It should not really matter how long it takes to boot your OS. Why? Because a good OS will start up once after a major system upgrade, then stay on forever after unless there is a power failure. If you want to shut down the computer you can hibernate the state to keep all your apps ‘running’ and files ‘open.’ You never really need to reboot, so why worry if a boot takes one minute, two, or three?
Unless, of course, your operating system is Windows. In this case, it matters a LOT how long it takes to reboot because you have to reboot the damn thing every time somebody’s hat falls off. So, Windows booting time is important.
JH at Linux in Exile has done a number of Windows/Linux boot comparisons. His results to date have always shown Windows to be slower. Many people insist that there are data that show that Windows is always faster. I personally know that Windows is slower because my wife and I have fairly similar laptops, but mine runs Linux and hers is a managed Windows system. When we both turn our computers on at about the same time, mine boots up, I do my stuff, I hand her my laptop and she does whatever web-based things she can do on my laptop, we make dinner, feed the baby and put him to sleep, maybe paint the living room, etc., then her laptop is maybe booted up, almost.
OK, I exaggerate but that’s what it feels like.
Anyway, JH has a new test that is slightly biased towards letting one of the two operating systems win the race to boot, and the results shocked me. Have a look.
Have you read the breakthrough novel of the year? When you are done with that, try:
or Paperback
*Please note:
Links to books and other items on this page and elsewhere on Greg Ladens' blog may send you to Amazon, where I am a registered affiliate. As an Amazon Associate I earn from qualifying purchases, which helps to fund this site.
I think that site is horribly biased. My 3 year old computer running Win7 takes a minute to boot up. Tops. Linux is a bit faster, by maybe 10 seconds or so.
My wife’s crappy laptop however, does take 1.5 -2 minutes to boot and load Chrome and I would expect Linux to show a significant difference there. But if you Windows 7 computer is taking 3 minutes or so to load up, you’ve done something to it wrong or you have a crappy computer. No other way around that.
John, the computer and Windows OS in question is that provided by a top notch technical support facility at a major institutor’s, so I would expect it is an implementation of the Windows 7 OS on a recently built laptop that is exemplary and typical. My wife’s case to which I refer is similar. But I admit my description was biased (as it is rather tongue in cheek, obviously).
So, maybe it’s your computer that is doing something different than typical or expected? Or, perhaps things vary from case to case? Either way, the Linux computers still seem to always be faster, either a few seconds or longer faster.
Remember, this is isn’t about how long it takes to just BOOT UP, but how long to boot up AND login AND start a browser AND display a page. Typical for booting a laptop to get to work.
Just booting to get into a desktop, sure I’d believe that this takes about a minute on Win7. But that’s a deceiving measure. Try timing how long it takes for you to do exactly what this page describes, and I’m sure you’ll have the same or similar timings to Linux in Exile’s.
I think boot times really matter. Sure we could leave our linux boxes on all the time, but if you are not using the computer (over night, weekend) then you really should shut it down. Even with the screen off, computers still draw power, and if you care remotely about the environment, you should do your bit to reduce energy consumption. So shut that PC off.
I just tried booting my HP laptop (on batteries rather than AC, if that makes any difference)and using Firefox to get the Google page and it took me 1 minute and 35 seconds (using Windows XP). This is not a new machine, and it was a low-end machine even when I bought it (Celeron M 1.40 GHz, 500MB RAM). The guy at the link had a Dell some 4 or 5 years newer with a Intel Core2 CPU (1.20GHz) with 2 GB memory and it took him 3 minutes and 7 seconds. I’m hardly an expert computer guy so it doesn’t seem likely that I’m so clever as to get my machine working so much better than his. Why my ancient machine boots almost twice as fast as his I have no idea, but it does make his numbers seem a bit off — his Linux laptop, also with better specs than my HP, took over 30 seconds more to do the deed.
I’m also not a Windows fanboy (I’m still nostalgic for my Amigas — 16 second boot time, BTW), and expect to use Linux if I buy a netbook next year; I’m looking forward to it because I do like to fiddle about with new stuff sometimes.
anthrosciguy, you may want to clarify that you are comparing Win XP boot time in your comment, since those are not the timings mentioned in the article.
I know the example he gave – and while I only use Chrome (which is faster loading than Firefox) as soon as my Windows loads up, it’s ready to go.
Again, Laptops, especially those with lower amounts of RAM load slower. Also, how much bloatware is loading up with your computer. Nortons or McAfee, both which are far slower and no better than freeware, add a significant time, especially McAfee. Check out how many unnecessary programs load up when you boot up your computer (From run type “msconfig” and then go to Startup and see how many things boot up. Most things in there are not required and can greatly slow down your startup times. By minutes actually. Most people I know have Adobe, Itunes, Norton’s, and about a dozen other things listed. All which do nothing but slow down your computer.
Richard, I agree with you about turning the computer off and stuff, but booting up can take one minute or three and it does not matter much for that desktop you are going to use for all day.
I have read recently, though, that this whole boot up thing is ancient technology and that we should expect computers in the not too distant future to start up in a few seconds.
Right after that, I think they’ll be issuing those jet packs we were promised….
“Unless, of course, your operating system is Windows. In this case, it matters a LOT how long it takes to reboot because you have to reboot the damn thing every time somebody’s hat falls off. So, Windows booting time is important.”
That’s a load of horse shit. I can’t believe you would post something as ridiculous and unfounded as that (actually, I can believe it because you’ve posted stuff like that before). Yeah, in 1995 that might have been true but it hasn’t been the case for a long time. It’s a myth perpetuated by Linux and Mac users because they can’t come up with valid reasons to not use Windows.
I expect more from scibloggers who are supposed to be (or, at least should be) on the side of truth and rationality. I only came back to reading your blog recently after a long absence due to previous idiotic statements I’ve read here.
57.77 seconds from choosing which OS I wanted (IE, not counting the Bios loading which is exactly the same for both Windows and Linux) to Google.com loading up on Chrome. Like I said, Linux is about 10 seconds faster (and free and open source), but the massive boot-times I have seen on friends computers have nothing to do with Windows but with Bloatware (though it is Windows fault for allowing them to force themselves into the boot up process).
@Greg – Google Chromium, Google’s Chrome Web-Based OS, is supposed to boot in seconds since it’s essentially just booting up a slighly souped up web-browser.
There is also a super small embedded OS (from cell phones) that some laptops have that allow you to boot up your super powerful computer in mostly brain dead mode, but really really fast, to do minimal computing (may be as an appliance or something to check your email, I’m not sure)
Greg’s already revised his statement on this, but I really don’t get why so many Linux users claim they never reboot their computers. I use Linux on my home computer, but leaving it on 24/7/365 doesn’t make sense.
1. Even Linux crashes.
2. Leaving a computer on when it’s not in use wastes energy.
3. Computers are more vulnerable to power surges when they’re on.
4. Laptops.
@MRW – A good reason to leave your computer on 24/7?
BOINC 🙂
I might not be able to contribute realistically to any major science field (above average smarts but nothing special). However, I can let the real scientists use my computer for real science work. My current flavor being Einstein@Home
The issue is complicated. There is a good argument that was advanced on this blog by Robert Thomson that he thermal shock to components is still a real problem, so simply turning a computer off whenever it is not used is not automatically the best thing to do. Robert, however, does not really think Global Warming is a big problem, so he may have a different opinion than others on how much energy is used.
In my original unrevised statement above (I’m not sure where I’ve revised …. it may be that “inconsistent” is more descriptive of my position on this than “revised”) I state that hibernation is not rebooting, which is not really true in some ways … and it is in other ways .. and in that statement is hidden a jab at windows that no one has (it seems) picked up on. The reason you want to reboot (not hibernate) windows is that the OS continuously breaks itself just by being used, and rebooting can unbreak some (but not all) of that. Linux does not do that. Even a badly behaved app can usually just be killed and you’re fine.
(Having said that, there are apps in Linux that can break things, but if you use a good distro and still with the repository that will be very very rare.)
Greg,
“Revised” is probably the wrong word. “Backed off?” Anyhow, I just put that in to acknowledge that I wasn’t strictly arguing against what you were saying, but rather against an attitude I’ve read far too many times elsewhere in the Linux community.
I remember reading somewhere that the function (= Linux command or DOS command) that would work from the Windows command line to give the number of days since last bootup was removed from the system somewhere around XP. Out of embarrassment, presumably.
I’m sure many of those Linux users really do turn their computers off more than they admit. It’s just that they don’t have to.
The boot time is important because laptops get shut down all the time.
Some people out there may even confuse the boot time with the “switch on” time in the case of laptops in “hibernate” mode.
@MRW: leaving a computer on all the time is only sensible if it’s a server or if it’s running calculations (or in contrived situations such as a cool environment with high humidity). It takes some effort to make Linux crash though. There are still some defective drivers which will cause the system to lock if you do the wrong things, and there are a few instances in which you can hang the local keyboard and screen (but you can regain control of the machine via the network, if you have a cable or WiFi and a second computer), but I haven’t seen Linux itself crash for over 8 years. As for WinDuhs, just last month I saw a BSoD on a computer running W7.
@Richard: I use suspend-to-RAM overnight to save power. Sure it takes a little more energy than shutting down all the way, but it saves enough boot time to be worth it for me. The reason boot times really matter for my laptop is that the battery tends to come loose while I’m carrying it from one place to another, causing a surprise reboot.
Incidentally, any idea why Ubuntu takes longer to come out of suspend-to-disk (hibernate) than to reboot entirely? Is there some easy way to fix this?
anthrosciguy, you may want to clarify that you are comparing Win XP boot time in your comment, since those are not the timings mentioned in the article.
One of the links in the linked post is about the Windows XP boot time, and that’s what I compared mine to. His was 3:07 and mine’s 1:35; I have no idea why his is so bad.
@anthrosciguy – It’s bad because he obviously has some serious bloatware on his system. Again, Linux has the upper hand here because it doesn’t let programs like itunes, Norton’s (which wouldn’t be on a Linux system anyway), quicktime, and probably a dozen other programs slow it down like Windows allows.
It’s an easy fix as I said. Just go to msconfig, then Startups, then disable anything that isn’t needed at bootup (most everything but an AV program, hopefully not Norton’s or McAfee) and Greg should see a significant difference.
John, you are not being realistic here. My wife’s computer (which I have not timed, but merely parodized, but trust me, it takes forever to boot) is a managed PC. SO, this is a combined issue: Windows sucks donkey dongs AND managed PC’s suck because your average IT manager is more interested in controlling the problems MS doles out than in letting users use their computers. The so-called bloatware is probably the stuff installed for network connections, anti-virus ware, etc. But whatever it is, it is what the wisdom of her employers IT experts have determined. Same with JH’s computer, IIRC. His Windows install was the official install of the place he works, and it was carried out by the IT department, and the system used for testing is the one they start you out with for professional use in a managed setting.
The IT-managed setting MUST be the standard for these things, because it is what you actually end up with in real live in all those places where you work where they force you to use MS products because some VP got a bribe. I assume.
Then put the blame where it belongs – with the IT people. Windows 7 does not naturally have a massive boot time, google it yourself, check out real tech sites for what results they get (here’s a hint, it’s under a minute). I haven’t magically super tweaked my computer, it’s simply running clean (which as I’ve said, is a flaw in Windows because it allows people to clutter it up horribly).
I’m just saying blame the IT people – because they are the ones slowing down the computer. Again, check any reputable tech site and see what they get for Windows 7 boot time or startup times. Don’t trust me, but don’t use one anecdotal blog to justify something either.
Managed computers boot slowly, and managed computers don’t get pwned easily. Some of the reasons are prompt security patches and daily updates to virus scanners. The big picture should compare total boot times, and downtimes due to updates and reinstallations to desinfect a machine. Like within a year.
But I must confess that I have no doubts about Linux winning the game.
Psst don’t spread Linux success story around please, let malware coders concentrate on Windows. 😉
I’ve had more hard hanging issues with Linux than I’ve had with Windows. JEdit would continually hang on me, reboot the only fix. And whoahboy, don’t try to mount an smb share over a flaky network.
@Aki – 2 Things. First this isn’t a “success” story. It’s from a blog that constantly shows at best, outliers, and tries to present them as the norm. Second, Linux typically will boost a bit faster, require less reboots, and install a bit faster (assuming a friendly distro). Third, Malware coders aren’t going to bother with Linux for several reasons (on the whole that is). First, Linux is a tiny part of the public market. I mean, less than 5% here. Only in the business/server market do you see significant numbers. Second, Linux is far more secure than Windows so is thus easier to create malware for people to infect their systems with. So I don’t think you have to worry much about Malware people bothering Linux to any significant degree.
Rob, I am calling you on this.
Jedit is a very simple reasonably well behaved program. It is not very likely it was “hanging” your Linux system. Something else was going on, or you are making it up.
“t’s from a blog that constantly shows at best, outliers, and tries to present them as the norm.”
Actually, no, it’s not. It’s a blog written by a well respected and experienced IT expert about the experiences he actually has. He does not select among multiple experiences and pick the most Linux-favoring, because he is not having multiple experiences.
@Greg – I’ve gone through his site and I’ve seen zero examples in his “tests” of how his Windows experiences are outside the norm. By having a “test” and not stating the results are atypical, he is promoting misconceptions. Otherwise, you’d say, Man, it took 6 hours to install windows 7. That’s not typical for most people but it was a nightmare. See, that is clearly stating all the facts and is being completely honest about typical experiences, instead of trying to provide “evidence” of how bad windows is.
Well, John, I’m sure JH will be happy to know it’s not Microsoft. It’s that Microsoft hates him personally. Because this isn’t JH reporting the most unusual of a bunch of experiences within his workgroup. These are JH’s personal experiences, and I happen to know he’s not holding back a bunch of stuff that makes Windows look good.
@Stephanie – I’m not claiming he’s holding evidence back, but as we all know, anecdotes are not evidence. It’s easy to find info about what typical performance is and it’s easy to see that 6 hour install times and 2-3 minute bootups are not typical on a fresh install.
By presenting this as evidence (he clearly says the numbers speak for themselves)without stating what the typical numbers are, he is implying heavily (since the “numbers speak for themselves”) that this the type of performance you should expect. That is my sole issue with this. Anyone simply reading that post will get the wrong impression about the differences in performance.
Think of it this way. If you are presenting something as evidence for Linux being faster at bootup (and in an earlier post install times) and then post stuff like “It’s hard to argue with numbers, people”, can you explain how that is any different than going to Wattsupwiththat.com and having people post stuff like “Well, for the past year my location has been trending colder – it’s hard to argue with the numbers, people”
In both cases, there is something missing, IE, the bigger and more accurate picture. In GH’s case, it’s that his Windows experiences are outliers compared to what you’d find on just about any tech site. In the WUWT case, it’s that the colder temperatures in one area are an outlier compared to the overall trend. Both statements present information that is factual, yet both mislead in the exact same way.
John, you’re asking for something that isn’t part of the mission of the blog and complaining about something that isn’t represented as part of the mission of the blog:
Now, maybe you want Linux in Exile to have a different mission, or maybe you’re reading in claims about typical experience that aren’t there, but neither of those is JH’s problem.
It might not be listed as part of his “mission”, but are you trying to deny that he isn’t trying to present info in his blog about how much better Linux is than Windows?
His own words speak for themselves…
“It’s hard to argue with numbers, people”
John [29] You have stated that JH’s results are cherry picked, and they simply are not. There is no reason to think of them as atypical. He has modern equipment purchased by an IT department and configured in a standard way for Windows, run of the mill redhat/fedora Linux installed normally on the same equipment.
Read his post about the install of windows. Yes, that’s probably not typical, and he’s not saying it is. What is typical about it is that it was a disaster. One disaster may require two hours to fix, another two days. It took me four days to install Windows 7 on a computer here, and I gave up, it wouldn’t work.
I think if you read his post on the windows install you’d see it is pretty straight forward.
[31] Well, actually, anecdotes are evidence. The statement “anecdotes are not evidence” simply is not correct. If that was correct, there would be no criminal investigations ever brought to trial.
Are they systematically collected data? Well, as an anthropologist,I assure you that they can be (that’s called ethnography). However, this is not what we have here. JH’s experiences are not data in that sense.
However, his narrative is. Read his blog, starting with his first post, or perhaps his about page, to see what he is doing here. He is a Linux systems engineer type person with vast linux experience forced to work on a Windows computer in his job, and he was relating those experiences, including his putting Linux via a USB stick on his and/or someone elses windows computer.
Then, he got a job where he is not in a position to be told that he must use Windows on his own work computer, and other he works with use Windows as well. And he is relating those experiences.
Linux In Exile is actually an excellent narrative of Windows vs. Linux, and there is real meaning there.
You are suggesting that JH’s posts re invalid because they are not relating experiences that you have pre-determined are the only valid experiences. But you are simply wrong about that. Linux in Exile is doing what is stated, it is valid, it is interesting, it is useful.
The numbers certainly do speak for themselves. He’s made a comparison between the boot time of two systems on the same exact machine and reported this.
As far as I know, a lot of benchmark testing that goes on is done the same way: PC Mag gets a standard computer, keeps the hardware the same, varies the software, measures. Reports. The numbers. They speak.
In my own personal experience, he’s getting similar results that I have with both Linux and Windows, by the way.
In both cases, there is something missing, IE, the bigger and more accurate picture. In GH’s case, it’s that his Windows experiences are outliers compared to what you’d find on just about any tech site.
Have GTL an JH been merge??? Kinda scary.
Anyway, Linux in Exile is a tech site. The hardware was specified. The software was specified. The test was run. The results were reported. You don’t like the results so you claim that these results are not ok, only the other results, that perhaps you like, are ok.
John, you suggested that JH was cherry picking in your first comment on this. But you are the one playing with the little red fruits!
Are you suggesting that for a tech site report on boot up time to be valid they need to use multiple instance of the same hardware just in case there is measurement error or something?
Here is a standard tech site making the comparison:
http://www.tuxradar.com/node/33
The only thing that there is more of is number of different tests. Otherwise, the same procedure as JH’s
The difference between JH’s and the Tux text is that JH’s windows machine was configured as the IT personnel have configured it, while the TUX machine is the out of box install. Which is an interesting poit that I’ve made above: Security wise, Linux is essentially off the rack, Windows needs a tech team to make it work and not require constant reinstalls, etc. And that may be what makes the boot time longer.
Here’s another: http://laptopsall.com/windows-7-vs-ubuntu-linux-10-04-boot-up-time-comparison
John [34] .. Ah, I see what you are saying now. Right.
So, JH is required to have the opinion that Windows is better than Linux because there are more “official certified” tech sites and magazines and stuff (like ZD’s empire, funded maily by microsoft related advertizing) that say it is. Otherwise JH is biased.
Or, perhaps, JH is require to not have an opinion about windows vrs. linux. Again, read about the blog.
JH has been a top-ranked computer tech professional for years. He has used Linux extensively, and Windows extensively. If he is of the opinion that Linux has multiple advantages over Windows, that is not irrelevant. Are you suggesting that if he has that opinion he must be wrong? That would be interesting.
The blog is called Linux in Exile. It is about a Linux guy stuck in a Windows world, who got stuck there because others with power over him drank the windows kool aid. He is a Linux supporter. He knows that Linux is the superior OS and he sees evidence of that every day and he reports it.
You don’t see many pro windows blogs like this because most windows users wish they were dead half the time.
The numbers speak to the mission of the blog, John. Was being pushed to use Windows better for JH’s productivity? Nope. Some of that had to do with learning new things, as it does for any user switching systems. He documented that. Some of that had to do with how the two systems perform under the conditions in which he was working. He documented that too.
If you think blogging about comparing the two OSs needs to be done a particular way, it’s probably time to set up your own blog and make your own comparisons. Either way, it’s time to stop harping on the “clear” meaning of one sentence, particularly since others are interpreting it differently.
Oh, and a favorite Windows startup note: I got to watch Windows Explorer crash on startup this morning. That’s always fun.
I hate Windows 7 and it hates me. But my piece-of-crap Dell laptop with McAfee and other bloatware installed boots and displays a web page in under two minutes, passwords and all.
Just another bit of anecdata for those who appreciate it.
I had the Windows Explorer crash-every-five-minutes problem, big time, but it seems to be fixed thanks to the suggestions of other users at MS problem forums, zero thanks to MS tech support.
There still are days when I am lost in the wonder and awe of how anyone could have shipped Windows 7 and managed to live with themselves, but things haved gotten a lot better and I am getting stuff done as well as I did on XP. Almost.
I apologize in advance for typos. On XP I could use the Google spell-checker, but last time I tried on W7, it ate my entire philosophical tretise … tretese … rant.
@Emile:
It would somehow end up in a state that would hang and would never recover – no other program would do it. I lay it more at Java’s feet, not JEdit. It was also an old version.
I’m guessing it was related to this. Would’ve been around the right version of Java when it was happening.
@John – #29
» I’ve gone through his site and I’ve seen zero examples in his “tests” of how his Windows experiences are outside the norm.
I think I’m reading that correctly, where you certify the validity of my methods, except on this last post. Gee, thanks! 🙂
» By having a “test” and not stating the results are atypical, he is promoting misconceptions. Otherwise, you’d say, Man, it took 6 hours to install windows 7. That’s not typical for most people but it was a nightmare. See, that is clearly stating all the facts and is being completely honest about typical experiences, instead of trying to provide “evidence” of how bad windows is.
Here’s the deal. It’s explained very clearly on my blog (read “About me” in the left column) that I’ve been a Linux user since 1993. And from 2002 to 2009, I was fortunate enough to run Linux on my desktop at work. So for 7 years there, I didn’t know squat about how Windows worked.
Then, my new boss makes it very very clear he wants us all to run Windows. So, I run Windows at work.
Since I find it interesting when long-time Windows users experiment with Linux for the first time, I thought it might be equally interesting for this long-time Linux user to blog about my first experience running Windows in over 6 or 7 years.
That’s what the blog is about.
So to your comment of “not stating the results are atypical” … how am I supposed to know these results are atypical? I haven’t used Windows since 2002. So I call the shots as I see them, and I call out broken behaviour when it’s there.
I can only share what numbers I’m seeing on the system that I’m running. I try to be completely open about the system spec, and who installed it (our tech support guys, not me) and that this is a managed corporate system (i.e. not a home system.) But those are the numbers I get.
Rob: so “I’ve had more hard hanging issues with Linux than I’ve had with Windows.” means “I used a broken version of Java and it fucked me up so I got mad at Linux”
I don’t think that this test make much sense.
There are lot of linux distributions and they have all different performance.
So it’s not possible to comapre windows and linux boot time
Alex, it does not make sense to say that because there is variation in systems you can’t compare. Please read the comments above by Linux In Exile to see why he made this comparison. Then, google “Linux Windows Boot Time” and look at the posts that come up and you’ll see that this result (Linux being faster) is the norm (Although JH’s case is more extreme, because IT personnel had their way with his Windows OS first, which of course lengthens boot time considerably, though it would not with Linux).
The quotidian rationale for shutting down a comupter is avoid wasting power. Sciencebloggers generally believe in the reality of climate change, right? Or is that just too yesterday?
Gene, what is your point? I ask because you are way off topic.
I am a Windows guy. I love my dell to pieces. It is the first computer Iâ??ve ever had that is mine and I do not have to share it. It does take a minute or two to load, buy I have an HDD password on there first, so I have to put that in before I can do anything. I have tried Linux before and honestly, there is not that big of a difference. I prefer windows 7 more, but I do not see a difference in the two when talking about boot time. I just sit and have patience and then I can use my computer. If it took over five minutes, then I would be angry or frustrated, but I cannot think of any computer that does that.
Any New Installation of windows is faster than an old one
But will always get slower and s-l-o-w-e-r and more unstable with use and especially when you have all the necessary anti-viral stuff installed on windows just to try and keep it alive.
I have used linux and windows and although linux isn’t perfect the lack of endless painfull re-boots for this update and that error are a big plus for linux
It is reasonable to measure performance issues (e.g. start-up times) on weak machines because sometimes one can fill every db access (of loading windows).
So however maybe interested in my start-up time tests: Ubuntu vs Windows xp
on an relative old laptop.
Someone surprised that ubuntu rules? ;))
Anyone tried using BootRace to measure the time Windows needs to boot. Would it be an accurate tool?