[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/g/ - Technology

View post   

[ Toggle deleted replies ]
File: 156 KB, 633x758, 1501439883169.jpg [View same] [iqdb] [saucenao] [google] [report]
64136736 No.64136736 [Reply] [Original] [archived.moe] [rbt]

>AMD is affected by 2/3 exploits
No no no no

>> No.64136758



>> No.64136762
File: 212 KB, 793x307, intelplsgo.png [View same] [iqdb] [saucenao] [google] [report]


>telling lies on the internet

>> No.64136782


i'll stick with the big kid brand. by the way, how much is amd's stock worth?

>> No.64136783
File: 18 KB, 267x297, 1514912174690.jpg [View same] [iqdb] [saucenao] [google] [report]

t. bryan

>> No.64136817

The two of which one has a trivial fix, and the other is unfixable but unlikely

>> No.64136838
File: 164 KB, 568x612, 1502747129350.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64136839

AMD isn't affected by Spectre 2, only 1 which already got patched.

>> No.64136846

>kid brand
Gtfo underage

>> No.64136854

Wait, what? I thought there were only two (Spectre and Meltdown). What's the third?

>> No.64136869

There are apparantly 2 versions of spectre

>> No.64136870

Spectre encompasses two exploits, the first two of 3 in the tables of 3 exploits you see. The last one is Meltdown.

>> No.64136874

spectre has 2 variants

>> No.64136878


More than yesterday, unlike Intel.


>> No.64136890
File: 178 KB, 882x758, 1491072951812.png [View same] [iqdb] [saucenao] [google] [report]


>> No.64136909

Intel is still worth more. Sad

>> No.64136921

>intel is literally in meltdown

>> No.64136922

I see, thanks. Is my understanding below correct?

Only affects Intel. Can't be fixed with a microcode update, requires software fixes that can decrease real-world performance by 30%
>Spectre 1 & 2
Affects all x86 CPUs, but can be fixed in software (or was it microcode?) with no performance hit.

>> No.64136927
File: 199 KB, 1500x844, 1515010258716.jpg [View same] [iqdb] [saucenao] [google] [report]


t. fake news

>> No.64136936

>Affects all x86 CPUs, but can be fixed in software (or was it microcode?) with no performance hit.
Correction, affects all RISC CPUs

>> No.64136946

spectre supposedly can't be fixed but is harder to exploit

>> No.64136949

So it affects all SPARC, ARM cpus as well as x86 because they're technically RISC under the hood?

>> No.64136972


>> No.64136977

One variant of spectre is easily fixed at no significant performance penalty. The second variant is unfixable.

>> No.64136990
File: 44 KB, 1207x303, .png [View same] [iqdb] [saucenao] [google] [report]

really makes you rub your hands

>> No.64137008

x86 is cisc tho

>> No.64137017

So what about RISC specifically makes this vulnerability work?

>> No.64137119

>big kid brand
So that makes AMD the adult brand then? Fine by me.

>> No.64137187
File: 17 KB, 200x200, 1503718186931.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64137197

Nothing. The real problem is out of order execution which provides massive speedups at the cost of security.

>> No.64137406

>>Affects all x86 CPUs, but can be fixed in software (or was it microcode?) with no performance hit.
>Correction, affects all RISC CPUs

But x86 isn't RISC, it's CISC.

>> No.64137440


also arm has admitted select arm processors in their lineup are vulnerable to meltdown

>> No.64137474

not how stocks work, retard

>> No.64137490

pure CISC CPUs do not exist, x86 has both RISC and CISC component in them.

>> No.64137562

>It’s worth pointing out that not only does Windows have KPTI/KVA shadowing enabled for AMD processors as well, it even has specialized shadow system call entry stubs for AMD vs Intel. This either suggests they know how to embargo properly or that Tom’s PR is not entirely accurate

>> No.64137642

Because I'm tired of seeing people not know what the fuck is going on, here's a quick explanation of the information available so far:

- Spectre
A class of exploits that attack any modern processor's branching execution pipeline. Modern CPUs can execute multiple "paths" of instructions simultaneously, when the right path depends on the result of a calculation that haven't yet completed. The CPU then discards the results it doesn't need and keeps the ones it does. This general idea is present in every modern processor design, and without it even our highest clocked CPUs would have the performance of some 90s-era piece of shit. But, the machinery that makes it possible is very intricate, and these attacks are meant to twist the way the CPU operates under the hood in ways the designers didn't expect. To completely eliminate this class of attacks will require a redesign of every architecture, to figure out how to prevent side-channel information disclosure from speculative branches. This will likely take years. There is no other workaround. It's possible that these techniques will be used to craft exploits in the wild, but the level of difficulty is probably very high.

- Meltdown
A specific attack that allows non-readable addresses to be read. It works by speculatively reading the memory and then quickly side-channeling it out before an exception occurs. This is easy to exploit (HUGE information leaks, reading disk encryption keys, keyboard input, etc), trivial to reproduce under most conditions, and definitely affects Intel processors. It potentially could affect AMD, but AMD claims that their microarchitecture doesn't allow even speculative reads that violate page protections in this manner. If this is true, it completely kills the vulnerability with no workaround required. The current workaround (KPTI/KAISER) works by unmapping kernel memory from the user address space. This is expensive, and workloads that need to call into the kernel frequently will suffer a lot.

>> No.64137643

>Microsoft forcing the patch on the AMD even though it's not needed


>> No.64137654

paid by the jews, of course.

>> No.64137797

Now, we wouldn't want AMD to get ahead, would we? L-let's keep the competition fair.

>> No.64138129
File: 1.63 MB, 1415x2326, 2018-01-03-15-18-07.png [View same] [iqdb] [saucenao] [google] [report]

>> No.64138135
File: 14 KB, 457x101, IMG_20180104_025011_419.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.64138170

I know a guy who works at intel. One of the higher ups, works out of santa clara... He's in his 40's, plays alot of a golf and is weird as fuck

according to him, shit is going down.. be ready for shit to hit the fan

>> No.64138234

implying Intel pajeet know how to read that

>> No.64138243

and my dad works for Nintendo

>> No.64138254

more details dawg

>> No.64138276
File: 162 KB, 633x900, 1515004274625.png [View same] [iqdb] [saucenao] [google] [report]

tfw not belonging to X86_VENDOR_AMD club...

>> No.64138332

Ty m8

>> No.64138843
File: 234 KB, 1200x784, 1200px-Pn-pravez-class-5.jpg [View same] [iqdb] [saucenao] [google] [report]

Jesus, is there anything that isn't at risk?

That old Soviet ternary computer is looking mighty attractive right now.

>> No.64138878
File: 8 KB, 481x442, 1513004891768.png [View same] [iqdb] [saucenao] [google] [report]

this but unironically

>> No.64139169
File: 74 KB, 1600x1240, maxresdefault.jpg [View same] [iqdb] [saucenao] [google] [report]

wait so my computer wont get any faster?

>> No.64139218


>> No.64139222
File: 65 KB, 726x781, 1515048929464.png [View same] [iqdb] [saucenao] [google] [report]

>> No.64140216

>tfw still have a Phenom X4 955 BE
I'm fucked

>> No.64140236

>not yelling YOLO and adding nopti to boot options.

>> No.64140356


imho you are going to see a bear dance like a ballerina. The massive company Intel is going to redesign their chips in record time. What this means in the short term to the profits/losses of intc and amd I don't think anybody knows.

>> No.64140362

Intel is the "big kid" brand, with "kid" being the verb and "big" being the adverb.

>> No.64140371

You do know why they have to use the term "near-zero", right? RIGHT?!?! Like there's a near-zero change the sun won't rise tomorrow.

>> No.64140460

>Vulnerability to Variant 2 has not been demonstrated on AMD processors to date

>> No.64140504

Nah, unless you using custom Linux kernal + custom bios setting that same as test. you be fine

>> No.64140667

>AMD claims

And nobody can prove otherwise, unlike the increasingly trivial Intel attack.

>> No.64140757

>intel implying that a process accessing it's own memory is on-par with a process access all kernel memory on demand

>> No.64140845

>attack any modern processor's branching execution pipeline
Stop spreading lies, zen is unaffected. Older amd architectures are subject to the attack but software fix is there.
>It potentially could affect AMD
No fucking way, stop spreading lies. You will get an access violation on AMD.

Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.