Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Windows 8 To Feature 'Fast Startup Mode'

Soulskill posted about 3 years ago | from the zero-to-minesweeper-in-10-seconds dept.

Microsoft 287

New story submitter CSHARP123 writes "Microsoft has posted details about a Windows 8 feature that is a hybrid between cold booting and waking up from a hibernated state. This feature is called fast startup mode. Gabe Aul, director of program management in Windows, explains: '[A]s 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).' The post contains a video as well, which shows Windows starting up in less than 10 seconds."

cancel ×

287 comments

Sorry! There are no comments related to the filter you selected.

And it took them *this* long... (0)

strangeintp (892348) | about 3 years ago | (#37350654)

to figure that out?

Re:And it took them *this* long... (1)

Gription (1006467) | about 3 years ago | (#37350934)

Your thinking of it backwards. It isn't so much that they, "figured this out". It is more of a case of, "We are stacking the cards this much higher".

Re:And it took them *this* long... (2)

bemymonkey (1244086) | about 3 years ago | (#37350966)

They didn't really figure out anything new - it's more like they're forcing a half hibernate on people who'd usually just shut down or reboot.

For those of us who already use the available sleep/standby states or hibernate, the difference will be unnoticeable, because the reboots we perform (i.e. after software or driver installation, or after Windows updates) will probably still require the full shutdown we've come to know and loathe. :(

Re:And it took them *this* long... (1)

ByOhTek (1181381) | about 3 years ago | (#37350992)

Actually, faster than hibernate, slower than sleep/wake (without the power trickle requirements). Unlike either, you don't keep a lot of application states.
Faster than a standard start from shut down.

For people that don't hibernate/sleep their systems, it will probably be nice.

Re:And it took them *this* long... (1)

bemymonkey (1244086) | about 3 years ago | (#37351122)

"For people that don't hibernate/sleep their systems, it will probably be nice."

Of course - still sucks for the rest of us though :p

I was hoping for an *actual* speed-up in the full boot process :(

Re:And it took them *this* long... (1)

ByOhTek (1181381) | about 3 years ago | (#37351320)

It is irrelevant to the hibernate/sleep users, because those processes are almost completely hardware bottlenecked dependent.

So, what you are saying, is it sucks for people who are dealing with hardware bottlenecked options, that the software won't fix it???

You should be looking for someone other than an OS vendor to fix that issue.

Re:And it took them *this* long... (1)

adonoman (624929) | about 3 years ago | (#37351344)

They're doing that too in the form of intializing devices in parallel.

Re:And it took them *this* long... (2)

neokushan (932374) | about 3 years ago | (#37351002)

If you read the article, the hybrid booting is only part of the upgrade.

It’s faster because resuming the hibernated system session is comparatively less work than doing a full system initialization, but it’s also faster because we added a new multi-phase resume capability, which is able to use all of the cores in a multi-core system in parallel, to split the work of reading from the hiberfile and decompressing the contents. For those of you who prefer hibernating, this also results in faster resumes from hibernate as well.

Re:And it took them *this* long... (1)

bemymonkey (1244086) | about 3 years ago | (#37351148)

Also results in faster resumes from hibernate as well... at least in systems which are CPU-limited (as opposed to IO-limited) when coming out of hibernate.

Those with SSDs are on the safe side, I suppose.

Re:And it took them *this* long... (1)

neokushan (932374) | about 3 years ago | (#37351286)

Not necessarily. Annoyingly, they almost glance over this point, but if we read it carefully -

which is able to use all of the cores in a multi-core system in parallel, to split the work of reading from the hiberfile and decompressing the contents.

This sounds more to me like they've implemented a sort of FIFO system that allows the contents of the file to be decompressed and loaded as it's still being read from disk. It also implies (but doesn't directly state) that the operation is CPU bound, even on typical mechanical hard drives. I'd believe this, as even a cheap, slow mechanical HDD is capable of read speeds of 50-80Mb/s, which probably takes longer than 1s for a single core CPU to decompress, parse and load.

Time to Usable (4, Insightful)

Scutter (18425) | about 3 years ago | (#37350670)

Can we start talking about "Time to a Usable Desktop"? My laptop boots to a login prompt in 15 seconds, but after login it's another 2-5 minutes before it's done thrashing the hard drive. There are precious few (useful) tools available to track down everything the system is doing, and even fewer to help you improve the situation.

Re:Time to Usable (1)

drinkypoo (153816) | about 3 years ago | (#37350698)

There are precious few (useful) tools available to track down everything the system is doing, and even fewer to help you improve the situation.

Soluto [soluto.com] does both, for Windows Vista and better, anyway.

Re:Time to Usable (5, Funny)

Anonymous Coward | about 3 years ago | (#37350796)

There are precious few (useful) tools available to track down everything the system is doing, and even fewer to help you improve the situation.

Soluto [soluto.com] does both, for Windows Vista and better, anyway.

So... that means Windows 7, XP, 2000, 98, 95, 3.1, but not ME?

Re:Time to Usable (1)

RebelWithoutAClue (578771) | about 3 years ago | (#37351242)

You're right, it does work for XP.

Re:Time to Usable (-1, Redundant)

Nimloth (704789) | about 3 years ago | (#37351086)

for Windows Vista and better, anyway.

... so it works on OS X?

Re:Time to Usable (1, Insightful)

grimmjeeper (2301232) | about 3 years ago | (#37351304)

for Windows Vista and better, anyway.

... so it works on OS X?

...and Linux?

Re:Time to Usable (2)

doobydoobydoo (446165) | about 3 years ago | (#37351104)

There are precious few (useful) tools available to track down everything the system is doing, and even fewer to help you improve the situation.

Soluto [soluto.com] does both, for Windows Vista and better, anyway.

I tried Soluto on my old laptop and was not impressed. I had lots of things in Startup that it either wouldn't disable or that I couldn't disable (lots of system processes, sound driver, etc. - basic stuff I'd need). I disabled tons of stuff but it didn't really seem to be any better (and, in addition, Soluto itself needs to start up, which slows things down quite a bit. I tried the "Delayed load" option, but that didn't seem to improve matters much, to be honest. Whilst being quite aggressive in what I turned off, there seemed to be marginal improvement at best. I think instead there's something fairly fundamentally wrong with a lot of Windows services, etc., and maybe the Windows scheduling, that just take an age to load.

Re:Time to Usable (1)

UnknowingFool (672806) | about 3 years ago | (#37350734)

Yep same here. Now if you are part of an enterprise domain, it seemingly takes even longer.

Re:Time to Usable (3, Informative)

andrewbaldwin (442273) | about 3 years ago | (#37350938)

Now if you are part of an enterprise domain, it seemingly takes even longer.

And if you have a corporate standard image with policies etc pushed out on each boot....

On a cold boot, I can wander off, make a cup of tea, come back and it may just be ready. On a request for a reboot after a system update (and why it has to reboot after a change is yet another gripe) I could walk into town, go to the supermarket, buy a box of biscuits, queue up at the checkout, walk back and still be waiting for a usable system.

Strange that all that downtime x the number of users never really appears in TCO calculations -- I guess that's what meetings were invented for (so we'd have something to do without access to the IT infrastructure

...and people ask me why I prefer Linux !

Re:Time to Usable (1)

ByOhTek (1181381) | about 3 years ago | (#37351054)

I think the problem there is more due to your IT mismanagement. I lose about 10 mins a month due to patches. Maybe another minute or two due to the network security software starting up, allowing me to access our network.

Even with EVERYONE in this situation, we'd still lose much more time to restroom or cigarette breaks (and in the later case, most people here don't even smoke).

Re:Time to Usable (1)

Reverand Dave (1959652) | about 3 years ago | (#37351266)

That's why we do all our patching at night or on the weekends. In the case of some updates you may have delay when you boot up, but that is the price of keeping your environment up to date and secure.

Re:Time to Usable (1)

ByOhTek (1181381) | about 3 years ago | (#37351342)

That's reasonable. Prior to using disk encryption where I had to be at the console to log in, I restarted my system when I left for the evening, and had no downtime (that affected me) from updates.

Re:Time to Usable (1)

GIL_Dude (850471) | about 3 years ago | (#37350958)

The tool you want to trouble shoot this is xperf (specifically xbootmgr.exe) from the Windows Performance Toolkit. That is part of the Windows SDK. This tool will give you a look into exactly what is going on during boot and what is hogging disk, CPU, and everything. It is very detailed. Our Windows 7 boot is about 35 seconds from "starting Windows" to being at the desktop with the network icon showing an internet connection and being usable. xperf helped us to get to that state. The Windows SDK is here: http://www.microsoft.com/download/en/details.aspx?id=8279 [microsoft.com] . With their web installer you can select the components you want and not have to download the rest - for xperf you just need the "Windows Performance Toolkit".

Re:Time to Usable (1)

Anonymous Coward | about 3 years ago | (#37350742)

You're still using XP I'm guessing? I hate Vista/7 with a passion, but this is one thing that MS actually fixed. I have 3 Windows 7 machines, all are "usable" within seconds of the desktop appearing.

Re:Time to Usable (2, Insightful)

Anonymous Coward | about 3 years ago | (#37350778)

Doesn't really matter yet.
There is a sweet spot between when you push the power on button to when the computer is usable.
If it is about 5s or less then the user will stay at the computer during boot. If it is in the 10s or more the user will go and do something else that probably takes several minutes while the computer is booting and in that range it doesn't really matter if the boot time is 30s or 3 minutes.
A 5s boot time will still make the user reluctant to use the computer while in a hurry and will cause stress and heart problems for simple things like looking up at time-table for the train and similiar things.
Below 1s and there is not really any need to optimize it further. (Unless it is a server that gets its power on signal from the router when there is incoming communication.)

Re:Time to Usable (1)

Wowsers (1151731) | about 3 years ago | (#37350824)

In Linux it appears most the loading takes place before the login screen, then what remains is the users desktop after login, but it appears usable as soon as it appears. Certainly on my distro the only thing that slows down the desktop loading in using Superkaramba apps because there are no decent KDE4 widget replacements.

My Win7 install is frustrating, just like other version of Win, where the desktop looks like it's loaded, but you don't really know how long it will take until the OS releases control to the user. Microsoft seem to have gone for the "show something is happening" to make it look fast, rather than actually being fast.

Re:Time to Usable (0)

Hatta (162192) | about 3 years ago | (#37350870)

Indeed. They're closing user sessions here to get a shorter boot, that's essentially curing the disease by killing the patient. If I have a long boot time from hibernate, I can go do something else for 5-10 minutes. But if I've lost my user sessions, I have to sit there and reopen all my windows before I have a usable desktop. They might technically have a shorter boot time with this strategy, but it will end up wasting more of the user's time.

Re:Time to Usable (0)

Anonymous Coward | about 3 years ago | (#37350928)

I highly doubt they are going to kill the original Hibernate feature.

Re:Time to Usable (2)

neokushan (932374) | about 3 years ago | (#37351040)

They're not killing hibernate, they're not REMOVING anything, they're simply making the computer start up and shut down operate differently by default. They've actually increased hybernate's resume speed as well because of this, so everyone benefits.

Re:Time to Usable (2, Insightful)

JanneM (7445) | about 3 years ago | (#37350900)

Here's an idea for Ubuntu to beat Windows: Take a screenshot of the desktop when the user selects shutdown. Throw up that screenshot as the boot splash screen. Presto - Ubuntu "booting" in just a second.

About as honest.

Re:Time to Usable (1)

Pope (17780) | about 3 years ago | (#37350984)

I think you just invented iOS :D

Re:Time to Usable (0)

Anonymous Coward | about 3 years ago | (#37351112)

Ubuntu like the Linux systems seem ready when they present themselves. And it's common for an HDD to boot around 10 seconds without hibernation. With all the updates that OSes get vs how rare reboots are, hibernation tricks won't help many.

Re:Time to Usable (2)

couchslug (175151) | about 3 years ago | (#37350940)

That much thrashing indicates something is wrong and/or you have too little RAM.
The first solution for any old Windows install is to nuke-and-pave (format and reinstall). It takes less time than troubleshooting. Update, add apps, and see if the behaviour recurs.

Re:Time to Usable (0)

Anonymous Coward | about 3 years ago | (#37351006)

And yet it happens on my parents Win7 machine with 4gbs of ram and a core 2 duo. It's unfortunately too hard to get them to switch to anything.

Re:Time to Usable (1)

Pope (17780) | about 3 years ago | (#37350978)

No kidding. Back when classic Mac OS was around, it booted to a usable desktop faster than Windows did. So one of the big "improvements" for XP was to get the desktop open faster. Great, but it sure as hell wasn't usable for 20 to 60 seconds afterwards.

Re:Time to Usable (1)

omnichad (1198475) | about 3 years ago | (#37351200)

I think MS decided to "speed up" the boot process in previous versions by showing the logon prompt sooner (before a larger percentage of windows components had loaded). It still took the same amount of time, but you log in earlier in the process with the supposed apparent effect of a faster boot.
 
This might at least tighten up that same delay a little.

Re:Time to Usable (1)

agentgonzo (1026204) | about 3 years ago | (#37350982)

This is not about 'time to usable desktop'. It's about shortening the existing boot process (I agree that Windows is far from usable the first half-minute or so after it gets you to the desktop).

This image [msdn.com] sums it all up nicely, without the waffle, video or text.

Re:Time to Usable (0)

Anonymous Coward | about 3 years ago | (#37351004)

I have noticed that Windows is by far the worst offender in this.

Re:Time to Usable (1)

EdZ (755139) | about 3 years ago | (#37351028)

There are precious few (useful) tools available to track down everything the system is doing, and even fewer to help you improve the situation.

Assuming you're running windows, MSCONFIG handles things that run at startup. It's pretty easy to look through the list and disable anything you don't want.

Re:Time to Usable (1)

omnichad (1198475) | about 3 years ago | (#37351210)

On a temporary basis, anyway. It tells you where the settings are and lets you "disable" them, but the settings can come back if you want to change it from "Selective Startup" back to "Normal Startup," meaning you still have to go into regedit or the start menu itself to clear the entries out.

Re:Time to Usable (0)

Anonymous Coward | about 3 years ago | (#37351116)

Oh, you mean like my Mac?

Re:Time to Usable (1)

jader3rd (2222716) | about 3 years ago | (#37351126)

There are precious few (useful) tools available to track down everything the system is doing.

That's because Process Monitor does all you need to do. There's no need for another tool beyond that one. The primary reason why are because so many applications start as part of start up (iTunes and Quicktime being the worst offenders). You can disable them using msconfig. Windows 7 did make things better by causing application which run as part of start up to run as low priority, which as upset many developers who chose to run as part of startup to make their application "appear" to have a fast startup.

Re:Time to Usable (0)

Anonymous Coward | about 3 years ago | (#37351264)

What's the problem in thrashing the hard drive? As long as your system responds quickly it's not an issue. Prefetch does that but it doesn't hang your pc. Besides it will make your applications start more quickly.

so I can make a comment (1)

lecheiron (2441744) | about 3 years ago | (#37350682)

in the 11 seconds or less

Re:so I can make a comment (0)

Anonymous Coward | about 3 years ago | (#37350764)

If you really had it, maybe you'd have made first post.

will they have a real reboot (1)

mikeabbott420 (744514) | about 3 years ago | (#37350714)

will they have a real reboot for when I inevitably need it

having the same bolixed kernel coming back after a necessary reboot seems like it would be an pleasant experience

Re:will they have a real reboot (1)

clownface (633478) | about 3 years ago | (#37350836)

RTFA: shutdown /s /full

Re:will they have a real reboot (3, Informative)

adonoman (624929) | about 3 years ago | (#37350840)

RTFA and find out..

Hint... The answer is yes. But note that they do re-intialize drivers even in the hybrid boot, so that takes care of a majority of kernel level issues

Re:will they have a real reboot (1)

mikeabbott420 (744514) | about 3 years ago | (#37350910)

who has time to RTFA, complain and make excuses?
not me so I picked my favorite two.

Re:will they have a real reboot (1)

CSHARP123 (904951) | about 3 years ago | (#37350850)

They do have "Real reboot" option similar to Windows 7 cold boot using the UI or you can also do so from the command prompt shutdown /s /full / t 0 to effect an immediate full shutdown.

Re:will they have a real reboot (1)

neokushan (932374) | about 3 years ago | (#37351064)

Shutdown/Startup are hybrid methods.
Reboot works like it does today (i.e. is a FULL Shutdown/Restart).
Hibernate works as it does today, only faster.

Re:will they have a real reboot (1)

Ex Machina (10710) | about 3 years ago | (#37351190)

If you need to actually reboot your machine, you can always run Windows Update.

so the rootkit stays alive (0)

Gunstick (312804) | about 3 years ago | (#37350716)

best to have the rootkit reside inside the kernel so it's there forever. Clever system.

Re:so the rootkit stays alive (1)

Haedrian (1676506) | about 3 years ago | (#37350756)

Didn't think about that. The hibernate file doesn't have any special permissions or encryptions or something does it?

Re:so the rootkit stays alive (3, Interesting)

Thornburg (264444) | about 3 years ago | (#37350826)

Didn't think about that. The hibernate file doesn't have any special permissions or encryptions or something does it?

It doesn't matter if the file is protected. If you can breach the kernel, and store your malware/rootkit/etc as part of the "session 0" data mentioned in the summary, then the OS will automatically save it all for you. No need to crack the file.

However, the file does provide another vector for attack.

Re:so the rootkit stays alive (1)

Haedrian (1676506) | about 3 years ago | (#37350894)

Was more thinking of a program when asked to shutdown (not sure how windows works, but I assume there's a sigterm of some sort), writing the executable code into the file so it boots up on next load.

Re:so the rootkit stays alive (2)

neokushan (932374) | about 3 years ago | (#37351074)

If you've got THAT much access to the system, you probably don't need to do anything fancy to keep it there. Besides, if the user does a restart, it'll wipe the hibernate file anyway.

Re:so the rootkit stays alive (1)

Wingman 5 (551897) | about 3 years ago | (#37351088)

The file is erased and rewritten after everything outside of ring 0 has been terminated. So the only way to infect the hyberfil.sys would be have ring 0 write the infected file.

Who cares about the technical details (1)

jefferies (2456878) | about 3 years ago | (#37350760)

As usual the computer scientist spends about 95% of the time talking about the technical details of how this feature is achieved, and 5% talking about why the user would actually want to use it. Seems like waiting 4 minutes and 15 second for your computer to start are two pretty different options. There is definitely room for a middle ground, 30-70% less than the maximum.

Re:Who cares about the technical details (1)

reashlin (1370169) | about 3 years ago | (#37350884)

you must be new here

Mac (0)

Anonymous Coward | about 3 years ago | (#37350762)

So,, they are actually thinking of imitating MAC

Should not be a feature they advertise but implement, who would chose between 10 second boot vs 35.

Another feature looking a lot like the boost feature on old 486 tower when it had the boost option, who was stupid enough to leave it at slow?

Re:Mac (1)

Bobakitoo (1814374) | about 3 years ago | (#37351046)

Another feature looking a lot like the boost feature on old 486 tower when it had the boost option, who was stupid enough to leave it at slow?

DOS didn't have gettimeofday(), games timing was base on cpu speed. I remember taking 'turbo' off to slow down a older games that was running too fast, or slow down a difficult game...

But the same question can be asked today. If your cpu is know to be overclockable and stable, who is stupid enough to leave it at slow? Even today, there is reason to leave it at slow. More speed is not always the solution, sometime less speed is better and some other time less heat is better...

Re:Mac (1)

mla_anderson (578539) | about 3 years ago | (#37351238)

DOS didn't have gettimeofday(), games timing was base on cpu speed. I remember taking 'turbo' off to slow down a older games that was running too fast, or slow down a difficult game...

I can't speak to the specific function (after all it's been 16 years already), but there were certainly more specific timers to be had in DOS than the CPU speed. The turbo button was primarily on the XTs running at 8 MHz to reduce them to the IBM standard of 4.77 MHz. When IBM XTs all ran at 4.77 MHz the use of loop based timing was common, Frogger comes to mind. By the time AT computers were common place loop based timing was mostly going the way of the dinosaur as it became evident that programs would be run on different speed machines.

However, one place I worked had to keep buying pallets of 386 motherboards whenever they could even after 2000 because they ran their older data capture platform with loop based timers. Fortunately for the platform I worked on they hired some real programmers who knew better.

Re:Mac (1)

larry bagina (561269) | about 3 years ago | (#37351070)

Turning off the turbo button was for the sake of games that expected to run on a slower cpu.

Re:Mac (1)

BitZtream (692029) | about 3 years ago | (#37351134)

So,, they are actually thinking of imitating MAC

No, OSX doesnt' do anything like this. It just boots fast. No need to save the last boot and resume like this.

And for reference, a full resume from hibernate on my Mac takes about 8 seconds, faster than a windows quick boot.

Another feature looking a lot like the boost feature on old 486 tower when it had the boost option, who was stupid enough to leave it at slow?

Anyone who had a reason to? So basically anyone who used an app that expected a specific CPU speed rather than looking at a real time clock or something like that to figure out how to do things at the proper speed. During that time, most things ran on bear hardware and assumed they had full CPU, so they could do things based on CPU timing. When the CPU timing suddenly changed, everything that had any sort of video or audio was horribly broken.

The turbo button turned that off so things would run slower, more like the program was expecting. LOTS of people used it, you were just too busy putting your hands down your diapers to notice or remember.

Way to totally not have any clue what you're talking about, go back to middle school and shut your pie hole.

This is fine and all... (0)

Anonymous Coward | about 3 years ago | (#37350766)

When are they going to come up with the technology that boots before I press the power button?

Fast Boot (1)

mfh (56) | about 3 years ago | (#37350770)

Skip the loading of anything unimportant, full of bloated crapware? Sounds familiar...

Relevant? (1)

Grizzley9 (1407005) | about 3 years ago | (#37350780)

While a welcomed improvement to the 60 seconds or so that my current machine reboots in (I don't even know really), I'm not sure this even matters. Booting has never been one of the slow downs in my computer and shaving a few seconds off a boot, which is rarely done as my machine is hardly ever turned off, is not something I even care about them improving.

But will this give corporate IT directors a reason to upgrade since they can count those few seconds as "saved" x the number of workers = profit! Even though in reality it won't make any difference.

Really? (-1, Troll)

BurfCurse (937117) | about 3 years ago | (#37350790)

Windows has a kernel?

Re:Really? (0)

Anonymous Coward | about 3 years ago | (#37350830)

You Linux homosexuals really do my head in. Any post about Windows and you're all over it dribbling bile.

BTDT (2)

tibit (1762298) | about 3 years ago | (#37350800)

So they have done what LISP systems have been doing for two decades or more? It's a standard thing for a LISP environment to initialize the environment and store a core image of it to speed up startup. Same thing can be done for LISP applications, effectively giving you hibernation of individual apps in a clean state.

It's all about perception (1)

Infiniti2000 (1720222) | about 3 years ago | (#37350812)

Some of you are missing the point. It's all about perception, not a usable desktop. If the user thinks or feels like his computer boots up in 10 seconds, then he's happier. Happier customers tends to mean more $. But, yeah, I'd like to know what exactly is available to do after 10 seconds. For example, how much longer to launch a browser and see the latest feed on /.?

Rebootinate proposed on Matthew Garrett's blog (1)

Sits (117492) | about 3 years ago | (#37350814)

Re:Rebootinate proposed on Matthew Garrett's blog (1)

dirtyhippie (259852) | about 3 years ago | (#37350962)

Doing a full reboot followed immediately by a suspend is nothing like this approach. What are you on about?

Re:Rebootinate proposed on Matthew Garrett's blog (0)

Anonymous Coward | about 3 years ago | (#37351178)

Except for the part where is exactly the same on the other side of that operation. Stop focusing on the front end of the operation and realize the back end is exactly what MS is doing. Its the same but different.

I don't "get" it... (0)

Anonymous Coward | about 3 years ago | (#37350842)

Why would you switch your computer off in the first place, in the middle of the day?
I mean yeah, when you go to bed, then the workstation can be switched off.
But in the morning when you switch it on, you won't sit in front of it anyway, until you have said good morning to others and gotten yourself something to drink. (Or showered, eaten, etc for home computers.) Which is never less than 1-2 minutes, is it?
And during the day, logging out and shutting it down would be stupid since you have your whole work session open, and closing and restoring that is just making yourself additional work.

So why all that virtual wanking over boot times?
Monkey see, monkey do? Imitation-mania again?

If you want to save power, get yourself a very fast SSD or other non-volatile storage the size of your RAM, and let it save its hibernation file in there while powered off. Done.

Nice, in theory... (0)

Anonymous Coward | about 3 years ago | (#37350844)

But the reality is that whenever you swap a hardware device resume from hibernation fails. Since I swap hardware quite often I have to instead select reboot and hit the power button when the system resets. The vast majority of users will never see this problem, or see it so infrequently that it will be a non-issue, but not having the option to disable this feature sucks.

SSD (0)

Anonymous Coward | about 3 years ago | (#37350854)

Or, you could get yourself an SSD boot drive. I did so, and Windows boots in 14 seconds for me, and is usable immediately upon login, no waiting for everything to finish loading. I'll never go back!

SSDs are a better overall solution (1)

davros74 (194914) | about 3 years ago | (#37350858)

My current Windows7 boots to login screen in about 8 seconds, and after logging in, it's about 2 seconds to useable desktop. Of course, this is on a new SF-2281 SSD, that pumps out about 511MB/sec read rate on SATA-III controller. If you want fast boot times, people these days should consider an SSD OS drive (120-240GB), and a spinning disk for everything else (data, games, photos, movies, etc). The SSD improves a lot of aspects of performance, much more than just the initial boot time. Of course, this fast-boot on an SSD should be darn near instantaneous start times - unless of course it's not possible to speed up that swirling MS windows logo on boot. (Is that the bootup time bottleneck? Heh.)

With the memory footprint of something like Windows7 64-bit these days, a partial hibernation might be a good idea since full hibernation may require writing an 8GB or larger file to disk depending on how many applications are open. If you leave everything open and then hibernate, cold-booting might be faster, especially on an SSD OS drive.

Re:SSDs are a better overall solution (1)

jjjhs (2009156) | about 3 years ago | (#37350954)

I'm going to get an SSD when the price per GB is the same as HDD. I'm not going waste money just to boot Windows faster. My boot takes up to 30s to desktop, it isn't loaded with the crap that makes it significantly longer for other people even with newer systems.

Re:SSDs are a better overall solution (1)

Pope (17780) | about 3 years ago | (#37351000)

Riiight. Because having Windows boot faster is *clearly* the only thing SSDs have going for them.

Re:SSDs are a better overall solution (1)

davros74 (194914) | about 3 years ago | (#37351276)

Then you're going to be waiting a while. Even before SSDs, there were obvious advantages to running more than one HDD in a system. Back when RAM was very expensive, it was usually worth it to have a second HDD just to hold the swap file.

Tiered storage is done in Enterprise and no reason a similar approach cannot be done for the home user. You can easily get 2TB drives today cheap. You will not see SSDs there at the same price point for a while. The performance benefits of SSD are not really needed to bulk storing your DVD rips. Application loading, or any other application that tends to do lots of small random reads/writes will greatly benefit from an SSD.

For me, the small (120GB-ish) SSD replaces my old stragegy since the mid-90s of using a smaller, more expensive 15K RPM SCSI disk as the primary boot drive, and a cheaper, larger, slower IDE or SATA disk for storage. These days even a 3-tier storage solution is very practical: fast SSD OS drive, larger SATA 7200pm HDD for data and games, and an external 2TB+ NAS for bulk storage, archiving, backups. I don't think $200 is unreasonable for a performance oriented drive. Trying to do a 15K RPM SAS for $200 is practically impossible (being that a good PCI-e SAS controller will probably be equal to or exceed the cost of a SATA-III SSD).

I guess technically I have a 4-tiered system, since after getting the SSD, I didn't junk my 15K RPM SCSI drive. But since it's only 72GB, it was inadequate for Windows7 + Linux dual-boot. So SSD for OS, 15K RPM SCSI for user data, video editing/encoding, internal SATA 7200RPM for normal data, games, photos, etc, and then 2TB external NAS. High performance drives are always going to carry a price premium (15K SAS) - but on a modern day Sandy Bridge system (easily run at 4.5GHz), the old spinning HDD is really the bottleneck in the computer for almost all tasks - but yes, it's too expensive currently to replace ALL spinning HDD in a system with SSD. But I would contend that a $100 price premium for something that could quintuple your disk performance, is not a bad investment.

If you are the type what wants just a single HDD in their system, then yes, SSDs (or any other performance oriented drive, e.g., 15KRPM SAS) will simply be too expensive if you want more than 100-300GB of storage in your system.

Re:SSDs are a better overall solution (1)

jawtheshark (198669) | about 3 years ago | (#37351008)

Has Microsoft solved the problem that it's really friggin hard to separate user data and system data on a Windows Vista/7 installation? This is the best way I found to do it [microsoft.com] . On XP it was much easier to do and possible without a reinstallation.

I might be overseeing something, but given the popularity of SSD, it should be two clicks on a live system to do this and a (normal, GUI selectable) option during a normal install.

SHEZE A GONNA CRASH CAPTAIN !! PUSH THE BUTTON !! (-1)

Anonymous Coward | about 3 years ago | (#37350862)

Sell !! SELL !! ZELL !!

Always with the Fast Boot... (1)

S810 (168676) | about 3 years ago | (#37350868)

It seems that every version of Windows has promised a "Fast Boot-up Mode". I wonder if this one will actually deliver on its age old promise.

Re:Always with the Fast Boot... (1)

jawtheshark (198669) | about 3 years ago | (#37351018)

Broken Promises [youtube.com] . Cracks me up every time I see it, and I'm not a Mac user.

tux on ice (1)

wirelesslayers (2014486) | about 3 years ago | (#37350932)

TuxOnIce do a pretty good job for me, takes like 10s ~ 12s. laptop specs are: core i3, sata HD and 4gb ddr3.

They're breaking the only reliable fix!!! (1)

Anonymous Coward | about 3 years ago | (#37351020)

This makes me think that the universal fix-all solution to Windows problems - rebooting - will cease to work!

Why are people not using hibernate? (1)

jader3rd (2222716) | about 3 years ago | (#37351072)

That's so strange that people aren't using hibernate. That's mostly all I ever do. I still know some people who think that hibernate will draw battery power and don't use it. I also notice that it's the default button in the Windows 7 start menu, when really hibernate should be the default. Perhaps if the UI explained it better that hibernate wouldn't draw any power more people would use it. I never want a "fresh state", generally the only thing forcing me to lose my user session state is patch Tuesday.

Hmm (0)

Anonymous Coward | about 3 years ago | (#37351098)

Sounds like it will be great for persistent ring0 rootkits that don't require a physical file to exist.

Boot time isn't Window's problem (3, Interesting)

grimmjeeper (2301232) | about 3 years ago | (#37351158)

The biggest problem I have when running Windows, especially in a corporate environment, is all of the crapware that doesn't start until I log in. Those are the programs that decide to do massive tasks as soon as they're started. They bog down the network connection and thrash the hard drive doing their startup scans. They make the desktop completely unusable for significant lengths of time after login.

I suppose the fast boot to a login screen is useful. I'm able to get to the login screen quickly and log in. Then I can go get my coffee and read the paper while the startup applications take forever to do whatever it is they are doing. But it still doesn't solve the core problem of having a computer that is up and useful to the end user in a reasonable amount of time.

Now, it should be obvious that the blame here is not entirely on Microsoft. They have no control over what crap the end user (or corporate IT monkeys) install on the desktop. They can't control what gets started up when the user logs in. Microsoft has no way to prevent an idiot from writing an anti-virus package that does a complete system scan (that bogs down the entire system while it's running) when it is first started by the user. There's nothing stopping a startup program from waiting for a slow network connection to time out, causing the entire startup process to basically hang. There's nothing Microsoft can do to prevent a program to rebuild it's entire search index at startup, thrashing the disk to the point where the entire system is unresponsive while it's running.

But Microsoft is not entirely blameless either. The root of the problem was the decision to make the console the central focus of operation.There is absolutely no reason why so much of the software has to start up as soon as the user logs in. There is no reason why it cannot be tied to the startup of the computer. And if that software was tied to computer startup there would be no reason it could not be identified for hibernation just like the kernel, resulting in not only a faster boot time but a faster time to actual usefulness of the desktop.

Re:Boot time isn't Window's problem (0)

Anonymous Coward | about 3 years ago | (#37351354)

WinPatrol's delayed start-up is your friend. I'm not sure how much it'd help with the scans (the IT "monkeys" will probably get mad if you kill the scans entirely) but I've noticed that the general start-up stuff isn't so bad if you delay half of it for about 30 seconds or so.

Of course, you could always just get more RAM. You'd be surprised at the number of things you can run at the same time with 8 GB of RAM (at least if you only just upgraded from 2 GB).

Just like Google? (1)

kurt555gs (309278) | about 3 years ago | (#37351182)

My Samsung Chromebook has REALLY fast startup mode. All the time.

Back to the future? (0)

Anonymous Coward | about 3 years ago | (#37351186)

I still don't get how PC laptop users shutdown before closing the lid.

I bought my first Mac 10 years ago now and have been using Mac laptops ever since. All this time I just close the lid and it goes to sleep and when I open the lid it's there waiting for me to type my password in to unlock it.

The only time I reboot is when Software Update demands it, which is a few times a year.

Why is it that after all this time, the IBM PC industry still can't get its act together with regards to sleep? I've never had one that would do it properly like the Macs do.

Re:Back to the future? (1)

grimmjeeper (2301232) | about 3 years ago | (#37351212)

It's not that users can't hibernate when they "shut down" for the day. It's that the common solution for solving so many problems in Windows is a reboot at least once a week. Users are so accustomed to this that they just shut down out of habit.

Just moving the problem really (1, Insightful)

dingen (958134) | about 3 years ago | (#37351192)

So instead of waiting for the system to boot up, you now have to wait for the system to shut down (because it is writing the files required for fast booting). What an innovation!

Reality, the theory (2)

Lexx Greatrex (1160847) | about 3 years ago | (#37351204)

When I was young my dad and I built a go-kart that used the power train straight out of my grandmother's electric wheelchair. It was fast and looked cool and for a time I felt like the Alain Prost of my entire neighborhood. There was one small problem though... my grandmother was still using her wheelchair at the time. So when she wanted to go out, we would put the battery and motor back in the wheelchair and when I wanted to use the kart we would swap it back. It took about twenty minutes and since she only went out once or twice a week, it wasn't too much of a hassle (for me at least.) It is at this point in the story that my mother pointed out something so ridiculously obvious that it would probably baffle, disorient and possibly even permanently educate Microsoft kernel developers... you know who you are... you have been warned...

My dad, being an enthusiastic amateur engineer improved the swap out time by mounting the whole assembly to a bracket that we could take out of one and bolt into the other in under ten minutes. I remember him and me being excessively proud of this rapid start up time. But then my mother dropped a bombshell that changed the face of go-kart engineering in our household forever. She suggested that we use the DC motor and battery pack from the Flymo lawnmower that sat gathering dust in the garden shed because my Dad hated using it. The mower motor and battery were both lighter and with a bit of gearing, the kart was faster than ever. I was happy, granny was happy, dad was happy and the startup time was reduced from ten minutes to zero seconds.

Malware friendly (0)

Anonymous Coward | about 3 years ago | (#37351240)

Great. This means Windows becomes even more convenient for malware. Now malware writers don't have to worry about trying to hijack registry entries and squirrel away executables in obscure locations to restart on system reboot. Windows does this for them.

That's worse than Windows XP (1)

Anonymous Coward | about 3 years ago | (#37351258)

Windows XP can wake up from hibernation in that time, and then your computer is pretty much directly usable.
But Windows 8 will have to still load the user session - so you get less functionality, in more time. Hip hip.

S3? (0)

Anonymous Coward | about 3 years ago | (#37351310)

Just curious since most pcs dont go unused for that long anyhow, why not just do a S3 suspend? Juet a thought...

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?

Submission Text Formatting Tips

We support a small subset of HTML, namely these tags:

  • b
  • i
  • p
  • br
  • a
  • ol
  • ul
  • li
  • dl
  • dt
  • dd
  • em
  • strong
  • tt
  • blockquote
  • div
  • quote
  • ecode

"ecode" can be used for code snippets, for example:

<ecode>    while(1) { do_something(); } </ecode>