Vista Users Still Won't Trust Sleep

Apr 10, 2007 · 3 min read  ·
Share on:

Last night, my computer fell asleep and I decided not to wake it. Experience taught me that Hibernate was much more reliable but I decided, eh, it's already Sleeping and I should be too.

This morning, after a public transit meltdown in San Francisco, I made it into work a mere 3 minutes before the managing partner was scheduled to call me for an important meeting. Still in a sweat from running, I arrived at my desk and attempted to wake my laptop from its sleep. I'm in IT after all, and I'll need it for this meeting. Well, my laptop didn't want to wake up. It sat there for a bit doing something.. not sure what, but it was powered on and the screen was blank. So I started poking around; I tried to eject the CD.. that's helped in the past to kick it into gear but it wasn't working this time. After about 30 seconds of waiting, I decided to it was time for a cold boot.

So now, I'm spending my last two minutes before 9:00 AM hoping that the managing partner won't decide to call a few minutes early. Microsoft, I'm sorry but resuming from sleep still doesn't work. And when it does, it tends to sit there for awhile, making me wonder what it will decide to do.

A few days before installing Vista, I came upon a PowerPoint created by one of the Microsoft Program Managers for the WinHEC (hardware) conference. In the presentation titled "Power Management in Windows Vista", Pat Stemen wrote the following::

Reliable Sleep Transitions

  • Windows Vista promotes the use of sleep as the default off state

    • Requires reliable, fast and deterministic sleep transitions

  • Failed transitions were the primary sleep adoption blocker in previous versions of Windows

    • Lead to great user frustration and distrust of power management
    • Investigations inlicate component vetoes are the primary cause
      • Appcication, service or driver willingly prevents the sleep transition

  • Sleep transitions will succeed

    • Vista will not query user mode components when entering sleep
    • Drivers may not veto sleep transitions
    • User-mode notification (PBT_APMSUSPEND) will continue to be sent
      • Timeout to process event has been reduced from 20 seconds to 2 seconds

  • Applications, services, and drivers must be prepared for this change

    • Proper design and test is imperative

I don't know where the development in Vista's Sleep went wrong but, it's still not reliable for me. Same goes for my friend Zach..he's the one that actually clued me into using Hibernate instead of Sleep. Even with writing up to 2gb of data to the hard drive, going into Hibernation is just as fast as sleep and coming back is much more dependable. Now there's a 9 in 10 chance that resuming will work instead of a 5 in 10 chance (if that.) 99.999% reliability would be best of course but I'll take what I can get for now. I really hope Sleep's issues are finally resolved in Vista SP1.

Update: They finally released this patch publicly (without you having to call/email). It's helped me a great deal. Note that VPN is also considered "PPP" so this patch isn't only for dial-up users. Also, visitor "wchp" said that this patch helped him with sleep issues. So far, about 250 people have found my blog while searching for a fix. Ya'll holler and let me know if you have an nVidia card too.