When are we going to drop the x86 32bit architecture?

Some “voices” inside me are asking when we’re going to drop the x86, 32bit architecture. Really, does it still make sense to support it? When is it going to not make sense anymore? 2012? 2013? Is it really disturbing us too much (I mean, having to maintain both arches)? Should we better focus on ARM?

Too many questions…


About lxnay

the creator of Sabayon Linux, Entropy Package Manager {Eit, Equo, Rigo}, Molecule release media buildsystem, Matter Portage buildbot/tinderbox and only God knows what else...

21 responses to “When are we going to drop the x86 32bit architecture?

  1. Radtoo

    Not sure when you want to drop it, but average people will be having a lot of old x86 machines for maybe yet another half-decade. Even some x86-only netbooks for instance are still available on the market.

    And yes, all that while the hobbyists and enthusiasts switched partially or wholly to amd64 more than half a decade ago.

    An ARM port would be kind of interesting, but the amount of additional work may be quite annoying. Figures more work would fall on you than Gentoo.org, if you offered such a thing. Besides devices with ARM on it can be annoying… not too few are locked down, don’t have a very generic BIOS, grub2 didn’t work on ARM last time I checked… and so on. Sure, maybe you want to do it for the challenge of it, but be prepared to sink quite some time into this, if you do.

  2. James Le Cuirot

    I know this relates to Sabayon but you might want to make it more obvious. This post has appeared on Gentoo Universe and I think many Gentoo users might get quite upset at such a suggestion!

  3. other thing to keep in mind is that intel 32bit isn’t dead- linncroft (released in 05/10) was 32b only. Frankly don’t much see the point in running most atom procs in 64b w/ memory available on most of those boards…

  4. alexrenzi

    I totally agree with Radtoo!

  5. klausman

    Also, there are quite a few embedded Chips (AMD Geode, eg) and derivatives.

    Granted, Gentoo is not necessarily the premier distro for those, but I’d hate to have to switch to something different for all my Alixes.

  6. Thomas

    We still deploy a lot of gentoo x86 servers for lots of different purposes that don’t need more than 4Gig of RAM, most of them are virtualized.
    So no, dropping x86 is not a good idea for the upcoming years.

  7. Stupot

    I think it depends on your definition of support.

    I can easily envision a future where DEs and apps are only made for modern computers, and since most modern computers are 64 bit, dropping 32 bit support for those.

    But I don’t think we’re anywhere near a stage where 32 bit support can just be dropped, if ever. I could see distros getting to a stage of ‘maintenance’ and ‘feature freezon’ for 32 bit arches, where only security updates and certain applications such as apache are still kept up to date and everything else remains frozen. But I can’t see 32 bit computers becoming negligible to the point where no support is given for at least a decade.

  8. Giacomo

    I think that instead of drop x86 it’s better to focus on one or max two DE(for exaple:keep KDE for the main edition and XFCE for a light edition). Most of the Chakra’s success come from the choice to matain only a desktop(less packages to mantain=clean desktop less updates smaller repositories). I know my propose will probably not be accetted but i think that the main problem of linux distro is the too wide possibilty of choice:most users don’t need 5 programs or 5 DE for do the same thing they need only one but that is stable and fresh updated and supported.

  9. lxnay , I agree with you too. Even Intel Atom is now support 64bit.

  10. tekwyzrd

    I use an x86 computer on a daily basis. It’s proven itself reliable and does what’s needed. I can’t afford a new computer or new components to build one. I would hate to see support for x86 processors dropped and to be honest it would force me to go searching for another distribution that plans to continue supporting 32 bit processors. I’m hoping to have a 64 bit computer assembled from used parts running soon but I plan to keep this one running until it’s hardware fails.

  11. Ramiliez

    I think i would be wise to count number of x86 users via update in entropy something like
    grep flags /proc/cpuinfo

  12. 4m

    Has the target audience of Sabayon changed? If it’s aimed at multimedia, keeping 32bit architecture would be like an F1 racer keeping an old slow engine.

  13. Kushagra

    I don’t understand we can’t just buy a new computer every second year.The later 32 bit processors are functioning just perfectly.Included the fact that they run more software than 64 bit today this decision must be delayed.

  14. It’s really not the case to drop x86. In my opinion, as far as there will be a light Sabayon release (e.g. LXDE) working flawlessly on old computers, mantaining x86 makes a lot of sense. I recently resurrected an old Athlon XP 2000+ with 256MB RAM and it works very well with LXDE. It’s not a ninja machine, but it’s possible to surf the web, edit documents, archive photos and so on. I also installed x86 Sabayon on various P4 >3GHz machines, with 2GB RAM. Those are in my opinion pretty good machines. And what about the wide number of netbooks with no 64bit support? And what for the wide numer of Centrino equipped machines?
    I can’t see the point of dropping this architecture. I could agree in keeping just a single light flavor, e.g. LXDE. We may claim that x86 is for old computers only, so we provide just an environment (others available but not supported) and focus our efforts on it.

  15. TheFuture

    Possibly do only 1 32bit version like a xfce4 or similar, small compact and upgradeable to gnome / kde on user input.
    Server, no idea why anyone would not be using at least 2 or 4 socket cpu’s as default and this would be 64bit AMD and Intel so might drop 32bit for servers.

    ARM is the future or ARM based chipsets with a few 10+ billion ARM chips sold each year there is a huge potential market and using crossdev and current “back online” ARM versions of gentoo should make it easy enough.
    ARM has the kernel source sorted and we are working on keeping it clean and decent from bugs for you guys.

    OpenCL latest news / would be a ideal solution but might be difficult to implement then you can cover both AMD and Nvidia cpu’s + the 100Billion ARM chips that is out there in tablets and mobile phones.

    We “not sabayon” have done 99% of the work for gentoo / sabayon already to make this switch easy even for the graphics of the mobile phones.
    So this ARM switch should be easy and Apple going to switch to ARM based laptop cpu’s or so the rumor goes and Intel trying to follow with whatever they can glue together.

    The last linked guys in this link below has a JetQi dual boot program for the mobile phone they use to get past Samsung touchwiz “ARM cpu” and this could be modified to work with Ipad1/2/3/4/5/6/7 and ARM MacBooks and many other as dual boot would be a great solution for most users.

    Installing Gentoo on Android chroot. (T-Mobile G2)

    Hope this helps.

  16. tekwyzrd

    The computer I’m hoping to have running soon (tomorrow if it rains) is a dual irwindale xeon with 6gb ddr2 pc3200, an 8 disk 3ware raid, and almost 1/2 tb of raid capacity (250gb of raid 1 and 240gb of raid 1+0), and an 8800 ultra. Not very impressive by today’s standards but not bad for less than $300

    I still plan on keeping this athlon xp running until it quits.

  17. BHReach

    I have been seeding Sabayon 5.5 torrents for months and it looks like 40-50% more people are downloading 32 bit than 64 bit. I think we should not drop 32 bit until a large majority stop using it, otherwise many current Sabayon users will have to switch to something else.

  18. Adam Jorgensen

    Pointless to drop x86. It still has it’s uses.

  19. Chris 'aka' Hopalong

    Whilst i agree, whilst thinking of those who develop (commercial and non-commercially) software and components, that maintaining multiple architectures is seemingly a lot of effort – but given how much widespread use of different distributions across different architectures (many and many, a bit old list in practise) when it comes to application software, modules, kernels and OS’s in general, then i think it’s definately a good thing to keep multiple architecture development going.

    Inside the Linux community, developers and users and admins alike, there is a fairly common ground where you don’t ditch the kit ’til it’s knackered or plain useless (and ‘plain useless’ to me, is a goldmine of opportunity to another) – and it’s the willingness to produce code that is ‘friendly’ to cross-archtitecture development that has been a big part of why open-source code has rapidly become to mainstay component in the development of many kinds of software and a wide spread of different forms in which Linux and other open software has become the norm.

    So whilst dropping the architectures that seem redundant and pointless may possibly speed up (through a narrow focus) future development, the friendly nature that embraces the wealth of architectures has been a big contributor to the sucess of Linux usage and propagation as it stands (and let’s face it, it’s far from over).

    To use a commericially toned example – do you honestly think that if development for many architectures and the willingness to adapt (and do the hard work, in practise) of others had anything to do with how Android has become the new and only really viable competition to IOS on devices and at the same time, become more widespread because it’s not narrow-focused or commericially strangled by the vision of one?

    Each to their own, now talking about developers – if a developer wishes to focus on x64 or whatever ‘not old’ architectures for their efforts, fair play – but each to their own, because whilst the willingness to embrace all remains in the dev world, more instances of ‘Linux’ (on whatever derivative you care to acknowledge) will continue to be the first choice (where it’s viable) for the devices of new and the future.

    Please don’t let us become too much MS and Sheeplike in philosophy and be focused on a vision of the few over what is right and proper for the future – the future is always unwritten (it’s a book being imagined and adapted by the microsecond into the screenplay of life) so it makes sense to let the ‘nature of the beast’ prevail, and if any particular architecture then becomes extinct in the evolutionary path of devices due it simply not being capable of being exploited effectively, then it’s time to maybe consider not supporting it any more – but let those who wish to, continue as forked developments.

    Just my highly devalued six million roubles on the subject :p

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

hello, twitter

Enter your email address to follow this blog and receive notifications of new posts by email.

Join 583 other followers


%d bloggers like this: