AMD Q414 results

[elemein]

-snip-

"We re talking of the definition of high perf and you said that the metric is single thread performance, but now seems that you re protesting that core count, TDP and so on have also a say in this definition...."

Quote me where I said that ST was the performance metric. Actually, quote me on all those insane things you said I said. It seems you completely forgot to reply to them. Was it because they weren't true? Funny.

The performance metric is that it performs well practically for what it is. No one cares if a CPU has the greatest MT performance to ever exist if they cant use it, and, you know what? The vast majority of computer users cant. Use. It. How many times must I repeat myself? No one cares about how a CPU could perform. They care how it does perform. And while the 8350 could perform very well, for the vast majority of people in the vast majority of usecases it performs worse than comparable Intel offerings. Don't know what's so hard to understand about "the metric is how it performs practically." It's a pretty simple metric that was stated right away but you had to make up metrics I never said to even try to argue.

"Why the need of all thoses metrics since ST perf is THE metric, at least according to your sliding rule former position...."

Now. Performance is the metric. And if you're gonna take TDP and etc. out of reason, then again: i7-5960X vs A4-4000. What? That's nonsensical? You're being silly. No it's not.

"Dont know either from where it comes but i thought that a purely artificial argument would be left unoticed..."

What? You put some gibberish in there to be left unnoticed? Uhm. Ok.

"More seriously power increase as a square of voltage, take a 2GHz CPU, increase voltage by a 1.414 factor without changing the frequency, the dissipated power at 100% load will be double."

Okay. Now what's that have to do with the argument at hand?

"Ok, then i can only advise you to better follow the physics courses..."

I'm in Info. Sec.. We don't have any physics courses for it. Except for communications, but that's not totally relevant to this argument.

 

[Abwx]

it performs worse than comparable Intel offerings.

Eventualy in games, a i5 46xx is more expensive than a FX8350, and significantly here in Europe.

Applications :

Games :

And they were sympathetic to castrate the FXs by using 1600MHz RAM since Intel only support this standard with the CPU s used as comparison, as if it was AMD s fault that Intel dont support officialy 1866MHz and to not give an "unfair" advantage to AMD according to their sayings, of course when testing the X99 plateform it didnt matter than DDR4 was faster..

Ypu have the details of the applications and games as well as the case by case perfs in this page :

http://www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html

 

[elemein]

Eventualy in games, a i5 46xx is more expensive than a FX8350, and significantly here in Europe.

Applications :

Games :

And they were sympathetic to castrate the FXs by using 1600MHz RAM since Intel only support this standard with the CPU s used as comparison, as if it was AMD s fault that Intel dont support officialy 1866MHz and to not give an "unfair" advantage to AMD according to their sayings, of course when testing the X99 plateform it didnt matter than DDR4 was faster..

Ypu have the details of the applications and games as well as the case by case perfs in this page :

http://www.hardware.fr/focus/99/amd-fx-8370e-fx-8-coeurs-95-watts-test.html

You know the extra 266 Mhz on the RAM wouldnt have made that big of a difference.

Also, can I ask how these benchmarks are averaged? A lot (like... 9/12 of these) of these programs are programs most people wouldnt use. Though sure, let's special plead and take the average of the supplied benchmarks.

In multithreaded performance, the 8350 scores higher in a lot of benchmarks than the i5. As expected. That's great. Unfortunately, and really, undeniably, most consumers don't use apps like MinGW.

Games. The i5 scores higher in almost all of them. Many more consumers play games, but admittedly a lot don't.

So what have we learned?

The 8350 has good MT performance and the i5 has better ST performance, which games tend to rely on.

Is this new information? No.

What kind of apps do most consumers use? Apps that rely on ST,

Is this new information? No.

 

[Abwx]

Firefox is single-threaded, isn't it? I use Waterfox, a 64-bit port, and it uses a rough max of 50% of my dual-core 1007U CPU, when a thread gets "stuck" and starts chewing up CPU time.

I opened the page to anwser you in another tab while checking teh task manager, both cores of my T4400 are fired at the same time and usage is balanced between the two cores, i already noticed that the four cores are fired when using a 4C Kabini, so this browser is obviously multithreaded.

Edit : Same thing when i clicked to send the post, both cores are used.

 

[Abwx]

What kind of apps do most consumers use? Apps that rely on ST,

ST tasks that use very few ressources, if this was the cases thoses softs would had been multithreaded, but i already asked you, what are thoses softs that are so demanding and that you pretend using by the truck loads, what are they..?.

As a consumer, myself and many (like many many... Like the overwhelming vast majority of computer users) use software that rely heavily to an incredibly high degree on singlethread performance.

Incredibly high degree of ST perf....

For what exactly, facebook, post at AT..?.

 

[elemein]

ST tasks that use very few ressources, if this was the cases thoses softs would had been multithreaded, but i already asked you, what are thoses softs that are so demanding and that you pretend using by the truck loads, what are they..?.



Incredibly high degree of ST perf....

For what exactly, facebook, post at AT..?.

Games, mostly. I agree, most consumer's things can be handled by an Atom with a good RAM config. But games are played by a lot of people, and that relies on ST mainly.

 

[Abwx]

Games, mostly. I agree, most consumer's things can be handled by an Atom with a good RAM config. But games are played by a lot of people, and that relies on ST mainly.

I was sure that the games argument was to be raised, ok, i pointed above that it was a stronghold of Intel generaly, but what else.?.

Another game.?.

Personaly i dont game and most people i know who do so use ridiculously underpowered solutions, typicaly the IGP, you think that in such a case Intel ST perf will be an advantage.?.

Not at all, it s the GPU that makes the difference, so Intel is better in games only with discrete solutions, for the vast majority of users Intel s game perf is well below AMD s APUs.

Anyway, what about the rest of thoses highly high ST perf demanding softs.??

 

[Ramses]

Incredibly high degree of ST perf....

For what exactly, facebook, post at AT..?.

This is a very important point.

You guys, while bringing up great points and making fine conversation and being generally very smart and informed, are akin to a bunch of Formula 1 drivers discussing the merits of a Toyota Camry.



Fast or slow has nothing to do with AMD failing to sell lots of CPU's beyond to a pretty small group that they weren't aiming at anyway, and another group that knows just enough to be dangerous (as in reading reviews and benchmarks and buying based on those instead of actual real world performance).
Absolutely there are people, and a concentration of them on sites like this no doubt, who NEED a strong i5 or i7, but it is NOT the majority of the buying public. It can't possibly be. You can count every registered user on every hardware forum you can find and there aren't enough people. We are a minority. And that's OK.


Most of the FX and APU's AMD has had for sale the last few years are plenty fast for at least 75% of PC users. If you were to swap out the i3 or i5 and in a lot of cases the i7 in most folks boxes and replace them with an A10 of FX or whatever, they'd never notice. If you were to put in a noticeably slower CPU, but add an SSD, they would think it was an upgrade. They don't understand cores and threading. ANY of these CPU's from the last few years will do what the vast majority of PC users do daily just fine. I asked what "high performance computing" or whatever was while ago since everyone keeps talking like they are constantly doing some sort of ungodly CPU intensive tasks, and you aren't. Neither are the masses. This is also why even among "enthusiasts", upgrade cycles have stretched out to years for a CPU. Years. The software isn't challenging by and large. Check the traffic, or lack of, on the FS/FT forums, any of them. Check the thread on this very topic on this forum from the last month or so. Look in your own computer and ask if it's really a burden, that CPU. Odds are it is not. And AMD's decidedly midrange stuff would have worked just as well. Lack of performance, even though it's clearly below Intels, is not why they didn't sell and making a new super-CPU isn't going to fix that. I'm all for it, but if I were deciding where to put money at AMD I don't know that R and D would get too much of it. Something else is wrong.

Single thread performance.
I am 36 years old, my first PC was a 486 hacked into an IBM XT case, and I ran NT4 with dual
pentium pro's and II's and III's and a few S370's for a long time. I have a little experience with
threading and software, or the lack of.
This has become if anything LESS important the last few years in my experience.
Who writing software is actively trying to not-use more cores/threads? Who?
Crappy cheap games?
Who is trying to use more cores?
Anyone who plans to have popular well received software
that performs well on modern hardware.
Big budget games scale and use several threads quite well. Crysis 3 blew me away with how
well it ran on my old slow FX, so did BF4.
This is why my 9590 or 8350 still games just fine all these years later.
BETTER in fact than it did a few years ago.
It's actually BETTER as it's aged since software has caught up to it, and I expect with
the DX12/Mantle bit in the pipe that will continue for a bit.
NO it's not amazing or better than Intel stuff, but it's a solid alternative for
the money, and it always was. (not the 9590, lol).
I have got to be at least an above average user, I run a VM, two browsers with 20 or 30 tabs each, couple monitors, bunch of crap running in the tray, etc, etc, etc same stuff I've been doing with computers for years and everything happens just next to if not instantly.
Instantly.
No wait, no lag, no bog, instantly.
It's boring watching my CPU usage.
And most folks are not using a computer nearly as hard as I am.
If you quit benching and start using, things are other than they might seem.

re: games there is a fine bunch of off the cuff benches here http://www.overclock.net/t/1534128/vishera-vs-devils-canyon-a-casual-comparison-by-an-average-user



The question to ask, and where the money should be spent imo, is finding out why those people that'd be perfectly OK with an APU of FX or whatever, aren't using one. And it's not because they are slow or have crappy single thread performance. It's not ideal, Intel does it better, and absolutely there is room for modest improvement (especially in power usage) but lack of performance is not the reason AMD is in financial trouble. It only makes sense for that to be the reason in our little closed world of nitpicky enthusiast geeks. While that world may be a heck of a lot more mainstream than it once was, it's still not THAT mainstream.

 

[VirtualLarry]

I opened the page to anwser you in another tab while checking teh task manager, both cores of my T4400 are fired at the same time and usage is balanced between the two cores, i already noticed that the four cores are fired when using a 4C Kabini, so this browser is obviously multithreaded.

Edit : Same thing when i clicked to send the post, both cores are used.

You DO know, that the windows scheduler moves threads between CPU cores to distribute the load around? So for a single-threaded app, your behavior is expected. That, in itself, doesn't mean that an app is multithreaded.

Let me guess, on a dual-core, Firefox never uses more than about 50% CPU time, and on a quad-core, not more than 25%...?

 

[VirtualLarry]

They don't understand cores and threading.

Someone here doesn't.

 

[Abwx]

You DO know, that the windows scheduler moves threads between CPU cores to distribute the load around? So for a single-threaded app, your behavior is expected. That, in itself, doesn't mean that an app is multithreaded.

Let me guess, on a dual-core, Firefox never uses more than about 50% CPU time, and on a quad-core, not more than 25%...?

I just looked when opening the answer page, total usage would exceed a single core capability, both cores peak at 65-70% before release, sure that a single core would do it since the peak is quite short, neverless it exceed the single core throughput.

 

[elemein]

I was sure that the games argument was to be raised, ok, i pointed above that it was a stronghold of Intel generaly, but what else.?.

Another game.?.

Personaly i dont game and most people i know who do so use ridiculously underpowered solutions, typicaly the IGP, you think that in such a case Intel ST perf will be an advantage.?.

Not at all, it s the GPU that makes the difference, so Intel is better in games only with discrete solutions, for the vast majority of users Intel s game perf is well below AMD s APUs.

Anyway, what about the rest of thoses highly high ST perf demanding softs.??

Weird how I go shopping for an hour and come back to intelligent replies. It's really refreshing. I was on another "tech" forum previous to this and it was nowhere near as civilized or intelligible. This is great. Love it.

Anyways, as for applications that only make use of ST performance and are intensive; there aren't many worth mentioning. True. But then again the outlook is really; anything a consumer does isnt CPU intensive at all except for games, which, Intel does better at as a CPU. Bringing in an APU and saying it's a better iGP than a lone Intel CPU is not only totally fine, but true. So yes, a discrete solution + Intel would be best of the best while an APU would be just fine for most folk. Most folk who are gamers only play LoL, DoTA, CS:GO or... Maybe Diablo III? I mean after those games what else if there that's worth mentioning... Maybe Minecraft but that runs on just about anything that can add a few numbers together reasonably quickly...

So... Fair enough. I don't think we "disagree" anymore (or atleast, exhausted anything worth arguing about which is... kinda the same thing.)

We have different views and have come to agree to disagree; or atleast I have.

 

[elemein]

-snip-

This is all true. Can't deny it.

Though if it werent for performance, TDP or power consumption, what is the reason AMD is in such a bind they are in now? Is it down to marketing? Intel getting more design wins? Simply market share multiplying?

 

[frozentundra123456]

I just looked when opening the answer page, total usage would exceed a single core capability, both cores peak at 65-70% before release, sure that a single core would do it since the peak is quite short, neverless it exceed the single core throughput.

Is anyone here advocating buying a single core cpu? I dont think anyone even sells one anymore. I tried it on a 3ghz i5 and max cpu usage was less than 30 percent. So even if you extrapolate back to a non-hyperthreaded dual core, that is not even close to full usage, and even that is just momentary.

 

[cbn]

I think Zen/K12 are two huge engineering challenges. AMD should be targeting slightly above A53 raw performance *and* performance per watt but comparable costs. This is the only way I see for AMD to develop the micro server market and still be able to get embedded business. Fail to reach the performance goal and AMD won't be able to get into embedded business with any advantage over the drove of ARM MPU designers, fail to reach the cost target and they won't be able to leverage on the chip to expand the embedded business.

The only chips that made money for AMD in the last 4 years weren't the big ones, but the small, IP friendly, cost controlled cat family. I don't see why AMD would throw this solid and profitable track record for the sake of daydreaming about a performance race they are ill-equipped to stay, let alone win. Whatever they are designing, it will have the business features and parameters of the cat family, not of the derp or K8 family. The design will be IP friendly, highly synthesis-able, efficient and with the smallest possible die.

Remember the big cores also suffered from expensive die size adding L3 caches that didn't add much in the way of performance (AM3+) or too large iGPUs (FM1/FM2/FM2+).

In contrast, the cat cores were a much more lean and balanced product.

 

[Ramses]

This is all true. Can't deny it.

Though if it werent for performance, TDP or power consumption, what is the reason AMD is in such a bind they are in now? Is it down to marketing? Intel getting more design wins? Simply market share multiplying?

Don't know. Shitty marketing? There was some discussion in another thread about the difference between marketing and advertising, or lack of both from AMD. The only thing I've noticed via personal experience with laptops is the AMD chips are in crappier quality laptops. I bought an A8 APU Toshiba that was a piece of crap as far as case/screen/keyboard/trackpad, spent 2x as much on an (overpriced)HP with an i7 that was much better. I'd have gladly spent a few hundred more for a better laptop with the same guts as the cheapy Toshiba rather than twice as much for an i7 that does the exact same things, but it wasn't available locally and I'm not especially brand loyal so...

Intel does what they do very well lately, and surely power/TDP is an issue for both OEM's and volume buyers, but they did make and sell SOME of them, so why not more?
Nothing really seems to fit and I think we are short on data as to why.

I've built PC's with FX's and A8/A10 APU's for folks in the last year and they are all thrilled with em. Money saved over Intel CPU went to better SSD and faster RAM and better PSUs as these were non-gaming boxes. I don't do it much anymore, it's a rare genpop person that will listen and learn about component quality when I explain I can't beat box store PC prices. Just quality.

I'm a lot more interested in how AMD is going to fix sales and marketing than I am in a new CPU nobody really needs.

 

[Abwx]

Weird how I go shopping for an hour and come back to intelligent replies. It's really refreshing. I was on another "tech" forum previous to this and it was nowhere near as civilized or intelligible. This is great. Love it.

Anyways, as for applications that only make use of ST performance and are intensive; there aren't many worth mentioning. True. But then again the outlook is really; anything a consumer does isnt CPU intensive at all except for games, which, Intel does better at as a CPU. Bringing in an APU and saying it's a better iGP than a lone Intel CPU is not only totally fine, but true. So yes, a discrete solution + Intel would be best of the best while an APU would be just fine for most folk. Most folk who are gamers only play LoL, DoTA, CS:GO or... Maybe Diablo III? I mean after those games what else if there that's worth mentioning... Maybe Minecraft but that runs on just about anything that can add a few numbers together reasonably quickly...

So... Fair enough. I don't think we "disagree" anymore (or atleast, exhausted anything worth arguing about which is... kinda the same thing.)

We have different views and have come to agree to disagree; or atleast I have.

Ok, it s nice to be in agreement for some things, notice that if i m talking of IGP it s not specialy because of games but due to all my old laptops being outadted mainly by the virtue of a crappy IGP, the only one that manage to be usable to this day for basic browsings is a Pentium 4 1.6 with a dedicated Radeon GPU, some with a better CPU but a mediocre IGP dont cut it anymore, not even for viewing Divx or DVDs.

Is anyone here advocating buying a single core cpu? I dont think anyone even sells one anymore. I tried it on a 3ghz i5 and max cpu usage was less than 30 percent. So even if you extrapolate back to a non-hyperthreaded dual core, that is not even close to full usage, and even that is just momentary.

Not at all, it was just to point that a single core can eventualy do the job when browsing, assuming of course that there s not 10 tabs opened with flash animations at will, also you used a 3GHz recent CPU, mine is a 2C 2.2 Pentium Penryn, extrapolating from your usage it would peak at a way higher %age in a same configuration as yours, that is, with a theorical 4 core count.
I get about 65-70% on two cores.

 

[frozentundra123456]

Well my point was that there is a huge middle ground between pure single threaded applications and heavy multithreaded apps that can load all 8 cores of an FX8xxx. And in that vast middle ground, intel has better lightly threaded performance and more than enough multi-threaded performance.

 

[frozentundra123456]

Don't know. Shitty marketing? There was some discussion in another thread about the difference between marketing and advertising, or lack of both from AMD. The only thing I've noticed via personal experience with laptops is the AMD chips are in crappier quality laptops. I bought an A8 APU Toshiba that was a piece of crap as far as case/screen/keyboard/trackpad, spent 2x as much on an (overpriced)HP with an i7 that was much better. I'd have gladly spent a few hundred more for a better laptop with the same guts as the cheapy Toshiba rather than twice as much for an i7 that does the exact same things, but it wasn't available locally and I'm not especially brand loyal so...

Intel does what they do very well lately, and surely power/TDP is an issue for both OEM's and volume buyers, but they did make and sell SOME of them, so why not more?
Nothing really seems to fit and I think we are short on data as to why.

I've built PC's with FX's and A8/A10 APU's for folks in the last year and they are all thrilled with em. Money saved over Intel CPU went to better SSD and faster RAM and better PSUs as these were non-gaming boxes. I don't do it much anymore, it's a rare genpop person that will listen and learn about component quality when I explain I can't beat box store PC prices. Just quality.

I'm a lot more interested in how AMD is going to fix sales and marketing than I am in a new CPU nobody really needs.

Well, here is a novel thought, maybe intel cpus actually *are* better for the way the vast majority of buyers use their computers. Nah, that couldnt be it.

 

[Ramses]

Only better if price is equal or less or you need more than amd stuff can provide. You know that is way too easy an answer though surely. None of the facts as near as I can determine them match the hypothesis.

 

[AtenRa]

I'm definitely not in the minority when it comes to thinking Zen will be a cat successor. Many people believe that.

8-core 95W Cat successor at 14nm FF ?? In order for this to happen this ZEN would be 6GHz+ not to mention it would only be 80-100mm2 SoC.
Dont thing so, ZEN is a big core. This must be an 8-core 16 Thread die coming from server, put on a package for FM3 socket to replace current AM3+ FX line.

 

[TurtleCrusher]

"Intel s better perf/watt is related to process and about nothing else,"

I hate when people say this. People are acting as if Intel using their process advantage which their engineers worked tirelessly over is some cheap, cheating, conniving and illegal trick that should be looked down upon. No.

Intel has that advantage. They worked for it. They deserve it. Whether "oh if PD had that advantage it'd be better!!" is true or not, that isn't how it is. So that argument is irrelevant.

And yes, I am saying it's not high performing because of the ST performance. As a consumer, myself and many (like many many... Like the overwhelming vast majority of computer users) use software that rely heavily to an incredibly high degree on singlethread performance. You can have an arch that has fourteen billion cores and have multithreaded performance the likes the Earth has never seen, but if no software takes advantage of it. It's useless. Irrelevant point.

The 2697 is Ivy Bridge and the 2650L is a 65W TDP CPU. What does that have to do with anything?

"Intel s recent 12 cores Haswell as not being high perf CPUs at all since due to frequency they have not better ST perf than an AMD FX83xx"

I don't understand what you're trying to say here. Are you trying to say that the FX83xx have higher ST performance than Haswell, while saying right before that they have less ST performance?

"You cant expect to have a bad GPU on the AMD side,"

Or I can expect it to have the iGP of the FX-8350. None.

"induce, what, only 32% higher TDP..."

What do you mean "only 32% higher TDP"? That's a whole 1/3. That's significant.

I'm sorry, your arguments previously were all explained and logical, but the ones detailed in your past post either don't make sense, conflict, don't have any sense of scale, or even, dare I say, a little desperate.

Newsflash: even web browsers are multithreaded. Just clock a multi-core CPU down to it's minimum frequency and load a single webpage and look at CPU usage. On the "Jesus's Middle Name Was Hume" thread in off topic, the FX-83xx series at 1400MHZ hits a utilization of 55% in chrome. That's over four cores completely utilized! And it's still very smooth and responsive. Almost all webpages are butter smooth on an FX-83xx @1.4GHZ. 4.2GHZ? Overkill really.

Content consumers need only an average APU at best. Content creators need cores and high throughput. Guess what? The FX series is overkill for one and a great performer for the other.

The "overwhelming vast majority of computer users" watch Netflix and talk to their friends on Facebook, not rely on single threaded tasks. The "overwhelming vast majority of computer users" are better off with an original Phenom/C2Q and an SSD than a heavily overclocked 4790k and a mechanical hard drive, even in the business sector. Be it encoding, compiling, decoding, streaming, compressing... Anything you use your CPU for with actual work is multithreaded. If not you and your IT department did a horrible job picking what software goes into your images. Multi-core in the professional space has been present since the Pentium days and even professional software then took advantage of it. If your software can benefit from an i5 or i7 above a Pentium, it can also benefit from an FX bulldozer or piledriver CPU.

If chrome can hit 85% utilization on a quad-core Bay-Trail, any system with 3+ desktop class cores is sufficient for the "overwhelming vast majority of computer users." Not only that, but most professional users as well. If you find that you're a part of that tiny niche group that needs single threaded software speed and that's all, microcenter is selling unlocked Pentiums for $49. But that's a niche, and should be treated as such.

 

[mrmt]

Remember the big cores also suffered from expensive die size adding L3 caches that didn't add much in the way of performance (AM3+) or too large iGPUs (FM1/FM2/FM2+).

The entire SNB-2C chips (iGPU included) would fit into the CPU area of Pilediver chips, meaning that even if AMD went for a small iGPU chip they would still have a cost handicap at comparable nodes. The Derp family doesn't fare well in terms of performance/watt and also performance/area, it is an extremely inefficient architecture, I can't really understand how it passed through AMD's internal controls and project check points and made it to product.

Bulldozer die size for a server chip isn't the problem. IBM is quite happy selling even bigger processors, and NVidia is also selling GPU accelerators bigger than AMD chips. AMD's sin was to design a poor chip which is also that big, meaning that it is both expensive to manufacture and fetches a low price on the market.

But the die size for an APU is a problem. AMD probably forecast that they would be able with the APU to replace a CPU + low end discrete, so they went for a big GPU + big CPU die and killed their bottom market SKUs. But once Intel directed OEMs to very thin designs and the 22nm finfet process outclassed whatever AMD had in terms of perf/watt, they were left in the cold with processors nobody wanted in the mobile market, because it was either too expensive compared to Intel's bottom market offers, or didn't offer enough CPU performance (and perf/watt) to fight for the mainstream market. To rub salt in the wounds Nvidia was very aggressive with perf/watt with Kepler and Maxwell, and as a result they got most of the mobile market for themselves, including the bottom market AMD arrogantly dropped.

To sum up, AMD failed to reach their intended performance targets, failed to take into consideration the shift to power efficient chips and failed to mitigate the risk of their iGPU chips not getting a premium for their bigger iGPU, all that in an extremely costly and complicated to validate package, very IP unfriendly. An unmitigated failure by all standards.

 

[mrmt]

Content consumers need only an average APU at best. Content creators need cores and high throughput. Guess what? The FX series is overkill for one and a great performer for the other.

Raw performance isn't the APU problem. Perf/watt and cost structure (the big die size) are the culprits you are looking. AMD has to throw a lot more of hardware at the problem than Intel, making their Derp chips a bad deal for both themselves and the supply chain.

 

[TurtleCrusher]

Raw performance isn't the APU problem. Perf/watt and cost structure (the big die size) are the culprits you are looking. AMD has to throw a lot more of hardware at the problem than Intel, making their Derp chips a bad deal for both themselves and the supply chain.

Am I talking to a child? Derp chips?