Intel: You don't need to disable Hyper-Threading to protect against the ZombieLoad CPU exploit
pcworld.com
Intel: You don't need to disable Hyper-Threading to protect against the ZombieLoad CPU exploit
James Adams
Other urls found in this thread:
techrepublic.com
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
kernel.org
twitter.com
Anthony Morris
>Intel says
John Moore
>The bad news? None of what we tell you is going to make you feel any better.
>Intel agreed with the exploit’s capabilities but downplayed the level of risk ZombieLoad imposed. Intel also decided to name the exploit Microarchitectural Data Sampling, or MDS. That’s a lot less scary-sounding.
>Chipping away at Hyper-Threading or turning it off completely would be a huge blow to the performance of Intel's processors.
>On the same page, Intel demonstrates a nothing-to-see-here attitude if HT is turned off.
>Our issue with Intel’s testing is that it doesn’t use particularly multi-threaded workloads. If Intel’s tests used Blender or Cinebench or other multi-core CPU tests, you’d see an immediate and massive drop in performance.
Has intlel finally run out of shill money?
Nathan Parker
And yet the actual researchers who discovered it, and who they've bribed, do. Really makes you think.
Isaac Morgan
>amd doesnt have a backdoor
are u this stupid
Easton Price
Damage control once again. Fuck Intel.
Brody Bennett
Intel shill BTFO
Brandon Moore
>Intel is not recommending that Intel HT be disabled, and it’s important to understand that doing so does not alone provide protection against MDS,” Intel said in a statement.
lmao. Even disabling hyperthreading still leaves you at risk. They're basically saying, you might as well leave hyperthreading enabled since NO FUCKING MATTER WHAT, you're absolutely fucked. Might as well be fucked while getting those extra fps.
Aiden Ramirez
>Chipping away at Hyper-Threading or turning it off completely would be a huge blow to the performance of Intel's processors. You wouldn’t believe it from some of the documentation Intel has released, however.
>Our issue with Intel’s testing is that it doesn’t use particularly multi-threaded workloads. If Intel’s tests used Blender or Cinebench or other multi-core CPU tests, you’d see an immediate and massive drop in performance.
lol. Intel damage control is real!!!!!!!!!!!!
Ryan Davis
This the the real problem: only when you disable hyperthreading can you have full secure mitigation against MDS variant attacks.
Google is fully aware of this and is why they have completely disabled hyperthreading in their linux based OS (ie read: more secure than wankblows).
techrepublic.com
The REAL kicker here is intel was fully aware of MDS for years which is why they have been releasing "i3" and "i7" processors without hyperthreading. I'm not shitting you, see for yourself:
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
cpu-world.com
Jacob Sanchez
>IntKEK says
COPE
Aaron Campbell
OH NO NO NO N-
AHAHAHAHAHAHAHAH
Logan Sanchez
Disable hyper-threading? Who said anything about disable cores? Why would you underclock
Nicholas Roberts
>says the company that tried to bribe people so this flaw doesn't even get publicised in the first place
Sure thing shill. Sure thing.
Angel Rogers
Oh shit.
Joseph Murphy
>AMDrone pretends that the PSP doesn't exist
(lol
Camden Watson
see
Cooper Gutierrez
JUST
Kayden Williams
Leaving hyperthreading on just makes the exploit run faster. Might as well make the death quick and swift.
Nathaniel Price
cant they just fucking fix these vulnerabilities as they release newer processors? why did they leave it unfixed, did they think people would never realize?
Blake Miller
This is satisfying, where do i find more videos like this
Hudson Cox
>intel chips full of vulnerabilities
>AMD chips are terrible at anything that's not encoding, compiling, or data center-related
The worst timeline has been reached.
Austin Thompson
Are you sure about that?
Jonathan Davis
Their "vulnerabilities" are the only reason their processors used to be so fast. They were trying to milk money out of consumers by releasing shitty insecure but fast processors and knew as soon as the goyim found out they'd be screwed.
Benjamin Sullivan
>believing intel in 2019
Thomas Taylor
what the fuck was that?
Brayden Sullivan
And I'm not gonna spam cherry picked benchmarks just to ride AMD fanboys but to say AMD processors have no place in gaming is a bit of a huge stretch.
I mean sure a 5GHz intel chip has a ~10% fps advantage over AMD at max 1080p gaming settings even when using very high graphics cards like the GTX 1080ti but at 1440p it literally doesn't matter if you use a 5GHz intel chip or a 4 GHz AMD one.
On that subject: how long are we to peddle the prehistoric resolution of 1080p? Are we still going to be clammoring about 1080p when 8K 60fps+ gaming graphics cards come out?
Christian Price
1080p is still the standard so we'll probably keep seeing 720p/1080p benchmarks because it's more elegant than running "low/off settings" benchmarks to reveal bottlenecks.
William Williams
it's how amerimutts are born
Andrew Price
Not him but what's the point in doing that when it doesn't reflect real life usage? Either:
A.) pc gaymur is a poorfag who got an i3/R3 and a $120 Rx 570 that can handle medium-high 1080p gaming.
or
B.) pc gaymur has a high end 1440p graphics card (ie GTX 1080 or better) meant to play on max settings and still achieve 60+ FPS + i7/R7.
In both scenarios, neither the GPU or CPU is the bottleneck. Also 1440 75Hz gaming monitors are becoming cheap as dirt for the high end gaymur. Are you stipulating they would cripple their 1337 build with a thirft store 1080p monitor they scored for $20?
Anthony Adams
this is the copening of our lifetime.
>you don't have to...
yes, you do
>you don' loos..
yes, you do
>who cares about security anyways, are you pedo?
Ayden Hill
Mercury in solution amalgamating with aluminium.
The solution is there to strip away the oxide layer.
Anthony White
Get raped and kill yourself, you retarded fucking faggot sack of nigger shit with down syndrome.
Jacob Lewis
Want me to believe them? Lmao I disabled my HT
Hudson Peterson
now post the RPCS3 benchmarks
Connor Morales
L1TF was enough to force people to disable hyperthreading.
Lucas Gomez
This depicts the process of Al and Hg amalgamation.
Gabriel Wright
SEETHING
Aiden Powell
cope
Luke Smith
>the crocodile says you can put your hand in its mouth because it isn't going to bite you
imagine being this retarded, I hope OP is falseflagging
Lucas Stewart
>this thread is brought to by Intel™ Technology™ PR-Department
record: corrected
thanks alot, Chaim!
Nathan Brooks
Colton Anderson
hit a nerve, i see. need a safespace, little 56er?
Nicholas Butler
>If Intel’s tests used Blender or Cinebench or other multi-core CPU tests, you’d see an immediate and massive drop in performance.
Have games or Blender/Cinebench benchmarks with the migation applied been released yet?
Juan Cox
based angry retardomutt
Luke Jones
kernel.org
>All MDS variants except MSBDS can be attacked cross Hyper-Threads. That means on CPUs which are affected by MFBDS or MLPDS it is necessary to disable SMT for full protection. These are most of the affected CPUs; the exception is XEON PHI
Connor Martinez
shit nvidiot HOMOSEXUAL detected
Ian Reyes
>talking about cpu
>jokes on intel
>nvidiot shill
dude, I'm a drone too, but ffs, read the topic of the thread at least
Jose Walker
ok mr. nvidiot HOMOSEXUAL, intel doesn't exist
Joseph Robinson
Intel got the report for MDS around Q3 2018. They knew, for a long ass time.
Samuel Turner
are you trying to bait or are you mentally challenged?
Joshua Robinson
>posts benchmarks from a game that's highly dependant on cpu
>hurr, y u no post benchmarks with gimped gpus. Everyone knows that if there is a gpu bottleneck, you can see the powa of the cpu
Lincoln Collins
> But Intel said that’s not necessarily the only answer for all PC users. In fact, Intel said that it’s really up to each customer to decide what to do. If software cannot be guaranteed to be trusted then yes, maybe you'll want to disable Hyper-Threading. If your software only comes from the Microsoft Store or your IT department, you could probably leave Hyper-Threading on.
Noah Diaz
>probably
That's really fucking reassuring, inhell.
Carter Turner
hey guys, intel is empowring their use base. They're letting them decide for themselves.
if they wanna get fucked or cucked.
truly the best company out there, letting people have agency over their own processors.
Jacob Barnes
The article doesn't explain it properly, but it is true that you don't need to disable HT entirely. As long as attackers and victims can't run on the same physical core in parallel, you're okay. I don't know why these "news" outlets keep parroting that hyperthreading is now completely useless. Of course, this requires significant changes to how OS and hypervisor schedulers work.
That's also not true. You need multiple mitigations in place to protect against ZombieLoad and RIDL. You need to protect against the hyperthreading side channel, and you need to protect against the context switching side channel.
Wyatt Moore
Topkek
Josiah Scott
I DON'T KNOW WHAT TO BELIEVE