManson]

It may be (a) Microsoft application specific, or (b) independant application specific, IMO. There are often hundreds of applications, be they shareware, freeware, or third party, that cause this issue, but lacking an exact basis for comparison between systems running each and every one of them, I doubt there is an easy answer, let alone a way to conveniently reproduce the bug.

Until more detail is offered as to the applications involved, there is no way to reproduce the bug, unless one develops psychic abilities, and can read the minds of every person using XP.

Re:More mirrors for Patch

[Mattsson]

Eum... Is this patch availible somewhere for other languages than English?
I would need a Swedish version. =)

(I can't belive that MS hasn't stopped using different binaries for different languages yet. It's idiotic. They even *rename* systemfiles and folders between languages, making some poorly made programs incompatible with other languages than English...)

Re:VM issue?

[stratjakt]

It is very obscure.

If it was the major flaw the slashbots would want you to think it is, you wouldnt have heard about it first on slashdot. (Did you hear about Code Red, Nimda, Slammer, etc on slashdot first?)

As an aside, here's another nifty coding flaw I found recently:

Slashdot requires you to wait 2 minutes between each successful posting of a comment to allow everyone a fair chance at posting a comment.

It's been -282 seconds since you last successfully posted a comment

I guess I posted from the future! (cue scary music)

Jeez, learn to read

[fm6]

It's a cute joke, but...

This problem may occur if the programs you run frequently allocate and deallocate large blocks of memory. Changes in Windows XP SP1 in the memory management system have caused this operation to take significantly longer than with pre-SP1 Windows XP.

It's "frequently allocate and deallocate large blocks of" not just "allocate". This is not a small nit to pick, especially for Linux people. For years, GNU libc had a memory leak bug that was triggered by frequent allococation and deallocation of small blocks of memory. It only became an issue back in 2000, when Borland ported their component libraries to Linux.

No problems..

[ball-lightning]

I haven't seen any slowdown personally...
I run:
Internet Explorer
Photoshop 6.0
Dreamweaver (yes, I am that lazy)
SimpleMu
Winamp
Windows Media Player
Office
Slowdown really bothers me too, so I'd expect if it was really noticable, I would have noticed. I suspect this may be more of a "benchmark thing" than anything else.

But...

[Lolaine]

Doesn`t the whole OS slow programs? I mean, since whe are running an OS, we know it will slow down our apps. Is the counterpart to run only an app, no network, and so on, but Why cant be developed critical apps for the PC (or other cheap platforms) with no OS? (I mean, apps that boot themselves) It would make OS crashes less critical for that critical apps and make response speed (since no scheduler, context switches, etc...) increased. It would be a pain to work in one of those apps from scratch, but they would not rely in nothing but hardware. I remember playing a game for the PC (80286 era) that selfbooted from a 1.44 Mb Disk....

It really makes a difference

[forged]

I have Apache 2.0 and MySQL 4.0 configured as services. Then upon logon, Mozilla 1.3 loads itself in memory. I often had to wait up to one full minute after system boot and user logon, for the networking subsystem to get started !!! This meant no Internet access until one minute after logon - not acceptable. With the hotfix applied, boot is faster (as in pre-SP1) and there was a sizeable difference in programs load time.

Re:Solution to your problem

[AsnFkr]

You work in a repair shop and you don't know how to get to Safe Mode?

You are kidding right? Of course I know how to get into safe mode. If you'd READ my post you would understand that the machine will not *allow* safe mode to be entered when this problem occurs. I hope this has restored your faith in the human race, as your comment has diminished mine. :)