Not sure what to make of this:
https://www.theregister.co.uk/2018/01/02/intel_cpu_design_flaw/
AMD Ryzen seems to be the winner if patches are selective and only affect Intel CPUs.
This was the update that was sent out on the 4th KB4056892. I use Windows Defender so that's why it installed.
I asked over on techspot how vulnerable we really are and this sums up what I really wanted to know..
Do you believe you will be targeted and infected?- Meltdown has low prevalence and if you have safe habits, update your OS regularly and use a decent security soft; chances you get compromised are very low.- Spectre is too complicated to setup, need physical access to your motherboard, you won't be hit by it.If you are not hurry you can wait a bit before purchasing any cpu and wait fixed versions. However since the issue, actual cpu's prices went down a lot. .It is your choice. " src="https://www.techspot.com/images/icons/new/prekesh-happy.png" alt="" />
What sums up what you really wanted to know, cos I'm not seeing anything?
Evidently, the full effect of the patches won't be known until the BIOS patch becomes available and is applied. And then I'm wondering if the I/O operations are going to be most affected, what does the combination of patches do to uploading and downloading, particularly with larger files? Things like skin uploads and downloads would probably be only minimally affected, but what about driver/firmware downloads from Nvidia, for example? Some of those can reach up to 140GB and more.
Somebody please correct me if I'm wrong, but as I understand this, it won't matter how fast your internet is, I/O operations, uploads and downloads will be slowed. Is that right? If so, then the impact of this may be greater than was first revealed.
I suspect network latency will still be orders of magnitude greater than the overhead of these patches on context switches (as the CPU swaps between user and kernel mode).
It not so much how much I/O happens but how many system calls a program makes to the OS that is a factor. Eg if some dumb program read/wrote (or sent/received) a large file one byte at-a-time then these patches would really hurt it's (already not very good) performance.
Ok it's the patch, not the vulnerability that will slow things down. Now as far as bits per second over the network that doesn't matter. Information will still transfer 100 gigabits per second for me. What will slow down is the computer itself not the network. As far as network the speed is whichever is the slowest computer involved. What we are talking is the computer if i'm understanding this do to the patch you processor will slow down 17-27%, and your ssd drive anywhere between 8-40 percent depending on what it is doing, and the brand of ssd drive.
That's what I've been trying to say, the performance hit, on even home user PCs , will be greater than first revealed by Intel or Microsoft. Yeah, we are told this is the way of the world, with hackers seemingly allowed to run rampant to do as they please, but does it really need to be this way? Surely these big tech companies have the resources and know-how to to combat these threats even before they emerge... like be proactive so that users can be 100% safe in using the products they spend big money on.
this is getting more and more screwed up.... still haven't got microcode/bios update for my motherboard (z170 based)... and now linux distros are reverting the microcode fixes (redhat, ubuntu, mint)
According to what I read, Intel is now saying not to install the patches is has released as they cause more frequent reboots. Intel is currently working with its partners to release updated patches since identifying the rebooting issue, so that may be why you've not yet received your microcode/BIOS update.
could well be... but more reboots is kinda meaningless? i could understand that to be a problem for data centres and servers... but home users? i'm assuming reboot means os prompting you to reboot (or increasing the frequency of scheduled reboots in them big boys) rather than kaboom and rebooting your machine all by itself
just goes to show how unorganised/rushed the whole thing is, regardless of whether the info embargo was prematurely breeched.
Whelp: Torvald's Take
Gulp: Intel's 1-22-18 advice
Intel: A major cock-up, so glad I went AMD for this new build.
I haven't built an Intel based computer for the past 20 years. AMD may not always be the best, but the are certainly far far cheaper and have nearly the same performance in most cases.
I followed a link on 'Tom's Hardware' and the article had various benchmarks which suggests that many of AMD's Ryzen series beat similarly rated Intel performers. Frankly, though, I don't much care about which manufacturer has the fastest CPUs. I'm happy with what I've got and am looking forward to firing up this AMD beastie once I have all the bits n' pieces to make it positively the best build I've ever done.
Hate to be "that guy". but...
"OH no the NSA's back door got found!" lol
Totally have no idea if that bug could be used for a backdoor. just thought it was ironic that a CPU level security flaw went "UN-noticed" for 10+ years? maybe they just didn't know....riiiiight
*takes out spoon.... stirs the pot"
Maybe Intel wants to make patches opt-in instead of opt-out. Intel likely doesn't want new CPUs tested with patches because they will look bad compared to old CPUs that were tested without patches.
The question is if Intel can do something with the current Core design. As for Spectre all CPU vendors need to redesign silicon to avoid performance penalty from software workarounds.
It will take years before we have CPUs which are safe without performance-sucking patches.
I'm torn between "I don't care, I want max performance from my hardware" and "it's better to be safe than sorry".
Right now I don't think there are practical exploits that one has to be very concerned about, but I don't really know. Browsers should be patched against Spectre.
This is just a headache, no idea to think too much about this - it's just a bad situation. I guess we could call it a lose-lose situation.
Hehe,... and efficiently patching said backdoor will just get the NSA bursting through the front door, jackboots 'n all. Oh wait a minute! What about the side doors? Probably for the C.I.A., right?
I do have to wonder, though, why it took 10 years to reveal these 'exploits'..... and was the Google employee tipped off about them or did he just get lucky? Like I said somewhere before, I smell fish.
i think chrome just got patched but you could have enabled a flag before.
and i think it's more like 20 years rather than 10? anything intel but itanium is vulnerable or some such. it's not just the google guy.. and they are building on someone else's research / speculation (last year or the year before i think. the kaiser stuff)
There are many great features available to you once you register, including:
Sign in or Create Account
