Me: At Jow Forums said Intel CPUs had a Backdoor due to the BIOS acting as a separate OS. Dude said that OpenBSD has known about this for years and disabled Intel threading. I googled Intel OpenBSD threading, and sure enough it's true.
Does Intel have a Backdoor or is this some Alex Jones conspiracy crap? If that OS OpenBSD has disabled Intel threading, which I googled but don't understand, than why that Theo dude disable it?
Other urls found in this thread:
raptorcs.com
en.wikipedia.org
en.wikipedia.org
en.wikipedia.org
youtube.com
youtube.com
youtube.com
twitter.com
Have you been living under a rock or did you just arrive from plebbit?
>install gentoo
>Does Intel have a Backdoor or is this some Alex Jones conspiracy crap?
Both. Yes, your intel chip has a backdoor meant for enterprise remote management. No, no one can prove it's compromised. Yes, you should assume that it is.
>your intel chip has a backdoor meant for enterprise remote management
>meant for enterprise remote management
Sure, buddy! Good goy.
How do you disable Intel CPU threading on Windows?
>Does Intel have a Backdoor
a) you can't know, since the firnware is closed, unless studies on it are made
b) multiple studies have indicated this for both Intel and AMD CPUs
so yeah, shit's fucked on a hardware level. how about that low-level server capability.
no, pol is full of shit as always
theo had some echo that intel speculative pipeline is buggy years ago, but there were no real info and fixes beside proper cache flushes on all syscalls
then meltdown and specter was released to public by google, since openbsd got no info beforehand, it took them too long to implement the fixes in comparison to other OSs
the restrictions put on the multi-threading was to never allow 2 cores running on separate privileges to share caches
Intel's hyperthreading is basically 2 core threads sharing the same cache so they can advertise twice as many cores despite it being pretty much useless beside very specialized number crunching applications
pretty much no one cared but theo who was convinced that this privilege difference happens pretty much on every syscall and saw no solution available thus disabled hyperthreading entirely
also you can't disable it properly, in the same way that a non-root user can't prevent root from doing something. it's deeper.
So, OpenBSD kills the rootkit in the BIOS? Wtf does hyper threading even mean brugh?
Fuck I should've bought that amd 2600 for $150 but I kept telling myself to wait for an i7 to go on sale.
jesus fuck, the information is one search away. how can you write so much and still be so misinformed?
Than that code was either leaked or that Theo De guy hacked it, how would he knows to disable Intel threading as a security risk.
BIOS has nothing to do with it, the bug is in CPU design and fix is in CPU firmware
t. Shlomo Goldstein
sorry for not larping
So... If I am running a large project in a program that utilises lots of threads, should I be getting Intel or AMD?
it's the same company, doesn't matter
I'm asking in relation to hyper threading and multi threading.
IBM Power9. It has double the threads per core than equivalent x86 processors.
Dude on pol said BIOS code loades in the Kernel and works like a rootkit? Is this crap true?
pol likes to use buzzwords without knowing shit
Arnt these servers? Or is there a way to get consumer level products from IBM, i.e - ones that can be installed in a typical desktop rig?
hyperthreading is Intel-only thing at it sucks balls
but AMD CPUs suck balls entirely
hard to decide
No, this isn't specifically related to OpenBSD. The hyperthreading vector is just a symptom.
AMD probably isn't much better, then again im pretty sure these backdoors are only used for catching criminals. You're not a criminal are you, user?
here's a trick I've learned that is 100% effective in telling if Jow Forums is talking out of their ass again: yes
Did you mean its the only thing Intel sucks balls at or only Intel does hyperthreading?
I am having latency issues when utilising many threads and with the whole processor war going on, its hard to determine which is best at the high end.
Every intel cpu runs a minix os inside their chips. The minix os grants level -3 access every hardware on your pc.
There's also a MicroATX mainboard that I don't see on the site. Supports only a 4 or 8 core chip though. It's the cheapest one.
>being this bluepilled about the world
oh my sweet summer child
Than why is hyper threading still disabled on that OS? Maybe dude is on to something that we don't know. If Jow Forums is right and the BIOS is loading a rootkit as dude said as a modual or hook than how would we even know?
>nothing to hide, nothing to fear goy :^)
Anything that "only works with Windows 10" and has motherboard manufacturers purposefully blocking Windows 7 has a backdoor. On both an OS and hardware level.
I'm assuming I cannot install Windows on these machines?
This is very interesting though.
Sorry to for asking for the spoon feed, but does IBM have anything that can contend with i9 or i9 XE series?
>shekel
Am I the only one who can't see this word as anything but *rubbeti rubb* happy merchant meme?
I don't even hate jews, sjws or marxists.
What is Libreboot?
not low enough
On every fucking chip? Man that's some BS if true.
Wtf you talking about? Are new computers only allow you to install Windows? How in the fuck is that even legal?
Yes
It has to do with UEFIs. Certain computers like the Microsoft Surface don't allow the end user to disable secure boot, which means you can't boot off of a flash drive, which means you can't install a new OS.
Google Intel management engine
Both Intel and AMD have them. Intel has ME, the management engine. AMD has PSP, the platform security processor. Both of them are ARM cores embedded inside the CPU.
That's what they're talking about when they talk about the "built in backdoor". They're closed source, but we at least know that the Intel ME uses Minix code for operating. It's unknown what AMD PSP uses.
The Hyper-threading exploits that OpenBSD went and disabled hypterthreading to protect against are something completely different and are part of the batch of exploits that were released last year.
>The Hyper-threading exploits that OpenBSD went and disabled hypterthreading to protect against are something completely different and are part of the batch of exploits that were released last year.
Spectre specifically.
The Intel management engine. An on chip computer that can't be controlled or accessed by the main CPU. It has full access to everything and runs even when the main computer is powered down.
Question if the BIOS can run and load hardware code, than why can't it load a Backdoor while the system is being booted?
Jews>indians
Intel wins again
And Meltdown.
en.wikipedia.org
And Foreshadow.
en.wikipedia.org
How does it run when the computer is powered down? Does it continue to run if I disconnected it from the power supply?
Are you telling me that a current i9 9970 cannot hyperthread?
If I am running a workstation machine that is never connected to the internet, is this vulnerability still an issue?
How big of a brainlet do you have to be to confuse IME backdoors with Spectre counter measures? The absoloot state of Jow Forums right here.
>How does it run when the computer is powered down? Does it continue to run if I disconnected it from the power supply?
Same way your NIC or always-on USB ports run when off. It will turn off if your machine is unplugged.
Yeah, the boot loader in my phone was locked, found an exploit on the web, unlocked the boot loader, rooted the phone an installed what ever rom I choose to use, why can't this be done on Windows?
It uses the bios battery. Also when you plug in your pc there's a hidden charger on motherboards now and a mic that listens to everything you do
Fuck off Jow Forums
Nobody reply to OP
>Are you telling me that a current i9 9970 cannot hyperthread?
If you're using OpenBSD, not unless you enable it. I'm not sure what Windows has done to mitigate the problem, but I wouldn't be surprised if they disabled or fixed it through microcode. Same thing Linux has done.
Stop being an ass, not everyone knows this stuff..
>Several procedures to help protect home computers and related devices from the vulnerability have been published.[73][74][75][76] Spectre patches have been reported to significantly slow down performance, especially on older computers; on the newer eighth-generation Core platforms, benchmark performance drops of 2–14 percent have been measured.[77][5][78][79] On January 18, 2018, unwanted reboots, even for newer Intel chips, due to Meltdown and Spectre patches, were reported.
>How does it run when the computer is powered down? Does it continue to run if I disconnected it from the power supply?
No, which is why I said powered down not unplugged. So long as your PSU is plugged in and connected to the motherboard the ME is running.
I know, but a lot of people are posting trying to inform those who don't know when they themselves don't know what they're talking about. It's fine if you don't know, but it's not fine if you act like you do.
So Windows running an i9 CAN hyperthread? An 18 core i9 will run a potential 36 threads?
Thank you, that is why I termed it that way to clarify as well. It can be hard infer through text.
>So Windows running an i9 CAN hyperthread? An 18 core i9 will run a potential 36 threads?
Yes. There have been microcode and compiler fixes for spectre, meltdown and foreshadow. Although current hardware takes a performance hit as a result because it is fundamentally a problem with the hardware architecture.
ZEN 2 and Ice Lake will have hardware changes to mitigate those three vulnerabilities.
No, you cannot run Windows. Only operating systems with support for Power9 (mostly Linux).
As to the second question, compete how? In speed, the top ones are equally as fast or faster as long as the software is optimized for the chips (lol). In scalability they're undoubtedly superior.
>
You Fuck Off Bolshevick!
I want the Blackbird system with the quad core from them. But I'm in the UK, so it'd cost about $1800 for the board, chip, cooler and 32GB of RAM. Or £1500 bongs. And then there would be the import taxes on top of that. There's VAT (which is 20%) and then customs tax (which fuck knows what that is...).
I wish they had a UK distributor.
>>Me: At Jow Forums [...]
>[...] did you just arrive from plebbit?
Yes, you should have stopped reading 3 words in.
Zero iq posts
>Random anonymous person on board flooded with misinformation says buzzwords that scare me
Pic related
Based and redpilled
For real time audio tracking and creation, a process which, apparently, applies one thread to each audio signal. The program I use, apparently, utilises the cores and threads in a serial fashion, even though its one thread per signal - so that if one signal uses more than the clock speed of the thread it will begin to use the power of the next available core.
This is my understanding, and I am well aware that I may be way off in my thought process.
If my understanding is correct, then I am often using over 50 "threads" - and considering I only have 8 cores and in turn 16 threads, I assume that the available processing per thread is being split amongst the signals that are not requiring entire cores to maintain themselves, which would explain latency issues and drop outs, right?
I guess "compete" is a stupid way to phrase anything. I mean how hard can it pushed in different tasks before experiencing performance drops, mainly in relation to running a task that requires more threads than is available - although I know this is a super specific line of questions.
If I have an 18 core i9 with a thread count of 36 - with over 50 signals each attempting to have their own thread, but being serially shared over the cores as they become available - am I going to get the same experience with latency and/or drop outs as I would with a machine with 16 threads, based on the premise that neither can deal with the 50 thread requests?
Sorry if I am absolutely rambling incorrectly, my understanding is pretty amateur.
Oh. It's a program you already use.
It's not going to run. The processors aren't binary compatible with intlel and amd.
Thank you.
The rabbit hole you lead me down was very informative.
Intel CPUs are so full of backdoors it's not even funny.
Just the ones we know of:
>Pozzed "random" number generator
>Actual "management" backdoor allowing full control of the system
>Tons of (((security bugs))) in out-of-order execution and speculative execution
There's a reason Chinese and Russians don't use this crap.
t. Dr Goldberg, Ph.D., inforation security analyst at Intel
>For real time audio tracking and creation, a process which, apparently, applies one thread to each audio signal. The program I use, apparently, utilises the cores and threads in a serial fashion, even though its one thread per signal - so that if one signal uses more than the clock speed of the thread it will begin to use the power of the next available core.
>This is my understanding, and I am well aware that I may be way off in my thought process.
>If my understanding is correct, then I am often using over 50 "threads" - and considering I only have 8 cores and in turn 16 threads, I assume that the available processing per thread is being split amongst the signals that are not requiring entire cores to maintain themselves, which would explain latency issues and drop outs, right?
Basically, any complex processing has operations that can be done in parallel and others that require being done serially.
In the case of audio, if you have separate sources (synthesizers, etc.) you can generally compute them in parallel but at the end, if you mix them and apply effects, this has to be done serially (and the effects have to be done serially, if you add echo to a signal that has some other effect, you cannot really compute the effects separately).
If your problems are only parallel, you can just use has many cores as you required and you're fine. That's how supercomputers work. They use regular Intel CPUs, there's just many of them.
But single-core performance is also important, precisely for the problems that have to run serially. Typically Intel is known to have the very best single-core performance (although with recent security issues we've seen that they basically "cheated" to achieve those).
You're a disgusting person.
K!kes and Pajeets stand side-by-side on shitting on this world.
He's disgusting but right.
>gee golly Bartholomew why is everyone else but me so uninformed on the intricacies of CPU-level vulnerabilities
go do something productive
the jews run this shit pussies, suck it
think of it like this
>intel security = swiss chease