Your Intel based computer is (probably) going to get slower this month (along with the whole world).

Peter FulfordPeter Fulford Posts: 1,325
edited January 2018 in The Commons

 

I say "probably" just to cover all bases, but it almost certainly will. Will you notice? Probably.

https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/

Basically - longstanding, unfixable CPU design flaw means serious security hole can only be fixed at higher level operating system level. Such workarounds at these levels lead to significant slowdowns.

64bit Windows, macOS* and Linux will all be affected.

 

AMD users safe, and may chortle (within the TOS, of course).

 

I rather like "Forcefully Unmap Complete Kernel With Interrupt Trampolines".

 

*The Register piece has been updated to state that macOS has already been patched since version 10.13.2

 

Post edited by Peter Fulford on
«1345

Comments

  • I have not updated my main WIN 8 pc for 3-4 years. It runs my music DAW, and ever since I started the DAZ hobby it runs DS too. I always disable any external interference in my workstation rig, keeping it stable and predictable.

  • *chortles*
  • AMD chips may not have the vulnerability, but will the OSes use two different modes (one patched for Intel, one original for AMD) or will they use the same for both, with the speed hit affecting both?

  • JamesJABJamesJAB Posts: 1,760
    edited January 2018

    Quote from AMD:
    "AMD processors are not subject to the types of attacks that the kernel page table isolation feature protects against. The AMD microarchitecture does not allow memory references, including speculative references, that access higher privileged data when running in a lesser privileged mode when that access would result in a page fault."

    Guess they've not so covertly informed the world that the issue has to do with Intel's preemptive code execution potentialy accessing kernal memory that the current processing level isn't authorized to access.....   lol

    Post edited by JamesJAB on
  • Good question, Richard.

  • IvyIvy Posts: 7,153

    Sad news indeed. 

    I guess it looks like I'll be turning off windows updates notices on my i7 server machine.and update the hardware manually.  

    my win10 laptop is so slow i doubt i would notice a 30% decrease. But my ivy bridge intel i7 4.4.4 ghz  is in a server set up mode with external gpus and is smoking fast . 30% decrease would be definitely be a hard benchmark notice when rendering animation and would not be worth the update to me when i can pull the system off-line or put it behind my firewall and bump up restrictions and feed content to it through my LP .  either way this old system won't be learning any new processor tricks. 

    sad news indeed

  • bluejauntebluejaunte Posts: 1,861

    Ouch? 

  • KerwinKerwin Posts: 124

    It's probably premature to predict how much impact will affect some processor intensive tasks like rendering.   For example, "... H.264 video encoding, timed Linux kernel compilation, and FFmpeg video conversion tasks didn’t lose anything."  (See https://www.pcworld.com/article/3245606/security/intel-x86-cpu-kernel-bug-faq-how-it-affects-pc-mac.html)

    Linus Torvalds wrote, "“Some loads will hardly be affected at all, if they just spend all their time in user space. And if you do a lot of small system calls, you might see double-digit slowdown..."

    Michael Larabel ran some games testing on early patched kernels, "Dota 2, Counter-Strike: Global Offensive, Deus Ex: Mankind Divided, Dawn of War III, F1 2017, and The Talos Principleon a Linux 4.15-rc6 machine with a Core i7-8700K and Radeon Vega 64. None saw a frame rate change outside the margin of error range."

    I think we'll need to wait a bit for the actual patches to roll out on Windows and MacOS before we can.

    I think most of the issue will be in I/O heavy tasks (think databases) which have to cross from user space to hardware drivers at very high rates.  We'll see.

     

  • TaozTaoz Posts: 9,724

    What everyone seem to ignore in matters like these is that the real problem is not the CPU, but those who want to exploit the bug (which wouldn't really be a bug if no one wanted to exploit it).

  • HorusRaHorusRa Posts: 1,662
    edited March 2019

    .

    Post edited by HorusRa on
  • AllenArtAllenArt Posts: 7,140
    edited January 2018

    Basically it means that Intel chips calculations right hand doesn't know what its left hand is doing for long enough for some hacker jerk to get info it's not supposed to be privy to and can exploit and that a firmware update from Intel is not enough to fix it. Your choices are: update your OS and take the potential performance hit or buy new hardware. Period.

    Well, or disconnect from the internet and don't install anything new or put anything  new on your computer.

    *sigh*

    Laurie

    Post edited by AllenArt on
  • KerwinKerwin Posts: 124
    HorusRa said:

    What is all this about in laymans terms, for us non IT folk? smiley

    There is a defect (which apparently has been around for about 10 years) in Intel's x86 processor lines including i3, i5, i7, i9, Xeon, etc.   The defect lies in the internals of the processor that helps the operating system regulate between user spaces (think of this as your own private world in the operating system) and privilaged spaces (like the parts of the operating system that control disk drives, USB, video cards, audio chips, etc.)  Recently, hackers have found a way to exploit this defect.

    Because the chip is hardwired when it was made, operating systems like Linux, MacOS, and Windows that rely on these processors will have to change how they protect data moving from user spaces to privilaged spaces, rather than depending solely on the chips support.   This will make some things slower.   How much slower, and how that will impact a complex program is still being measured, but things that have a lot of data going out to disk drives (like databases) seem to be the most affected, while things that privately run lots of calculations, like audio and video encoding seem to be largely unaffected.   Games, which use the video card seem largely unaffected in terms of frame rate, which is a relatively good sign.

    HTH

     

  • Peter WadePeter Wade Posts: 1,603
    HorusRa said:

    What is all this about in laymans terms, for us non IT folk? smiley

    As I understand it there is a flaw in the design of Intel CPUs that makes it possible for a hacker to access parts of the operating system they should not be able to see. This could help them break into computers. Windows has to be modified to stop them doing this, but this means Windows has to do more security checks which take up CPU time so there will be a bit less CPU time for running applications. 

    To any experts reading this, please correct me if I have got it wrong.

  • GatorGator Posts: 1,268

    AMD chips may not have the vulnerability, but will the OSes use two different modes (one patched for Intel, one original for AMD) or will they use the same for both, with the speed hit affecting both?

    Good question.  Sounds like it's up in the air.  It's easy to filter for the OS, they can just filter AMD CPUs but I heard they may not.  If they don't, I'd expect AMD to put up a big stink about it.

  • nicsttnicstt Posts: 11,714

    Makes me glad I bought a Threadripper

  • "I'm extreme about health, I know, but I'm not alone with this," he told the Times. Evans and other proponents argue that "raw water" is healthier than regular water, which they see as tainted by chemicals like fluoride.

    I think it's time someone introduce them to the wonders of reverse osmosis and distillation.

  • nonesuch00nonesuch00 Posts: 17,926

    It's basically a bespoke trojan defender running in the kernel at all times to inspect HW SW instructions so I believe this speed hits of 17% - 30% is accurate regardless of computer activity.

  • dragotxdragotx Posts: 1,134
    Kerwin said:

    It's probably premature to predict how much impact will affect some processor intensive tasks like rendering.   For example, "... H.264 video encoding, timed Linux kernel compilation, and FFmpeg video conversion tasks didn’t lose anything."  (See https://www.pcworld.com/article/3245606/security/intel-x86-cpu-kernel-bug-faq-how-it-affects-pc-mac.html)

    Linus Torvalds wrote, "“Some loads will hardly be affected at all, if they just spend all their time in user space. And if you do a lot of small system calls, you might see double-digit slowdown..."

    Michael Larabel ran some games testing on early patched kernels, "Dota 2, Counter-Strike: Global Offensive, Deus Ex: Mankind Divided, Dawn of War III, F1 2017, and The Talos Principleon a Linux 4.15-rc6 machine with a Core i7-8700K and Radeon Vega 64. None saw a frame rate change outside the margin of error range."

    I think we'll need to wait a bit for the actual patches to roll out on Windows and MacOS before we can.

    I think most of the issue will be in I/O heavy tasks (think databases) which have to cross from user space to hardware drivers at very high rates.  We'll see.

     

    That's what I'm worried about honestly.  At my office we run a database that it's not unusual for it to pass 1 million IOs a second, and we are very heavily virtualized on the server side.  If those get the potential 30% performance hit, we are going to feel it in a bad way.

  • nonesuch00nonesuch00 Posts: 17,926

    So what generation of Intel CPUs is the 1st generation without this bug? It sounds like this was a known and well-kept secret that an Intel employee exposed.

  • dragotxdragotx Posts: 1,134

    So what generation of Intel CPUs is the 1st generation without this bug? It sounds like this was a known and well-kept secret that an Intel employee exposed.

    I haven't heard anything definite, but it sounds like it's essentally all of the i# line of processors, so i3, i5, i7 are the ones that seem to be the most likely.

  • AllenArtAllenArt Posts: 7,140
    edited January 2018

    "I'm extreme about health, I know, but I'm not alone with this," he told the Times. Evans and other proponents argue that "raw water" is healthier than regular water, which they see as tainted by chemicals like fluoride.

    I think it's time someone introduce them to the wonders of reverse osmosis and distillation.

    "Raw water"???! o.O Are they nuts? With all the crap one can catch nowdays? The very thought gives me sympathetic stomach cramps...

    I guess they don't care that their "raw water" might also be tained with "raw sewage"....maybe it's just a "raw" thing..... o.O

    Laurie

    Post edited by AllenArt on
  • nonesuch00nonesuch00 Posts: 17,926
    dragotx said:

    So what generation of Intel CPUs is the 1st generation without this bug? It sounds like this was a known and well-kept secret that an Intel employee exposed.

    I haven't heard anything definite, but it sounds like it's essentally all of the i# line of processors, so i3, i5, i7 are the ones that seem to be the most likely.

    Wow! That's a doozy!

  • AllenArtAllenArt Posts: 7,140
    edited January 2018

    So what generation of Intel CPUs is the 1st generation without this bug? It sounds like this was a known and well-kept secret that an Intel employee exposed.

    I dunno if I'm ready for this. All three of my computers are Intels and none were made before 2012 *sigh*

    Laurie

    Post edited by AllenArt on
  • ByrdieByrdie Posts: 1,775
    edited January 2018

    I think you'd have to go back to sometime around 2007 or thereabouts to find one without the bug. My Dell XPS which is just two years old surely has it, so bad news for me on the one hand. On the other, the new laptop I ordered during their Boxing Day sale has AMD chips. Guess I was lucky that even on sale the Intel laptops were out of my league budget-wise.

    Post edited by Byrdie on
  • nonesuch00nonesuch00 Posts: 17,926
    AllenArt said:

    "I'm extreme about health, I know, but I'm not alone with this," he told the Times. Evans and other proponents argue that "raw water" is healthier than regular water, which they see as tainted by chemicals like fluoride.

    I think it's time someone introduce them to the wonders of reverse osmosis and distillation.

    "Raw water"???! o.O Are they nuts? With all the crap one can catch nowdays? The very thought gives me sympathetic stomach cramps...

    I guess they don't care that their "raw water" might also be tained with "raw sewage"....maybe it's just a "raw" thing..... o.O

    Laurie

    Flouride is not so bad but chorine tastes so aweful I only make tea with tap water and use flouride infused mouthwash to make up to for it (I once very quickly found out how much flouride really does protect the teeth years ago when I moved from USA to Switzerland)

  • QuasarQuasar Posts: 574

    So this will be something that will affect the 8th gen i7 possessor in the new computer I ordered and am picking up today? That's not good news. *sigh*

  • AllenArtAllenArt Posts: 7,140
    AllenArt said:

    "I'm extreme about health, I know, but I'm not alone with this," he told the Times. Evans and other proponents argue that "raw water" is healthier than regular water, which they see as tainted by chemicals like fluoride.

    I think it's time someone introduce them to the wonders of reverse osmosis and distillation.

    "Raw water"???! o.O Are they nuts? With all the crap one can catch nowdays? The very thought gives me sympathetic stomach cramps...

    I guess they don't care that their "raw water" might also be tained with "raw sewage"....maybe it's just a "raw" thing..... o.O

    Laurie

    Flouride is not so bad but chorine tastes so aweful I only make tea with tap water and use flouride infused mouthwash to make up to for it (I once very quickly found out how much flouride really does protect the teeth years ago when I moved from USA to Switzerland)

    I only drink bottled spring water and we make the coffee with distilled. The only time I use tap water is for cooking and washing dishes. There's flouride in the toothpaste and moutwash. And while I will agree that chlorine smells and tastes awful, it's better than picking up a shigella germ or worse. Gah.

  • kyoto kidkyoto kid Posts: 40,561
    Ivy said:

    Sad news indeed. 

    I guess it looks like I'll be turning off windows updates notices on my i7 server machine.and update the hardware manually.  

    my win10 laptop is so slow i doubt i would notice a 30% decrease. But my ivy bridge intel i7 4.4.4 ghz  is in a server set up mode with external gpus and is smoking fast . 30% decrease would be definitely be a hard benchmark notice when rendering animation and would not be worth the update to me when i can pull the system off-line or put it behind my firewall and bump up restrictions and feed content to it through my LP .  either way this old system won't be learning any new processor tricks. 

    sad news indeed

    ....I turned off Windows update on my W7 system on Sept 30th, 2016 when MS switched from single individual update files (that you could "pick & choose from") to the bundled update format they now use (which forced you to accept all or none of the updates just like in W10).  Yeah, I render on the CPU in both 3DL and Iray which, considering the what I gather from the article, would mean longer rendering times due to the constant calls made to the CPU during the process, no thanks.

  • kyoto kidkyoto kid Posts: 40,561
    edited January 2018
    nicstt said:

    Makes me glad I bought a Threadripper

    ...I would have considered that except then I would have to downgrade to W10.

    Post edited by kyoto kid on
Sign In or Register to comment.