New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Comments
Funny that we rarely or never get to read some constructive article from you but can be almost sure that you will be there to criticize and to (seemingly) know better ...
And no, that is not a theory. Just look at diverse sorts of malware, viruses, etc. Virtually all of them go against x86-64.
Hey, don't demand from Mr. "I have nothing constructive on offer but will try to rip everything apart and to look like I knew better" to act like civilized people!
That's funny you say that specifically, because one of those criticisms is that you give your opinion instead of just pointing to a useful article. It's like you can't accept any of your own shortcomings and blame others because you made an attempt and you should get a gold star. Get a grip. Many of us get our tech news direct from news sources, we don't need your strange rants on things.
All I did was point out that you CAN'T know if they are currently under embargo. No argument to that point from you (so you didn't need to reply). You just saw this with Intel (they thought they solved it in hardware fixes), so that shows your bias. It's like you learned nothing from that.
You're the worst for attacking the person and not having constructive responses. That's hypocritical.
Long on hypothetical hysterics, short on evidence. By your logic, I would like to point out that we CAN'T know if you are currently mentally challenged, so we should talk without assuming you are normal.
Hey fishy, does JSG give you free reach arounds or are you just banging him for the status?
sniff sniff smells like someone is butthurt.
You're an academic? Seriously? Seriously? I'm literally laughing out loud. SMH stay in school.
Seriously, did you not understand how an embargo works? It's like you need to be smacked with the logic stick yourself. There's nothing difficult to grok that you need to call "hypothetical hysterics". What I said was possible literally just happened in jsg's Intel thread.
Anything is literally possible. You might as well have said water is wet.
You don't even know what you're arguing, then? Just didn't like jsg being contradicted? That's fine, just say that.
I know what I am talking about. You on the other hand have shown no evidence unlike @jsg. When you demonstrate ability to reason with evidence, we can discuss.
Classical @TimboJones reaction, one gets used to it. He is not about the matter at hand, he is all about "criticizing" and then self-defense by new attacks, incl. personal ones.
The information about Power9 having been fixed, even in hardware, can be found even in his own source. He completely ignores that however and instead chooses to "defend" himself by mounting new "but it could be!" and finally ad hominem attacks.
As for logic, not so much. Blabla about logic of course and lots of it but painful inconsisteny in discourse. Example: It seems that only the Power9 people could have embargoed ugly info. Funny that when considering that Power9 has a better track record while intel *was actually caught in the act" of abusing both an embargo and the fact that they sponsor universities which then seem to take a lot while bending over.
In fact this article is about intels new vulnerability and new attempt to keep it hush hush.
Another problem is that TJ lacks relevant knowledge of the field (which can not be gained by clicking on this or that article). Therefore he thinks "intel, ergo AMD too, and why not power9 too?" - but that is wrong.
Simple reason: Former attacks like e.g. Spectre relied on common mechanisms and would hence work on any HT processor or at least be easily portable to it. Zombieload however does not rely on a mechanism but on implementation (mainly of the LFBs) which is significantly different on different architectures and even processors. Hence to the best of my knowledge not even AMD Zen is vulnerable but only intel processors. Power9 can be considered immune against zombieload as it has even more significant both architectural and implementation differences.
Another factor to consider is target markets, because different markets react quite differently with quite different consequences. intels markets are wide spread and range from Joe and Jane to mid level professional. Those markets are tightly held (control decreasing now due to AMD) and can be quite easily manipulated because neither Joe and Jane nor most x86-64 customers have a deep technical understanding, plus they are used to exposures and exposures getting somehow magically fixed. Power9 on the other hand addresses a very different market, a market where playing dirty games and serving gross lies can lead to doom. It is therefore not at all surprising that the Power9 people take vulnerabilities very seriously. And they usually come up with actually working fixes quite early.
Finally note that I talk about an existing, working, and proven vulnerability while TJ's speculations are pretty much based on a magical crystal.
Funny side note: The majority of vulnerabilities of Power9 during the past couple of years were actually GCC, GNU, and linux implementation problems.
Do you guys hear a whooshing sound? That's the sound of you not getting it.
Poisson, you're self admitted not a technical person, you're just a cheerleader with nothing to add. You don't even know what is being argued. Hint, you asked for evidence when my only point was possibility. If they say they fixed all problems and then get news they do, they have to walk that statement back, looking really bad. And since there's been past exploits, it's reasonable for them to assume there will someday be more. That's the point about not throwing stones. Don't go crazy saying you don't have problems when they cannot be 100% certain they won't get their own exploit someday.
I said:
"How would you know if they are not under some embargo and just lying they are free from all known vulnerabilities?"
He wouldn't know if they are under an embargo. They are not public.
How does jsg know they don't think they have it fixed but don't actually? That just happened with Intel.
Seriously, just have discussions without getting all emotional about shit.
@TimboJones the one using emotional words are you, not me. But you are right in that I am emotionally happy snarking your Dunning-Kruger syndrome. I am not a technical expert, but what @jsg said makes a lot more sense than you frothing at the mouth with your tin foil hat on.
Stay in school, man. Jsg said, "Plus Power9 allows you to claim in big fat letters "free of all the known vulnerabilities!""
You want to say they can 100% claim that exact phrase? It's a good idea to market in big letters JUST after having a bunch of exploits that required silicon to fix? And it won't blow up in their faces if/when another exploit is found?
You're the king of irony. For one thing, this isn't even a complex discussion and it's well over your head.
If you don't know the discussion, and you insert yourself and continue to not know your head from your ass, you should rethink pulling the Dunning-Kruger effect.
Well, reality shows that quite a few companies said things like that even when they knew that it was a lie.
And btw, there is already yet another intel vulnerability. You are not on exactly solid ground.
This thread and this discussion isn't about what TimboJones thinks of this or that person - although you seem to see and treat many threads like that.
Let's discuss matters, not persons.
What does that even mean? I don't own Intel stock, I haven't bought new intel CPU'S in at least 5 years, and I never ever said they wouldn't have more vulnerabilities. So I'm not sure wtf you're talking about. Continually finding vulnerabilities was my basis of why you don't go around touting your security don't stink. Hello, McFly?!?
Did you mean to address that last part to poisson for being off-topic and not talking about the topic, which I fucking did? Because then that would make sense and not be ironic.
using Intel is like hosting at ColoCrossing now
Don't get emotional. That's not quite @manish
The emotion you're referring to is confusion; confused by what jsg is going on about, and confused by how you don't seem to grok anything but unaware of it.
@TimboJones
"You're the king of irony", "Stay in school" are just two examples for what seems to be a mix of ad hominem and seeming to feel superior. And in fact in another thread you even clearly stated that you are or feel superior to @poisson.
My point isn't to disect phrases or to attack you. My point is based on the fact that I saw communities rot away and die more than once because discussions were not anymore on matters but increasingly on other users.
I think, all of us, including you and myself should be trying to be and act as civilized as we can, which btw also includes to not belittle others or feel or paint oneself as superior.
And to just let go of something is sometimes a more civilized and smarter choice than to always fight to the end.
Keep in mind that I wrote this OP with good intentions and to inform others in our community. I don't own stocks or work for intel, AMD, or any closely related entity and I have no general bias towards or against either one nor am I blindly an IBM fanboy although I worked for them a loooong time ago.
And btw THAT, how much someone contributes to our community or at least honestly try to do so, is how we should judge other users.
Time to nut up or shut up.
Austrian universities are financed by state and federal gov. Any funding above 25% from external is not legal and barely any - including medicine - is taking more than a few % sponsorships.
TSMC has, AMD has nothing. Samsung also fabs in 10nm. Samsung and TSMC 7nm are not the same as Intels 10nm which is more like 8nm.
Nothing, the PCIe 4.0 is just there to provide double interconnect speed between 2 CPUs on EPYC and the chipset - That external lanes support it is just by design; why downgrade functionality. This was not designed to benefit the customers, same as more lanes are just a side effect.
No, it is a costly failure to use IBM. The prices are insane. Power9 also has a veeery deep (= inefficient) instruction pipeline and not much cache plus barely any out of order execution (which is not solved by more cores.).
I don't think that the researchers lied. They quite openly said that intels "sponsoring" was a (probably even the) decisive factor in sticking to the embargo although intel had broken the agreement.
Yes and no. Yes, TSMC has the production but it's not like AMD only had funny ideas and not a lot more. AMD does have the 7 nm technology and does a very significant part of the chain between basic design an tapeout. AMD could change their fab with reasonable effort and costs just like they could change the process, e.g. from 7 nm to 5 nm or from TSMC's 7 nm to Samsungs.
I don't agree. For a start PCIe is just a serdes protocol that also lends itself well to connecting CPUs and chipsets/bridges.
And no, more lanes are certainly not just a side effect, they are complex beasts and not that easy to put into a chip in numbers, plus for many customers the number and capacity of PCIe lanes was/is a significant bottleneck. Having plenty means that lots of USB 3.x, NVMes, 10 Gb/s and even multiple 40 Gb/s ports aren't a problem.
Btw, intel has already announced that some new processors will support PCIe 5.
The (not at all) good old flame war with company A's product declared all sh_tty vs company B's fine products.
Plus, it's BS. I'm talking about current Power9 with pretty much the same cache cap. as high end intel and AMD processors. Yes its instruction pipeline is longer than other (but considerably shorter than that of its predecessor) but almost everything in computing is about trade-offs and IBM has the results to show that it's decision was a good one. Plus the architectures are very different; what's good for a CISC processor can be bad for a RISC processor and vice versa.