Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
×
Chrome Software The Internet Hardware Technology

Chrome Will Soon Lose Support For Some Ancient CPUs (techspot.com) 141

If you're one of the few people still using a PC with an x86 processor more than 15 years old, here's another reason to upgrade: the devices will not work with future Chrome releases, starting with version 89 of the world's most popular browser. TechSpot reports: The Chromium development team announced that CPUs older than the Intel Core 2 Duo and AMD Athlon 64 would not work with Chrome 89 and future versions as they do not meet the new minimum instruction set requirement of SSE3 (Supplemental Streaming SIMD Extensions 3) support. So, if you are still sporting an Intel Atom or Celeron M CPU, you'll soon be counting Chrome as one of the many programs that are incompatible with your potato-like rig. The devices will no longer attempt to install the browser, while running it will result in the software crashing. It's noted that the change only affects Windows as Chrome OS, Android and, Mac already require SSE3 support.
This discussion has been archived. No new comments can be posted.

Chrome Will Soon Lose Support For Some Ancient CPUs

Comments Filter:
  • by cypherljk ( 201011 ) on Wednesday February 10, 2021 @07:55PM (#61049196)

    The problem with software today when your browser, a darn web browser dictates what computer / cpu you can use.

    • by Anonymous Coward on Wednesday February 10, 2021 @08:20PM (#61049266)
      I find it hard to believe someone on this site has never complained because they had to support some ancient piece of shit software or hardware.
    • That's the problem?

      It's noted that the change only affects Windows as Chrome OS, Android and, Mac already require SSE3 support.

      You'd think that a /. "editor" would notice something missing in there and maybe call it out.

    • >> a darn web browser dictates what computer / cpu you can use.

      The darn browser _IS_ the "computer" these days, for most users.

    • by mykepredko ( 40154 ) on Wednesday February 10, 2021 @10:33PM (#61049492) Homepage

      I'm trying to figure out a reason why any software should support all the hardware out there and I'm not coming up with any good reasons.

      Maybe there are some government/military systems which have to use 15+ year old hardware (IO buses or software requirements) but, if that's the case, then I would think that the people responsible for the systems don't want people surfing the web on them and making them available to being compromised.

      As the person who supported IBM Token Ring NetBEUI from 1986 to 2006 I can say it's really frustrating keeping up with new OSes and processors when the sole purpose of doing the work is so that a marketeer can put that support is available for the latest OS/hardware when there is literally NOBODY asking for it.

      If running the latest version of Chrome is that important to you (Chrome 87 will still run, as will FireFox and others), then download Chromium and do your own builds for your non SSE3 system. If you're not the only person with this requirement, then you'll find others that will help you with updating the code base.

      • by tlhIngan ( 30335 ) <slashdot&worf,net> on Thursday February 11, 2021 @04:26AM (#61050206)

        I'm trying to figure out a reason why any software should support all the hardware out there and I'm not coming up with any good reasons.

        Maybe there are some government/military systems which have to use 15+ year old hardware (IO buses or software requirements) but, if that's the case, then I would think that the people responsible for the systems don't want people surfing the web on them and making them available to being compromised.

        There's a lot of hardware a lot younger than 15 years old that will not work. A lot of Atom processors, for example, I have more than a few Atom tablets or micro PCs that are well under 5 years old.

        Blame Intel all you want for making processors like that.

        • by AmiMoJo ( 196126 )

          According to Wikipedia every Intel Atom every made has supported SSE3.

          https://en.wikipedia.org/wiki/... [wikipedia.org]

          In fact if you look at the page for SSE3 and the list of CPUs that had it, it seems that TFA may be wrong. Intel Core (not Core 2 Duo) had it, AMD Althon 64 had it. It's been common since 2005 and I couldn't find any CPUs made in the last decade that didn't support it.

          https://en.wikipedia.org/wiki/... [wikipedia.org]

          • There's some extras the SSSE3, which often gets lumped in and which are less widespread, and aren't supported pre-bulldozer, those just scrape in within the last decade, though they were probably in the distribution channel a while longer.

    • Re: (Score:3, Informative)

      by KalvinB ( 205500 )

      Some people act like the world should just stay static as soon as they get comfortable in it.

    • The problem with software today when your browser, a darn web browser dictates what computer / cpu you can use.

      No, I'd say you are wrong. That's a problem only if you are inflexible and MAKE it a problem. Some points:
      1) Chrome is not the only browser in the world.
      2) Anyone running Windows on these 'ancient' CPU's (Personally, I consider an 'ancient CPU' being something like 6502 or Z80) and connecting it to internet is an utter, complete moron.
      3) If you use open source software, that will never be a problem; you can always compile the program in question to suit your CPU. Well, trying to compile Chromium on a 15-yea

    • Or a paint program.

      Oh, wait ...

    • by ledow ( 319597 )

      No, it's people putting 15-year-old machines and OS onto the modern net and expecting to function and perform in an identical manner into perpetuity.

      P.S. There's literally nothing stopping you using Chromium and compiling it without SSE3 support.

    • by antdude ( 79039 )

      It's not just web browsers too. :(

    • The problem with software today when your browser, a darn web browser dictates what computer / cpu you can use.

      How is that a problem of software today? Software has always dictated what computer / CPU you can use. This has literally always been the case and is why software is published with system requirements.

      Also worth mentioning Chrome is a bastard to use on a 4GB RAM system. The last desktop CPU which didn't have SSE3 support only had a 2GB max RAM cap. A user would have to be a special kind of masochist to want to run a modern Chrome build on that machine.

      If you spent $50 at a local pawn shop you'll have someth

    • [sarcasm] How dare people decide what hardware they want to support in their own code. They should code on whatever we decide. Next thing you’ll tell me is that I have access to their source code and I could fork my own development. [/sarcasm]
    • Blame developers and corporations for being lazy and cheap because they want to use the browser as a general purpose user interface which necessitates the user of massive JavaScript libraries that make a web page larger than a complete install of Windows 95.
    • by eepok ( 545733 )

      Well... that's compatibility for you. It's been a thing for a very long time. Some things are coded to require certain types of hardware. That's life.

  • by Penguinoflight ( 517245 ) on Wednesday February 10, 2021 @08:10PM (#61049244) Journal

    Supplemental Streaming SIMD Extensions 3 is known as SSSE3, which according to the tweet is not what is currently being targeted.

    SSE3 is, and is supported even on "ancient" CPU's like the newest P4s, the Atom, and the oldest Athlon 64's.

    Yes this distinction is quite confusing, but given that SSSE3 wasn't supported on AMD until 2011, google probably made the right choice here.

    • Incompetent Techspot "journalist" is incompetent. I don't know where they find these people; clearly not from /. Then again, writing about what you know nothing about, and having zero research skills...that's the norm.

      • Yeah, not uncommon, but surprising they couldn't even bother to read the original information where the devs are discussing the differences between the two, when they were implemented etc. It's bad enough to just clickbait someone else's reporting but not bothering to read it seems to be a new low.

  • A few fossil Thinkpad collectors may be annoyed but they can use another browser and many likely abandoned Chrome as a resource hog already.

    • they know the number of machines it will impact, and they link to an analysis in their document, but the document says "sorry, Googlers only". it's a literal secret.

    • >A few fossil Thinkpad collectors may be annoyed but they can use another browser.

      Could you rattle off a few?

      I have a working Thinkpad that is otherwise useful. Many websites are still browser agnostic and play fine but popular ones, like Youtube, won't work with the last supported version of firefox or Palemoon. IE6 is a non-starter. In spite of it's age I find it MUCH faster that my corporate laptop running Windows 10 on an i5 multi-core, multi-Ghz processor.

      Newer is not always better.

      And I
      • by havana9 ( 101033 )
        Problem is that nowadays JavaScript heavy pages could reduce an old machine on its knees. I have an old Windows 2000 machine with a Pentium IV. It's perfectly fine to use fot Firewire transfer or reading and writing floppies, but having to use Firefox ang o on a generic website is a pain, even if Fireox itself used to do ftp transfers from another machine it's perfectly fine.
        Another reason I still have that machine is because it's great to play DOS era games on the real hardware
  • by Don Bright ( 6770394 ) on Wednesday February 10, 2021 @08:21PM (#61049268)

    https://docs.google.com/docume... [google.com]

    tldr is that they want to use fast features of modern processors features like SSE3 and to do that while also supporting older CPUs they claim they have to use dynamic dispatch, (in other words loading different binary code at runtime), and the "Engineering Cost" is too high, and there are a "very small number" of processors that will be impact - then they link to a secret document that shows exactly what number "very small" is.

    Alphabet only has about 150,000,000,000.00 dollars cash on hand so it is understandable they need to cut down on costs.

    • by Bert64 ( 520050 )

      Given how bloated websites have become today, it's unlikely that these older processors would make for a very satisfactory browsing experience in any case. Aside from low performance, such processors tend to have fairly low limits as to the size of memory they can use too.

      But runtime detection of SSE3 support is for precompiled builds, if you're compiling it yourself you could build it to not require SSE3. There is almost certainly an architecture-independent implementation of every function in order to sup

      • by dryeo ( 100693 )

        Watching Firefox around the time they required SSE2, first they would have 2 implementations of some functions, then once they required SSE2, they removed the non-SSE2 version of the function.

        • by Bert64 ( 520050 )

          SSE only exists on x86, they are going to have implementations for other architectures and likely a generic compilable C implementation to aid porting to new architectures.

    • by AmiMoJo ( 196126 )

      "Engineering cost" does not refer to monetary cost, it refers to the amount of effort required and the compromises that must be made to support something.

      In this case it sounds like they would need to maintain parallel versions of the code, which means fully supporting both, keeping them secure and so on. All for the sake of supporting some very old CPUs in systems that will probably struggle to run Chrome anyway.

      If it's a problem for you then I suggest you ask for a refund.

    • Alphabet only has about 150,000,000,000.00 dollars cash on hand so it is understandable they need to cut down on costs.

      This isn't about cost in dollars. Engineering cost includes among other things distraction and work overhead which quickly can lead to mistakes, additional bugs, and distractions from relevant work.

      Before you pickup the pitchfork maybe also consider the "engineering cost" here is the poor end user who somehow thinks it wise to run Chrome on a machine with 2GB of RAM (assuming the user actually maxed out the RAM on the chipset of the day). Maybe they are doing the world a favour by reminding them that their

  • This is simply what it is, plus a certain amount of hand woven witch crafted SSE3 instructions ..

    And I highly doubt the hand crafted part.

  • by OrangeTide ( 124937 ) on Wednesday February 10, 2021 @08:32PM (#61049302) Homepage Journal

    But it can run on hundreds of ancient CPUs. And nobody can get it to mine bitcoin with javascript/wasm/NaCl when I use lynx. Plus I don't spend all day watching YouTube videos on sharpening knives and slam ball exercises.

  • Fark, I had a laptop with a core 2 duo in circa 2006/7. That's 2 laptops ago since I rarely upgrade. Throw it into the sea already.
  • What if I turn off "Hardware Acceleration" in the settings. Won't it still work?

    This makes no sense.

    • What if I turn off "Hardware Acceleration" in the settings. Won't it still work?

      No, "Hardware Acceleration" means it uses your GPU. This change will be compiling the main binary to use new instructions (SSE3) which did not exist on older CPUs. If you have a very old CPU then it simply will not run.

    • Re: (Score:2, Funny)

      by jensend ( 71114 )

      It is rather unlikely that you're actually using a machine that doesn't support SSE3. The summary was wrong; Atoms and most anything else people are still holding on to are fine.

      If you really are running Windows 7+ on a 2003-era CPU, here's the proper procedure for hardware acceleration:

      1. Open the nearest window.
      2. Unplug your computer.
      3. Defenestrate your computer.

      You have now reached the proper hardware acceleration: 9.8m/s^2 as it descends.

  • by linebackn ( 131821 ) on Wednesday February 10, 2021 @09:50PM (#61049434)

    Crap. I was presently surprised to find out they still had a 32-bit version that ran on 32-bit Windows 10. Will have to check to see if it will still work or not.

    But chrome has broken so much lately, Palemoon/Newmoon/Serpent/Firefox are still much better choices.

  • Yes, I'll not be able to use Chrome, and therefore, lose Vikky for alexa support. But guess what? Even though I am personally worse off, I think that, over all, this is a good thing.

    I wish windows/linux followed suit, and started to (progresively) require more advanced CPU/GPU support, to (slowly) weed out ancient machines.

    For example, windows could move (progresively) from requiring SSE2 to SSE4 (for enhanced performance). From requiring a WDDM 1.0 to WDDM 2.4 (for added stability and performance), and fr

  • > It's noted that the change only affects Windows as Chrome OS, Android and, Mac already require SSE3 support.

    Strange, my Lenovo Yogabook (previous version) runs Android and has an Atom processor. Mind you, it isn't anywhere near 15 years old ... but if they're going to mention Atom processors by name they should be more specific.
  • by jensend ( 71114 ) on Wednesday February 10, 2021 @11:54PM (#61049674)

    Unfortunate that TechSpot didn't check the facts before publishing.

    All Atom models support SSE3 - even the document linked mentions that march=atom implies not only SSE3 but also SSSE3. The Celerons that are impacted are those based on Banias or Dothan; Dothan is a 2003 design. Yonah, Merom, and Penryn Celeron M processors support SSE3. Any Intel Core (not just Core 2) supports it, as do even Prescott Pentium 4s.

    This will impact almost exactly nobody. Chrome dropped Windows XP support 5 years ago. If you're running Windows 7 or later, here in the year 2021, you are pretty certainly not running it on a 2003-era CPU. No reason to be whining here.

    • by thegarbz ( 1787294 ) on Thursday February 11, 2021 @06:59AM (#61050576)

      Actually it's quite possible to run Windows 7 32bit on a pre-SSE3 CPU. That said if you're running Windows 7 32bit on 2GB of RAM and running a modern Chrome release at the same time, that's a level of masochism that would make even a BDSM dominatrix say "Daaaam! Too hardcore for me."

    • I think they meant SSSE3 in general. SSE3 is not interesting, SSSE3 actually has some additions that makes it worthwhile over SSE2.

      • by jensend ( 71114 )

        While SSSE3 would disqualify more CPUs, it is very clear from Google's document that they're requiring SSE3, not SSSE3.

        And that makes sense. While Intel introduced SSSE3 in 2006 and not too many people are using Chrome for Windows on processors older than that, AMD didn't introduce SSSE3 until 2011. The Steam survey shows 100% of Steam users with SSE3 but only ~99% with SSSE support. If Google wants to make use of newer instructions they'll do runtime detection or something.

        • Then it is entirely pointless, there is almost nothing worthwhile in SSE3. Some now deprecated fetch and horizontal add.

          SSSE3 has the all important shuffle instruction.

  • Let's be real here (Score:4, Insightful)

    by Sitnalta ( 1051230 ) on Wednesday February 10, 2021 @11:55PM (#61049676)

    If you run CPUs this old, Chrome not working anymore is the least of your worries.

  • Surely the effort to support "older" cpus is significantly less than losing all those advertisement eye balls.
    • Yes. i know many will upgrade, but many wont and when the earth has billions of people, that small with old machines who dont upgrade is a lot of people.
  • If I can't buy an old machine on eBay anymore that I can use to check that the software is working, then it becomes unsupported / untested. At some point I will change my software so that it refuses to run on that old hardware anymore, typically when supporting the old hardware causes work for me.

    My boss is more willing to stop support for older machines, and he pays me.

    There is a problem since I make a living writing iOS software, and it's kind of impossible to get phones running iOS 10, iOS 11 or iO
  • Calm down (Score:4, Informative)

    by Artem S. Tashkinov ( 764309 ) on Thursday February 11, 2021 @04:30AM (#61050214) Homepage

    The whole issue is blown out of proportions by the /. crowd.

    Let's be honest: no one expects computers from the early 00s to be able to browse the modern web as most PCs from that era contain less than 2GB of RAM and are often completely usable for browsing the web anyways unless you're OK with watching a slide show and taking minutes to load a web page. We are talking maybe about 0.01% of running PCs. At this point you'd better buy a smartphone with 4GB of RAM which will cost you less than $100.

  • It just wouldn't die...running Mint and using it in the kitchen for recipes, gets wet, oily, floury....maybe Firefox will still keep on.
  • Comment removed based on user account deletion
  • Nawh. Another reason to not use Chrome.
  • Since new stuff always has security holes, I'd be inclined to go with the old ones that no one sane would target...
  • I remember old RedHat distros switching to i686 from i386 (which was probably i486 in practice). Compilers and programs move on, and uses newer instruction sets, better cache awareness, and of course more registers.

    The old CPUs are still supported in source. I think it might still be possible to find i386 Linux. But almost all mainstream ones are now x86_64 only.

    (Side note: Best option was Gentoo, where I could set flags for exactly what my machine has. However I gave up on them when their repository got ha

Top Ten Things Overheard At The ANSI C Draft Committee Meetings: (10) Sorry, but that's too useful.

Working...