Hacker News Reader

Respectfully, you have no idea how Windows 8's fast startup mode works. You also don't seem to know that Fast Startup is the default (so when you hit "shutdown" it actually hibernates most of the system's state).

Essentially in Windows 8 (out of the box) if you hit "Shutdown" then it logs the user out and then "hibernates" the kernel, the registry, services, drivers, and all other "system" level processes.

The computer is then ordered to power off (just like hibernate). When you turn the machine back on, a notice appears on the screen which reads "Resuming from Hibernation" and it loads the Hibernation file into active memory.

Fast Startup is essentially Hibernation without a logged in user.

Windows 8's boot times with Fast Startup enabled is very competitive to Windows 7's boot times using Hibernation (instead of "Shutdown"). It is also very comparable too as Windows 8 doesn't truly shutdown in the traditional sense.

If you boot Windows 8 without Fast Startup enabled then the boot times are very similar to Windows 7's boot times.

As I said, Windows 8 lies to you. It isn't faster, it is just pushing Hibernation in a new deceptive way in order to appear faster.

"This is provenly false"

Citation needed.

OP claims that a Windows 8 uses a feature which without exception causes 1. extra IO workload, 2. reduced performance and which 3. persists entire sessions to 1. reduce IO workload, 2. increase performance while at the same time 3. session is not persisted.

I point out how this argument makes no sense.

I don't think I'm the one you should target with "citation needed".

Now here’s the key difference for Windows 8: as in Windows 7, we close the user sessions, but instead of closing the kernel session, we hibernate it. Compared to a full hibernate, which includes a lot of memory pages in use by apps, session 0 hibernation data is much smaller, which takes substantially less time to write to disk. If you’re not familiar with hibernation, we’re effectively saving the system state and memory contents to a file on disk (hiberfil.sys) and then reading that back in on resume and restoring contents back to memory. Using this technique with boot gives us a significant advantage for boot times, since reading the hiberfile in and reinitializing drivers is much faster on most systems (30-70% faster on most systems we’ve tested).

Source: Microsoft, "Delivering fast boot times in Windows 8", Building Windows 8 blog, September 9, 2011 [1]

[1] http://blogs.msdn.com/b/b8/archive/2011/09/08/delivering-fas...

That's certainly interesting, and thanks for proving parts of my argument wrong. I do appreciate being corrected. However OP made the following claim:

Windows 8 lies to you when you "Shutdown." Windows 8 actually hibernates instead of shutting down for real; and you can get the same effect by enabling hibernation on Windows 7

Note the part that where he claims that you can get the same speedup on Windows 7 by enabling hibernation. This is provenly false and anyone who has tried it would know.

You are correct that "enabling hibernation" should have no effect on Windows 7 startup or shutdown performance, though I have no idea how quickly Windows 7 can enter and wake from hibernation with no user sessions open. While this would be easy enough to test, I also don't care: Windows Server 2008 R2 cold boots to a usable desktop in twelve seconds on my machine, which should be fast enough for anyone.

If you hibernate in Windows 7, and then boot, it takes approximately the same period of time as "booting" in Windows 8 (with "Fast Startup" enabled).

I have tried it. I do know. In fact if you measure it "to desktop" then Windows 7 full hibernation beats Windows 8's fast startup, because Windows 7 caches user programs in additional to Microsoft's side (i.e. registry, services, kernel, etc).

Interesting. Now allocate 20GB of non-paged pool and try again.

That's a fairly obscure scenario. Plus, if that allocation was within a service or driver then both Windows 7 hibernation and Windows 8 Fast Startup would have to page it to the Hibernation file.

"Interesting" was my main point; the non-paged pool comment was mostly kidding. Interesting in that it could indicate a "de-hibernation" process that's nothing like the "inverse of hibernation" one might naively think it should be: while hibernation must write 100% of allocated physical RAM to disk, "de-hibernation" need only restore non-paged pool plus whatever else is necessary to reinitialize the virtual memory subsystem (including whatever drivers are necessary to support it, of course). This done, the balance of the hibernation file could effectively be treated as a page file.

Hibernation over shutdowns massively improves the shutdown and startup times of my work PC.

Primarily because I have SQL Server on it for development - the services have a extremely long shutdown and startup time, which doesn't need to happen for hibernation.

Interesting. I rarely bother to keep MSSQL loaded because it starts so quickly...

  PS C:\> (Measure-Command { Start-Service MSSQLSERVER }).TotalMilliseconds
  1127.5208

Alternatively, if you want SQL Server to start automatically without materially impacting startup performance, why not set its startup type to "Automatic (Delayed Start)"?

As for shutdown, while it depends on how you're using it immediately beforehand, I've never seen it take more than a few seconds on a development box.

I do the same thing because of Visual Studio and MSSQL - on Windows XP...

Sure, Microsoft fans never need to cite anything. Being Microsoft is enough. Oh btw, all that nice features already exist in other OSs since a very long time ago. Oh and also: can we finally stretch the Windows console horizontally?

You can already increase the width of the console, you just need to use Properties menu.

> So if you end up with an empty session, it's provenly not hibernating, it's booting.

Not necessarily. You can hibernate the machine after the user is logged off and restart from that point.