Announcement

Collapse
No announcement yet.

x86 FPU Optimizations Land In Linux 5.2 That Torvalds Loves But Worries Of Regressions

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • starshipeleven
    replied
    Originally posted by Buntolo View Post
    Is Linus referring to Pentium-era or modern Pentium CPUs?
    Pentium-era.
    Has this gone through lots of testing, particularly with things like
    FP signal handling and old machines that don't necessarily have
    anything but the most basic FP state
    (ie Pentium class etc)?


    That's old crap in x86 arch (so the historical Pentiums) and slightly less old crap in ARM arch (ARMv6 or v5 I think, where the FPU was optional and you could get away with anything you decided to call FPU).

    Leave a comment:


  • starshipeleven
    replied
    Originally posted by ThoreauHD View Post
    I just started getting some 'MODSIGN Can't find UEFI db list' error on 5.1.1. Not sure if it's because they suck or ubuntu mainline sucks, but somebody is sucking.
    Ubuntu mainline definitely sucks, but I'm getting that error since 5.0 at least on OpenSUSE.

    Leave a comment:


  • schmidtbag
    replied
    Originally posted by wizard69 View Post
    At this point nobody should care if Pentium era hardware fails. You can’t improve Linux if you are tied to old technology, sometimes you just have to move on. I’m pretty bad in this regard too as I just sent an old 486 based machine to recycling. After awhile though you need to say is it really worth it ?

    If old hardware issues are found I hope mainline developers have the strength to eat oh well time to upgrade.
    Also worth pointing out that I'm pretty sure the average Pentium user doesn't care about Linux 5.2. Most people running that hardware are deliberately using era-appropriate software.

    Worst-case scenario, they use a slightly older kernel. Not like they'll notice the difference.

    Originally posted by coder View Post
    Bug fixes, new features, and compatibility with some more recent userspace software that you need it for, are reasons enough.
    To my understanding, the only untested hardware is pre-i586. I highly doubt anyone with i486 (or older) hardware is seeking the improvements brought out in recent kernels. As far as I'm concerned, any of the kernels released within the past decade should have all of the stability and compatibility someone like that could actually take advantage of. Remember, we're talking pre-USB hardware here, where 4GB hard drives were considered massive. The kinds of people attempting to run anything newer than Linux 3.0 on these computers are the ones who compile their own kernel. If they get an FPU problem and for whatever reason really care about running Linux 5.2+, I'm sure they can figure out how to revert this patch.
    But I wouldn't knowingly break some old hardware without good reason.
    Improved performance or efficiency among all newer (and frankly relevant) platforms isn't a good reason?
    Last edited by schmidtbag; 05-13-2019, 08:48 AM.

    Leave a comment:


  • Jabberwocky
    replied
    Originally posted by Buntolo View Post
    Is Linus referring to Pentium-era or modern Pentium CPUs? I'm asking because he used the words "Pentium-class", which may refer to newer lower end Core i CPUs. Being lower end, they may only have some basic FPU functionality and thus have issues with these optimisations.
    I would like to know too, here's my guess. He mentioned "old machines" so I assume he is talking about pre-Core CPUs, up to Prescott/Cedar Mill era. However I'm sure it would be useful to test on more modern (Core i and later) Pentium too.

    Anyway remember that there is very old hardware still being used today, because it's very well known (e.g. aerospace, ATM, assembly line machines, etc.), so there may be situations where you may want a new kernel on old hardware.
    For general purpose computing (normal consumers) I understand the desire for new kernels on old hardware, but not so sure about specific industries. For example, many ISPs in my country use old versions of (free/open)BSD. Most if not all ATM machines in my country runs a form of Windows. Some tax systems here still runs on SCO.

    It is my understanding that Linux is used for most web servers and cloud providers, these industries upgrade hardware quite frequently. I wonder which major industry relies on running (newer) Linux kernels on legacy hardware.

    Leave a comment:


  • Michael
    replied
    Originally posted by FireBurn View Post
    Is this change for all x86 or just 32bit?
    all x86

    Leave a comment:


  • FireBurn
    replied
    Is this change for all x86 or just 32bit?

    Leave a comment:


  • rmoog
    replied
    Originally posted by Termy View Post
    personally i think it's the right thing that Linus pulled it despite the slight chance of regressions on really ancient hardware. It's unlikely that PCs from the Pentium-era need new drivers and most of them probably aren't even connected to the internet i would assume
    No. Me and my friends have machines with Pentium 1-3 and K6-K8 CPUs that do various things. These aren't business-critical things since they're all leisure machines. Nevertheless, we need internet connection on them on FreeDOS and Linux to share source code and abandonware.

    Leave a comment:


  • Buntolo
    replied
    Is Linus referring to Pentium-era or modern Pentium CPUs? I'm asking because he used the words "Pentium-class", which may refer to newer lower end Core i CPUs. Being lower end, they may only have some basic FPU functionality and thus have issues with these optimisations.

    Anyway remember that there is very old hardware still being used today, because it's very well known (e.g. aerospace, ATM, assembly line machines, etc.), so there may be situations where you may want a new kernel on old hardware.

    Leave a comment:


  • dkasak
    replied
    Originally posted by ResponseWriter View Post

    As a web developer who has to try to keep up with these trends I agree with you. This is where I find Otter browser to be a great asset because it's easy to disable scripts, cookies and images. When I depend on connecting via overpriced cellular data plans it's a massive cost saver.
    Hehe. Using web development as a yard stick is perhaps going just a MASSIVE BIT too far. What's the lifetime of a JS framework these days? Months?

    Leave a comment:


  • ResponseWriter
    replied
    Originally posted by creative View Post
    Even on intel 7th gen, I disable javascript 97% of the time. Why? Why do I need it? So it breaks a bunch of web pages functionalities I don't want? Good! Was just wanting to read text anyway, not see a bunch of fancy graphics, code and images I did not ask for.
    As a web developer who has to try to keep up with these trends I agree with you. This is where I find Otter browser to be a great asset because it's easy to disable scripts, cookies and images. When I depend on connecting via overpriced cellular data plans it's a massive cost saver.

    Leave a comment:

Working...
X