AMD Dragged to Court over Core Count on "Bulldozer"

[qubit]

@FordGT90Concept That's one battered box, lol. And good point with the task manager display.

 

[blobster21]

 

[HCT3000]

Pretty sure he's going to win. I don't think there's any nomenclature to properly describe Bulldozer's design and even if it had existed, AMD wasn't using it.


x264 HD Benchmark runs on GPU and AMD undeniably has a stronger GPU in FX-8150 than Intel has in i7-2600K. The problem stems from floating point operations executed on the CPU. If you heavily load the FPUs in one core, the FPU performance of both cores will effectively half.

FX chips have no IGPU.....

 

[FordGT90Concept]

I know that. Displays need some kind of GPU driving them and software can use DirectCompute on them to accelerate computing.

 

[R-T-B]

You say potato, I say potato core.

No, this is a potato core:

 

[the54thvoid]

No, this is a potato core:

View attachment 69076

No, that's a Potato Chip!

Thank you, you've been a wonderful audience - I'm here all week.

181 posts to set up TPU joke of the year. Thanks go to Ford and his tireless efforts at core definition.

 

[xChoice]

cof cof im not reading the whole comments, and i doubt someone will read mine.

Just in case no one mentioned it, what happens with this 8 core arm processors that some companies are advertising while there are in fact 4 cores + 4 cores gpu, i mean, they are "cores" but hey! they are not the cpu itself.

As for the top of the line fx processor, I thought they were real 8 complete cores, not 4 complete cores + some left overs that were good enough for running something like a hardware HT......

 

[FordGT90Concept]

Just in case no one mentioned it, what happens with this 8 core arm processors that some companies are advertising while there are in fact 4 cores + 4 cores gpu, i mean, they are "cores" but hey! they are not the cpu itself.

http://www.techpowerup.com/forums/t...count-on-bulldozer.217327/page-3#post-3367326

As for the top of the line fx processor, I thought they were real 8 complete cores, not 4 complete cores + some left overs that were good enough for running something like a hardware HT......

That is the reason why this thread has 183 posts: it is subject to debate.

 

[xChoice]

nono im not talking about big.little, those are 8cores, with 4 that have more resources and work faster, and 4 that are more... littles .

anyway, if its an 8 core chip, you should be able to power gate those true 4 cores, and let the other 4 online, if they are full complete cores they should work.... isnt it that way?

 

[FordGT90Concept]

http://www.realworldtech.com/bulldozer/10/

Llano is also the first announced CPU from AMD to use power gating. The power gates are implemented as a footer ring of NFETs around the periphery of a core and L2 cache, using the package plane as a virtual ground. While Bulldozer’s hierarchical and shared microarchitecture improves area efficiency and throughput, it does complicate power gating. Conceptually there are five major circuit regions of each Bulldozer module – the shared front-end, the two integer cores, the floating point cluster and the L2 cache. Unfortunately, if a single core is active all of these regions (perhaps save the other integer core) must be active. The benefit of power gating a lone integer core is not worth the complexity of the implementation problems, especially since the operating system scheduler should be power-aware. As a result, Bulldozer’s granularity of power gating is at the module level. Each Interlagos die incorporates at least 4 power gates, one for each Bulldozer module.

It can't drop to 7, 5, 3, or 1 logical processors like a real 8-core processor, if that is what you mean.

There are some articles that say it can power down unused components (like the FPU) but it was phrased in a way that was theoretical so I don't know if it really does.

 

[xChoice]

*cof* *cof*

module= core and the integer cores= excution units??

But those integer cores have an alu, a control unit, and registers, and L1 caches... sooo they are microprocessors on their own...... arent they?

<.< i starting to fuck up my brain

 

[FordGT90Concept]

According to AMD, FX-8### and FX-9### has four "modules" and each module has two "integer cores" for a total of eight. The plaintiff alleges that it only has four cores.

The "integer cores" do not understand x86 instructions. They have to be decoded and dispatched to FPU/"integer core" first and in Bulldozer's design, that happens in a shared component. The "integer cores" without the rest of the module, therefore, is quite useless.

 

[xChoice]

well I agree with you in that, they are not x86 cores.

so in context, that is what we are buying, an x86 cpu that has n ammount of x86 cores, as those are not x86 cores by their own... well thats it for me

 

[Pill Monster]

No, it has 4 siamesed cores who's performance suffers when running two threads on one. I tried to sympathize with you, but you got one hell of a shit attitude. I gave you a reasonable explanation of how I see it, so all you had to do is discuss it in a reasonable manner with me and show some respect. I'm not going anywhere.

Yeah, go on, just get off this thread like you were gonna.

Bullshit buddy. Do you take me for an idiot?

First u gave me an underhanded insult insinuating I was pissed off due to people disagreeing with me, and made out i was the only one arguing the point. . Which is a garbage as you know. I don't mind a good healthy discuission at all. This isn't.

I was pissed off because I was being trolled and right at the time I was over this debate And it's not the first time it's happened either.
So don't act like you're the victim here..

 

[Roph]

I know that. Displays need some kind of GPU driving them and software can use DirectCompute on them to accelerate computing.

x264 does not use directcompute, it can (optionally) use OpenCL. You must specifically tell it to though or swap in an OCL build, and in many cases it results in slower encoding. Few parts of x264 w/ OpenCL are accelerated.

I encode without OpenCL "acceleration" as for example, when doing 1080p to 720p with OpenCL enabled I cap out at about 70fps. CPU only, about 130fps.

 

[qubit]

Bullshit buddy. Do you take me for an idiot?

First u gave me an underhanded insult insinuating I was pissed off due to people disagreeing with me, and made out i was the only one arguing the point. . Which is a garbage as you know. I don't mind a good healthy discuission at all. This isn't.

I was pissed off because I was being trolled and right at the time I was over this debate And it's not the first time it's happened either.
So don't act like you're the victim here..

I was not insulting you and was just being nice to you as I could see you were pissed off. I then quite reasonably explained where I think AMD has gone wrong here and hoped to have a reasonable discussion with you about it.

Obviously you've misunderstood my intention and have given me one hell of a douchebag response instead of responding reasonably - and you're compounding it by continuing it. Yes, it's you making a personal attack against me not the other way round. You're the one getting all drama queen threatening to leave the thread, but then just can't let go, lol. Get a grip FFS.

Whatever, I don't give a fuck. This discussion is over.

 

[Captain_Tom]

A "core" is a complete computing unit. AMD proves it is not complete in their 6-"core" Bulldozer processors. The two units packaged together are inseparable or they would have sold 7-"core" Bulldozer processors having only gated off the one that was defective.


Because compression is mostly integer-based where Bulldozer performs more like an 8-core processor. Even considering the widely different architectures and Bulldozer having a design ideal for it, it doesn't win by a very large margin. The lawsuit is about the worst case scenario (saturated FPUs) and you're citing the best case scenario (saturated ALUs) where Bulldozer's non-traditional design shines. The latter doesn't forgive the former.

Hey genius, then why are there no 3-core intel CPU's? By your logic Intel not making some 3-core i5/i7 proves that Intel doesn't have 4 cores in an i7.

 

[Deleted member 6693]

If you right-click on the graph, left click and highlight "graphs to show >," and select Logical Processors, it might show 16 little graphs of "FordGT90Concept please insert what you want to call it in here because you feel strongly that AMD must burn in a fire for their fraudulent misrepresentation of cores" to further prove your point, Vulkan. If it doesn't show 16 little graphs, then on an ironic side, it would support Ford's point. Maybe? Doubt it.

Now that I think about it, looking at your performance tab, Vulkan, maybe AMD just markets it's CPUs as having X amount of Logical Processors, but they call it "Core Processors. Intel market CPUs based on y-amount of Cores which are really 2 Logical Processors to 1 Core. I think that's one part of where the problem occurs between FordGT90Concept and others having their disagreements.

If we go back to the original Topic of Discussion, the Lawsuit is probably dead on arrival. I don't believe the court is going to award damages. At best, it's a marketing blunder on AMD's end.

 

[FordGT90Concept]

Hey genius, then why are there no 3-core intel CPU's? By your logic Intel not making some 3-core i5/i7 proves that Intel doesn't have 4 cores in an i7.

Good point. Intel could do that but they don't. They're sold as dual-cores. AMD presumably sells them as six-cores.

 

[Captain_Tom]

Good point. Intel could do that but they don't. They're sold as dual-cores. AMD presumably sells them as six-cores.

Wrong - something in Intel's architecture doesn't allow them to do that, because they clearly would. Same with AMD's construction archs - how they built it allows them to save TDP, money, and die space by having the cores share some resources.

 

[Aquinus]

No rebuttle to my point made earlier? People seem to be content spouting off the same thing they've been saying for pages but not following up anything substantive...

http://www.techpowerup.com/forums/t...count-on-bulldozer.217327/page-6#post-3367789

 

[FordGT90Concept]

Wrong - something in Intel's architecture doesn't allow them to do that, because they clearly would.

Sure it does. I distinctly remember being able to disable 1, 2, or 3 cores via the BIOS on Core i7-920. The reason why Intel doesn't sell tri-cores and why AMD didn't sell Phenom II X5 Thubans is because they a) need enough volume to warrant doing so and b) would have to generate another SKU for it (creates a lot of hassle from additional UPC code to the creation of store front pages at vendors like Newegg and Amazon).

No rebuttle to my point made earlier? People seem to be content spouting off the same thing they've been saying for pages but not following up anything substantive...

http://www.techpowerup.com/forums/t...count-on-bulldozer.217327/page-6#post-3367789

Every point was already rebutted on its own so I felt no need to repeat myself. The majority was about performance which is why I replied the way I did: 4-core-like performance is cited in the lawsuit for damages and it's awkward design can be at least partially faulted for that.

 

[qubit]

Wrong - something in Intel's architecture doesn't allow them to do that, because they clearly would. Same with AMD's construction archs - how they built it allows them to save TDP, money, and die space by having the cores share some resources.

No, I don't think that's the case with Intel. In the BIOS I can set the core count to 1,2,3 or 4 with separate HT on/off. I've experimented with that before and the computer worked just fine, therefore, it would be trivial for Intel to fuse off a core if they wanted to. They already do that with two cores to make dual core CPUs.
Why they don't sell tri core CPUs you'd have to ask them.

@FordGT90Concept Dammit, you beat me to it!

 

[Frick]

And when Ford rebutts something it stays rebutted. Especially when he has qubit backing him up.

EDIT: If it wasn't clear I was sarcastic.

 

[Captain_Tom]

No, I don't think that's the case with Intel. In the BIOS I can set the core count to 1,2,3 or 4 with separate HT on/off. I've experimented with that before and the computer worked just fine, therefore, it would be trivial for Intel to fuse off a core if they wanted to. They already do that with two cores to make dual core CPUs.
Why they don't sell tri core CPUs you'd have to ask them.

@FordGT90Concept Dammit, you beat me to it!

No you can set it to 1,2, or 4.