Ubuntu Vs. Windows Boot War (On an old budget laptop)

Video is ready, Click Here to View ×

This video compares the boot times between Ubuntu 16.04 and Windows 10 on my 2010 Acer Aspire 5734z.
(With a dual core 2.3GHz Intel Pentium T4500 processor, 4GB DDR3 Memory, and a 1 TB Seagate SSHD. Windows quick startup is disabled.)

It is very important to note that this is not a test to show anything or show any favoritism one way or another. Both operating systems are on a 1TB Seagate SSHD, which means I have no control over what is on the SSD and what is on the HDD. This might very…


  1. Windows 10 is such a bloated POS that even SSDs sometimes bottleneck it to the point of annoyance. I don't know how people consider its performance on a HDD to be tolerable. Starting with Vista, MS started the trend of splitting everything into smaller and smaller files and Windows 10 (and especially AppX) continue the trend. It's difficult to defrag effectively when the OS has been reduced to files less than the size of a cluster.

  2. Actually, to do the test properly, you need to start some applications.

    Because, with Ubuntu, once you see the desktop, then you can get going.

    But, with Windows, it can be extremely sluggish initially, as it still does an awful lot of work AFTER the desktop shows. Which is kind of a deliberate cheat.

    Now, in both cases, the "notification area" programs will load only after login – because they're a per user thing – but there is a fundamental design difference between the two OSes that gives Linux the edge here.

    In Linux, you have package management. Software comes in "packages" and this governs installation / uninstallation and updates. The upshot of using package management is that updates are handled by the system itself. The system itself knows what software packages are installed, it knows their dependencies on other software packages and it'll keep it all up-to-date with a single system-wide update process.

    Windows doesn't have this standardised system, so what happens is that every single application has its own installer / uninstaller program, and every single application has to deal with its own updates. In a thoroughly "free for all" manner.

    What this tends to mean is that applications individually add their own update daemons to the startup process, and things get sluggish, while pretty much every application in your system all simultaneously piles onto your Internet connection to look for updates (and, indeed, if there are updates, then install those).

    And Windows itself is not exempt from this either, because Windows' own Windows Update daemon kicks in just after boot.

    I wrote some cross-platform code, you see, and I wanted to test its performance on Linux and on Windows. The thing is, to properly gauge performance, I have to let both systems fully settle down, so that things like background update processes aren't skewing my results.

    So, you know, I'm not being evangelical here – I really don't care what OS you want to run, as long as you're happy with it – but just detailing, from actual practical experience (and much frustration waiting), that you can't simply measure "power on to seeing the desktop" as a genuine gauge of boot performance.

    I mean, any OS could conceivably show you the desktop graphics almost instantly. Just load the desktop wallpaper and show that to win this "race" and then take your time to actually complete the boot, during which the user can't realistically use their PC but. hey, it superficially looks like it wins this simple test.

    So what I'd say is redo these sort of tests but rig them to start up some applications – let's say Thunderbird email and a LibreOffice document (so that we're using the same applications on both for a fair test) – as well as waiting for all the notification icons to fully load in, so that you're measuring not to the point where you can see your desktop wallpaper, but to the point where you can actually start working.

    Trust me, I use both Ubuntu Linux and Windows 10 all the time at work, and, on the measure of how fast I can start working, Linux wins because, when you see the desktop, it really has loaded (at least some 90+% of the way there anyway that a tiny bit of extra loading is no impediment). Windows, though, is pulling a few psychological tricks on you to make you think it's finished before it genuinely has and, really, its "boot sequence" should be measured up until the disk and network access stops thrashing about wildly.

    (Plus, on Ubuntu, I could further install something like "preload", which monitors your application use and preloads, in otherwise idle time, your most used applications to pre-emptively have it ready for you. So the edge it has already can be made slightly better again, if you want, and are happy with a little RAM vs. load speed trade-off on the side, as you've got plenty of RAM to spare.)

Leave a Reply

Your email address will not be published.