AMD Can't Shake Off US Class-Action Lawsuit over Bulldozer "8-Core" Advertising

Normally "car analogy" is a euphemism for "completely irrelevant anecdote”.

And I’ve never seen so many outrageously poor car analogies as in this thread. Sometimes they’re at least slightly relevant.


Here’s my attempt! No I lied, I don’t think there is a car analogy that fits this situation at all.
 
And I’ve never seen so many outrageously poor car analogies as in this thread. Sometimes they’re at least slightly relevant.


Here’s my attempt! No I lied, I don’t think there is a car analogy that fits this situation at all.

yes there is

a 4 door truck has 4 independent doors and extended cab truck has 4 doors but the rear doors need the 2 doors in front open first to open.

ergo an extended cab is not a 4 door truck because the doors are not independent and can't open in their own.

that's the closest think we'll get.
 
How is it not 8 cores though? There are 8 integer units.

The more accurate automotive example (though still inaccurate) is like getting a Ford 4.6 V8 and then wondering why it doesn't perform as well as a Chevy 6.2 V8 because all V8s are equal, right?



That's simply lazy programming in Windows, and should not be used as proof that it isn't 8 "real" cores. Windows only did that to prevent overloading the front end. The hyperthreading design is having two front ends feed a core. Entirely different beasts.

you are assuming that i am saying something that i am not saying
i never said it does or does not qualify as an 8core proc. i only said there was a more apt analogy than being so focused on the performance aspect alone and that this may be one of the times where the class action suit is valid because you're not allowed to just make up marketing bullshit and call it fact. im not qualified to say it is or isnt, but the whole thing does kinda sound like a bunch of smoke and mirrors.
 
you are assuming that i am saying something that i am not saying
i never said it does or does not qualify as an 8core proc. i only said there was a more apt analogy than being so focused on the performance aspect alone and that this may be one of the times where the class action suit is valid because you're not allowed to just make up marketing bullshit and call it fact. im not qualified to say it is or isnt, but the whole thing does kinda sound like a bunch of smoke and mirrors.

What part of the marketing lied? It had 8 cores. In certain workloads, it could match Westmere, and that's exactly what they showed if I recall correctly.

Marketing is always about painting your product in the best possible light while not crossing the line of outright lying. I didn't see AMD outright lying with Bulldozer.
 
What part of the marketing lied? It had 8 cores. In certain workloads, it could match Westmere, and that's exactly what they showed if I recall correctly.

Marketing is always about painting your product in the best possible light while not crossing the line of outright lying. I didn't see AMD outright lying with Bulldozer.

sigh.
have a good evening sir.
 
Looks up judge Haywood Gilliam.

Education, Bachelor of Arts and Juris Doctor degrees.

Seems like the right man for the job...
 
  • Like
Reactions: N4CR
like this
sigh.
have a good evening sir.

If you don't get it yet, the point is this case should have no legal merit whatsoever. There was no deceptive marketing other than the usual. Anyone that says otherwise are either complete computer nincompoops or AMD/Bulldozer haters.

Edit: Well, maybe complete nincompoops is a little harsh, but as others have stated in this thread numerous times, there is no official definition of what a core is, and whatever definition you want to come up with, there's almost always an exception except for the most basic of definitions, which Bulldozer easily fits as an 8 core.
 
I agree with the person you quoted. So, what very specific parameters define what a core is then?

If you don't get it yet, the point is this case should have no legal merit whatsoever. There was no deceptive marketing other than the usual. Anyone that says otherwise are either complete computer nincompoops or AMD/Bulldozer haters.

Edit: Well, maybe complete nincompoops is a little harsh, but as others have stated in this thread numerous times, there is no official definition of what a core is, and whatever definition you want to come up with, there's almost always an exception except for the most basic of definitions, which Bulldozer easily fits as an 8 core.

Perhaps after litigation there will be a more precise definition. The term, I'd argue, is too important to be left open to interpretation.
 
If this lawsuit holds ground what's it say about pentium's and celeron's cores it's not like those aren't **** as well. Better yet what about some of the early Atom cores? They are worse than Bulldozer probably core for core.
 
Perhaps after litigation there will be a more precise definition. The term, I'd argue, is too important to be left open to interpretation.

Counterpoint: Why does the definition of a core matter when the only thing that matters is its performance in the workloads it is designed for?
 
Perhaps after litigation there will be a more precise definition. The term, I'd argue, is too important to be left open to interpretation.

It's not important at all. By trying to define it as anything other than a loose group of execution units, you give merit to the idea that more cores equals better performance. But I'll take a dual-core I3 over an octa-core Samsung processor for desktop use.
 
Counterpoint: Why does the definition of a core matter when the only thing that matters is its performance in the workloads it is designed for?

It's not important at all. By trying to define it as anything other than a loose group of execution units, you give merit to the idea that more cores equals better performance. But I'll take a dual-core I3 over an octa-core Samsung processor for desktop use.

Any processor advertisement made to the general public will have Series, Model, Clock, and Core count.

Series and model are fabricated. i3 - means nothing. i7. Ryzen. Phenom...whatever. You could substitute literally ANY name and model. Clock and core count (physical and logical) are much more quantifiable. Imperfect, yes, but it is what we have. Want to buy a 12-core 8700k? Oh, sure I'm just counting all processing units, physical and logical, but hey - why not? That's exactly the type of thing I'd like to avoid. We'd have 128-core EPYC's running around and such.

It's not that I'm saying cores = performance. No one on [H] would say that without some qualifications. Probably a lot of them. So, if core count means SOMETHING, and I maintain that it does, then it needs to be understood what does and does not count. I'm not even making a statement as to what AMD did here is right or wrong, but rather, I see it as a path towards unification and clarity of *very* commonly used metrics.

Obligatory car analogy:

I just want to make sure the MPG calculation of one manufacturer is the same as the others. Otherwise, it is a truly useless metric.
 
I think the only way this case has any merit is if AMD wasn't clear about what they were counting as a core and from what I recall that was talked about at launch almost as much as the poor performance.
 
Any processor advertisement made to the general public will have Series, Model, Clock, and Core count.

Series and model are fabricated. i3 - means nothing. i7. Ryzen. Phenom...whatever. You could substitute literally ANY name and model. Clock and core count (physical and logical) are much more quantifiable. Imperfect, yes, but it is what we have. Want to buy a 12-core 8700k? Oh, sure I'm just counting all processing units, physical and logical, but hey - why not? That's exactly the type of thing I'd like to avoid. We'd have 128-core EPYC's running around and such.

It's not that I'm saying cores = performance. No one on [H] would say that without some qualifications. Probably a lot of them. So, if core count means SOMETHING, and I maintain that it does, then it needs to be understood what does and does not count. I'm not even making a statement as to what AMD did here is right or wrong, but rather, I see it as a path towards unification and clarity of *very* commonly used metrics.

Obligatory car analogy:

I just want to make sure the MPG calculation of one manufacturer is the same as the others. Otherwise, it is a truly useless metric.

You are literally asking for the impossible because there's absolutely no flipping way every architecture's core will be created equally, and you wouldn't want it to be either. Also, that car analogy is beyond horrible, and is in no way representative of the situation.
 
The FX series had 8 actual cores. Yes, they were grouped into units of two, and shared an FPU (which could be split in two parts when needed) but they had two actual integer cores.

If this lawsuit doesn't get defeated it's just another in the long line of lawsuits decided on by people who don't have a clue.
 
The FX series had 8 actual cores. Yes, they were grouped into units of two, and shared an FPU (which could be split in two parts when needed) but they had two actual integer cores.

If this lawsuit doesn't get defeated it's just another in the long line of lawsuits decided on by people who don't have a clue.

A lawsuit getting defeated on the merits is one thing. Getting thrown out before any presentation of facts is another.
 
who the f purchased a bulldozer in 2015 at a premimum?

the plaintiffs ...

"The plaintiffs, who sued back in 2015, argue that they bought a chip they thought would have eight independent processor cores – the advertising said it was the "first native 8-core desktop processor" – and paid a premium for that."
 
Core is only part of the mesurment and that's fine. the buyer is responsible for doing the research on what they buy. Cores are not and have never been the same. I have 72 core 288thread CPUs the would be shitty to game on but that was understood when I bought them because the information to what I was getting was out there. Arm cores are another example of this. So are ibm cores. Yet they are still marketed correctly as elsewhere there is more information on what cores it has
 
A lawsuit getting defeated on the merits is one thing. Getting thrown out before any presentation of facts is another.

This lawsuit shouldn't even go to trial. It has no technical merit whatsoever, and serves to do nothing except waste AMD's money and fatten lawyer wallets. There will be nothing good that comes out of this lawsuit, and a technically incompetent judge-defined definition of what a core is wouldn't do anyone any good either. A technically competent judge would throw this case out on the first day.
 
Thw coffee was brewed at too high of a temperature and severely burnt the plaintiff. It isn't a joke.

Yeah it happened but do you really "need" a warning on the coffee cup ? Does that fix anything?
The fact that it was brewed too hot would still be brewed too hot even with the notice on the cup...

Warning doesn't say coffee should be within XX to YY temperature so what does that warning do ? Paying for her burnt is ok and needed but in the end was caused by coffee been our of specification if I may say, bulldozer wasn't OOS, it was designed like this and couldn't be exploited for many reason.
 
Yeah it happened but do you really "need" a warning on the coffee cup ? Does that fix anything?
The fact that it was brewed too hot would still be brewed too hot even with the notice on the cup...

Warning doesn't say coffee should be within XX to YY temperature so what does that warning do ? Paying for her burnt is ok and needed but in the end was caused by coffee been our of specification if I may say, bulldozer wasn't OOS, it was designed like this and couldn't be exploited for many reason.
You know that hot water from a faucet is hot, right? But that doesn't mean businesses can just turn their water heaters up to 200 degrees and burn people's skin off their hands when they use a faucet that they "knew was hot."
 
Perhaps after litigation there will be a more precise definition. The term, I'd argue, is too important to be left open to interpretation.

After means nothing at all. I asked specifically now!
 
This lawsuit shouldn't even go to trial. It has no technical merit whatsoever, and serves to do nothing except waste AMD's money and fatten lawyer wallets. There will be nothing good that comes out of this lawsuit, and a technically incompetent judge-defined definition of what a core is wouldn't do anyone any good either. A technically competent judge would throw this case out on the first day.

I have to wonder if this is something that has been instigated by Intel, without anyone knowing about it.
 
  • Like
Reactions: N4CR
like this
I wish people would stop using this as an example. Look up that, case. The lady almost died from third degree burns. It practically melted the cup, she didn't tilt a little coffee on her and sue. That is negligence
Add to that disaster, that she didn't get paid anywhere near 12million or whatever shit.
Companies just looooooooooove how so many simply stab their own necks in thinking oh my God, she got 12 million, blah, blah, blah we got to reform things! everyone is sue sue sue! (that is not true either).
In truth most people are like she was:
Please help me with medical bills, I burned my self badly with your coffee.
McDonald is like FUCK YOU!!!
Only then people go like da fuk?, this ain't fair...
Or like the hospital, its like well you left me all this shit inside and its giving me pain and an infection, well, okay just take it out... then.... you get billed for the correction of their mistake!!!
WHO WOULDNT SUE THEN?!
JESUS?
Serious people, change your mind set, no individuals or even class actions are abusing no freaking companies... they loooove when you believe that though, they want to take more and more rights away.
And no, this AMD lawsuit or not, the class action system is not 'abused' (no I won't give a shit about 'examples') nor it should 'go away' that is just stupid.

Think of this: Tesla fine for a tweet that affected "Wall Street" : 30 million, and management shake up.
Tesla fine for worker safety: 30K.

Think of THAT next time your brain is arguing in favor of a company or anyone in power already gaining even MORE power, and you less justice, or less access to justice.

Yeah, I like AMD, and I buy and will buy their products.
I don't think the lawsuit has merit and will probably die when someone with a better understanding takes it on, however, even if they lose, millions, I don't give a shit, it says nothing on class actions, let alone a need to reform it.. plenty of injustice to go around for the little guy.
 
This lawsuit shouldn't even go to trial. It has no technical merit whatsoever, and serves to do nothing except waste AMD's money and fatten lawyer wallets. There will be nothing good that comes out of this lawsuit, and a technically incompetent judge-defined definition of what a core is wouldn't do anyone any good either. A technically competent judge would throw this case out on the first day.

I completely agree, this is not serving anyone.
I fully expect a series of amicus briefs to be filed explaining the facts to the judge.

A very similar case existed with Intel's P4. It had a very high clock rate for the time, but throughput was quite a bit lower at those clocks than other architectures. Should Intel be sued because it wasn't really a 3 GHz part? No, it really did run at 3 GHz. Bulldozer really does run 8 wide. Neither is just terribly great at what it does.

Some designs being better than others has ALWAYS been the case in the industry. And just like all industries, the buyer should actually understand what they are buying, if they care. Insert car analogy here, if desired.

For computing performance, you'd look at reviews to see what the throughput is on the workloads most similar to your own. That has, and always be what people should base purchasing decisions upon.
 
core.png


here's the lawsuits definition of a core.
 
Yeah it happened but do you really "need" a warning on the coffee cup ? Does that fix anything?
The fact that it was brewed too hot would still be brewed too hot even with the notice on the cup...

Warning doesn't say coffee should be within XX to YY temperature so what does that warning do ? Paying for her burnt is ok and needed but in the end was caused by coffee been our of specification if I may say, bulldozer wasn't OOS, it was designed like this and couldn't be exploited for many reason.

You probably never installed a water heater before, and that's fine. I've installed a couple in my home and they all come with a chart and warnings clearly spelling out what degree burns you get at what temperature setting. Tankless water heaters with an electronic controller are locked at 120F so that people don't burn themselves. You can unlock it on your own responsibility up to 140F. The difference is huge when you wash your hands. For a commercial setting you can unlock it up to 180F but you need to install a special controller. This goes for Rinnai, Takagi, and all major brands.

The reason why I used this example is because the same goes for brewing coffee in a restaurant. There are safety guidelines, and the only reason why they disregarded them was so the coffee stay hot longer. It's stupid. That lawsuit had merit, but McDonalds and other corporations financed media campaigns to downplay the damage and ridicule the victim.

The AMD situation is different, as the lawsuit had no merit. But AMD is not Intel, so they will get their asses handed to them. Intel never had to pay anything for the damage they caused with their flawed CPUs, with Meltdown and Spectre. AMD doesn't command the same influence, not even close.
 
You probably never installed a water heater before, and that's fine. I've installed a couple in my home and they all come with a chart and warnings clearly spelling out what degree burns you get at what temperature setting. Tankless water heaters with an electronic controller are locked at 120F so that people don't burn themselves. You can unlock it on your own responsibility up to 140F. The difference is huge when you wash your hands. For a commercial setting you can unlock it up to 180F but you need to install a special controller. This goes for Rinnai, Takagi, and all major brands.

Sounds like I’m staying away from tankless heaters, I like my showers scalding. Interestingly in the 80s all the scalding incidents were typically caused by the low flow showerheads spiking temperatures 30-40 degrees higher with the change in pressure. Science!
 
This is what it boils down to--judges with *no* technical acumen are allowed to judge technology. Oh, brother. A definite weakness in the system, imo.
 
Back
Top