Return to

Intel FUBAR ... again - Kernel memory leak in nearly every Intel CPU of the last decade (Spectre hits everyone, Meltdown still Intel exclusive)



Either can’t be bothered, or paid intel sponsorship. I’m not even kidding.

I try to call it out. If AMD (or ARM/Apple/Commodore64/etc.) may be impacted, sure - say they MAY be impacted. Don’t imply otherwise (i.e., “suspect” they are impacted) without any facts or architectural discussion (that AMD or someone more knowledgable could directly respond to or validate) to back it up.

If anything, i’d give AMD the benefit of the doubt, because thus far they’ve shown to be slightly less retarded than intel regarding the performance vs security stuff as demonstrated by meltdown.


I try not to go conspiracy routes, normally things are just dumb f*cking luck / coincidence.

What I do know is that this suspicion will be harming Amd adoption in data center and it works to intels advantage as it sand bags the opponent.

If there was no ‘suspicion’ then amd would be the cure, but with the suspicion there… people are going to think that swapping platforms doesnt really matter from a security standpoint.

That is why all these articles that use the suspicion word, with no evidence ever presented to back up that suspicion… annoy the f*ck out of me.


AMD is always called out. Not sure why. So far I see this.


Periodic reminder that this is a newly discovered class of attacks that have not been considered and new specific exploits of similar general variety will keep coming out because CPUs are staggeringly complex. Don’t expect the end to be in sight.


So seems like intel CPU’s with Hyperthreading is taking a big hit in performance with the 4.20 kernal due to new mitigation. Threadripper looking even better now.


20+ years of intel sitting on arse with security paying off.


If only Threadripper will solve the IOMMU situation with the chipset. I want to be able to use separate NICs for each VM.


If only Threadripper would be more flexible when it comes to choice of o/s instead of catering to Win 10 the way it does.


Be more specific? Performance gains in the 2990WX are up to 2x windows performance when running Linux.


I have what some might call an old PC. I can run Win 7, Win 8, even Win XP on it if I wanted to and all 64 bit versions of these at that. Yes, I also run Linux on it and I find that old triple boot comes in handy from time-to-time. I like my options but I don’t like Windows 10. I could run that abomination on this build too but for me, that would be like putting Wallmart budget tires on an old Ferrari. Ain’t gonna happen. So why would I put 'em on a brand new Porsche? Windows 10 or Linux. Those are my choices? Surely AMD can do better than this. A work station should include the ability to use optional operating systems of all kinds, in my opinion.


Threadripper can run Hackintosh with a Clover patch.

And my way of dealing with Windows 10 is to sandbox it’s networking by running it in a VM and a pfsense VM at the same time, with only the pfsense VM having access to the physical NIC. There’s no use editing the hosts file, there’s hosts definitions hard-coded into the kernel. Using pfsense to block connections is pretty much the only assured way.


Sure because pumping money into a soon-to-be-EOL OS that doesn’t even get real support from its developers (which BTW need to patch the kernel in the first place, and not AMD) is a smart business choice.

CPUs like this didn’t even really exist in the enterprise when 7 was current, nevermind consumer space where 7 was targeted.

It’s not AMDs job to make every OS to work on it. Sure they will provide support if they are asked for it, but they can hardly contribute kernel code to a closed OS, can they…



@CHESSTUR, hardware doesn’t support software. It’s the other way around.
Also OT…