Sponsored By Aspose - File Format APIs for .NET

Aspose are the market leader of .NET APIs for file business formats – natively work with DOCX, XLSX, PPT, PDF, MSG, MPP, images formats and many more!

VPC 2007 Beta: Centrino 915 Chipset Issues (Updated)

As you'll recall from a previous post, I've moved (lock, stock and barrel) to Windows Vista using VPC 2007 for development purposes. This has brought to light an issue that I never experienced in Windows XP using VPC 2004. It seems that certain Intel chipsets (such as the Centrino 915 used in my Dell D810 notebook) use a power management technique (SpeedStep) which adversely affects the mouse, keyboard and other things in the "guest" operating system. To date, Microsoft has issued a hotfix which (according to some) only partially addresses this issue in VPC 2004 but so far, I haven't been able to find an answer to this issue in VPC 2007 running on Windows Vista. A work-around exists as part of this hot fix where adding a certain <enable_idle_threads> key to the options.xml file used by VPC 2004 and VPC 2007 "should" alleviate the issue. It does, by maxing out the CPU at 100%, which creates a whole range of other issues for the "host" operating system. This is not what I'd call a very "robust" solution, and quite frankly, I'm very surprised that this is the best Microsoft can offer after knowing about the issue for over eighteen months. I realize that this is a "hardware" related issue but for VPC 2007 (and VPC 2004 for that matter) to be successful, it must be compatible with widely used and very popular Intel chipsets.

As most of you know, I'm a very (VERY) strong supporter of Microsoft technologies and Microsoft in general, so this isn't just another "bash MSFT" post. I consider this bug to be a "blocking issue" to the quick adoption of Windows Vista and VPC 2007 by serious developers.

If anyone reading this post knows of a solution to this issue, PLEASE post a comment or ping me via my blog. I'll update this post as I (hopefully) get additional information about a solution to this issue.


I've spoken with several folks from MSFT in the newsgroups and via email and this is the best work-around right now.

Add the <enable_idle_threads> element to the Options.xml file found in the Users > username > AppData > Roaming > Microsoft > Virtual PC folder in Vista. See below for an example. Please keep in mind that this will cause your host (Vista) to run at 100% cpu and host application performance may suffer. I really dislike this work-around but it does work.

  <enable_idle_thread type="boolean">true</enable_idle_thread>

I hope this helps till RTM.

[tags: Virtual PC, Vista]

This entry was posted in Uncategorized. Bookmark the permalink. Follow any comments here with the RSS feed for this post.

11 Responses to VPC 2007 Beta: Centrino 915 Chipset Issues (Updated)

  1. Kevin says:

    Worked for me very well. Thanks for solving a large issue for me right now.

  2. Pascal says:

    Same problem on a Dell Laptop 2.0Ghz Core 2 Duo, 1Gb RAM
    The problem appears when the CPU is slowing down.
    The only workaround I have found is to write a program (Delphi in my case) which makes the CPU
    run at full speed (one of the core is always at 100%) !
    It works fine.
    Hope it can help.

  3. William says:

    Has anything been done to resolve this problem with VPC 2007? I just tried installing/using it on my new Dell laptop (2.0Ghz Core 2 Duo, 2Gb RAM) and the mouse movement on the XP virtual machine is ridiculously slow and erratic. I’ve scoured the Internet for the last day, but I can’t find any permanent solution.

  4. adancs says:

    Yes, the work-around works! Many thanks for your update. I’ve been investigating this issue for a long time, and finally I surrendered. I have just discover that a new version of MS-VirtualPC was released, and I was trying it… with the same bad results.
    I dislike that my CPU is always at 100%, but it works, now I can use my virtual machines.
    Many thanks, and sorry for my bad english.

  5. Petr says:

    FYI all these SpeedStep problems were introduced in VPC 2004 SP1. First release of VPC 2004 (without SP1) never had such problem. I found a posting in newsgroups from someone from Microsoft explaining the issue. There were some changes related to multi-cpu machines introducing this problem as a side-effect. There is supposed to be a test for multi-cpu or so, if false it should switch to “old timing” as before SP1. Unfortunately it seems as the test fails. I don’t understand why Microsoft does not allow to set it manually instead of stupid fixes like “disabling” the SpeedStep by idle thread consuming 100% of CPU time.

  6. jlynch says:


    Things aren’t running well in VPC2007 but it’s usable till RTM. In the Dell D810 bios, disabling SpeedStep runs the cpu at the lowest rated speed instead of the highest. Pretty strange!


  7. camera says:

    Glad to hear things are running well. I’m surprised to hear that disabling SpeedStep leaves you in the low-end MHz because the opposite should happen — it should run at the high-end (max) MHz when disabled.

  8. jlynch says:

    See the update in my post for a work-around.


  9. jlynch says:


    Thanks for the suggestion and link. In WinXP I used SpeedSwitchXP to make sure my notebook was running at it’s fully rated speed and never experienced the issues with VPC 2004. Unfortunately, this program won’t run in Vista and disabling SpeedStep in my Dell D810 bios slows the cpu to 800 MHz. I’ve also set the power management scheme in Vista to “High Performance” but the issues persist.


  10. camera says:

    Sorry to hear you’re having problems. SpeedStep has been around since the Pentium III. You can disable it in the BIOS I believe (I don’t have my D600 in front of me) which may get you going again. There’s a good FAQ at http://www.bay-wolf.com/speedstep.htm

Leave a Reply