HardOCP Readers Ask AMD Bulldozer Questions

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,535
HardOCP Readers Ask AMD Bulldozer Questions - AMD came to HardOCP after the lackluster Bulldozer desktop launch and wanted to reach out to the hardware enthusiast community by answering questions that many of us had. We posted a HardForum thread and allowed [H] readers to ask their questions. AMD culled through the questions and AMD staffers answered 10 of the questions.
 
Almost all of the answers are marketing speak, they don't own up to the failures of BD in any way shape or form. Notice the 'why should I buy BD over Thuban' answer, they practically made up the answer. Thuban and BD are identical in performance. The efficiency answer is the same, they gave a 'we're forward looking' generic answer when in fact it is matched by a processor that's on an older 45nm node. Also look at the 'why is the pipeline longer' answer, they're basically saying that the same thing that made Netburst an epic fail is going to make BD an epic win, which is ridiculous.
 
Last edited:
To be honest....those answers are totally marketing bullshit.

Everyone wanted BD to rock...it honestly does not
 
Considering the number of layoffs there recently, I wouldn't be saying anything bad or revealing about a recent product launch either. I'm not surprised that they aren't very forthright in some of their answers.

It will be interesting to see how, if at all, the updated Windows 7 scheduling improves things.
 
I'd like to see AMD do better, but those were some pretty BS answers.
 
"It is also important to note that the "Bulldozer" architecture is configured and optimized for server throughput. The two integer execution cores present in Bulldozer are designed to deliver area- and power-efficient multi-threaded throughput."

Thank you and good night. At least they came out and said it. It would certainly explain some of the benchmarks we've seen. I really hope they start to focus a lot more on the single-thread aspect of computing in addition to multithreaded software/hardware. I *hope* they've realized that general use in the desktop encounters a whole lot of single-threaded, poorly threaded and legacy applications, but judging by these answers I'm not very hopeful and can see Piledriver being another disappointment.
 
Last edited:
"It is also important to note that the "Bulldozer" architecture is configured and optimized for server throughput. The two integer execution cores present in Bulldozer are designed to deliver area- and power-efficient multi-threaded throughput."

Thank you and good night. At least they came out and said it.

Which is ironic because even in this one area where AMD says BD is supposed to shine, Sandy Bridge does the job 11% faster while using 50% of the power.

http://www.tomshardware.com/reviews/fx-power-consumption-efficiency,3060-13.html
 
Which is ironic because even in this one area where AMD says BD is supposed to shine, Sandy Bridge does the job 11% faster while using 50% of the power.

http://www.tomshardware.com/reviews/fx-power-consumption-efficiency,3060-13.html

The anandtech review on the server-side bulldozer chips was also disappointing. It provides a small increase over previous-gen opterons but, much like the desktop, requires a nuclear power plant to provide the juice.

At least they've said they designed it for server's first -- which many people have been saying for a while now -- and regardless of how it performs there, it simply doesn't make sense on the desktop.

I do think it's honorable that it wasn't just PR guys answering the questions. Hats off for that. I mean that sincerely.

The reason why people are so upset is due to the processor's obvious underwhelming performance in many normal desktop-related tasks and workloads and the atrocious power consumption figures. I would have liked to see AMD at least come out and say that they're not too happy with these issues either. Not acknowledging the shortcomings really sets you guys up for some whacking on account of pissed off AMD fanboys and Intel fanboys and those who just want the best bang-for-their-buck.

Everyone wants AMD to succeed, but I think everyone but AMD believes that this just isn't the right direction.
 
Last edited:
This was quite a disappointing read. I have to agree that it has marketing speak all over it and the answers were extremely shallow and superficial. I actually felt as though they were fibbing a bit when they attempted to talk up performance per watt but I honestly don't care enough to go double check the reviews. I'm really surprised the marketing guy did not find a way to get a plug in for their "Radeon Edition" RAM. :rolleyes:

Perhaps AMD consulted Netflix on how to handle their customer relationships.
 
Those are pretty sad answers in the face of such an epic failure. Is their outright denial of just how bad this chip is supposed to be for our benefit or theirs?
 
Thanks for taking the time to get these answers ( Even if it wasn't the Droid we were looking for) Kyle. Long live the OCP!
 
Well they are right in that applications are going to be more multithreaded in the future thus they want to get their technology out there now to push that initiative forward. That's the type of thinking that will pay off in the long run. But it still seems that there shouldn't be that much of a disparity between Intel and AMD on single threaded applications. The gap is just too large to justify ushering forth the future while damning the past.

Also I would like to see them fix the bios issues that still keep us from playing certain Steam Games on the newer optimized bios. As it is right now we have to use the prelaunch bios to allow Deus Ex, Saint's Row the Third, etc to run. If we use the much faster updated bios we BSOD our systems. That's unacceptable as I actually had to call Microsoft to reactivate Windows because I have flashed my bios so many times trying to find a miracle. It seems each flash gives my PC a new identification as the processor's name is changed.

End users should have to tolerate such things over a month after launch.
 
From the comments already posted I'm not going to even click on the link to read the answers...why pull the scab off needlessly so it can hurt again! :p
 
So basically: "The bulldozer you buy today will kick-ass in two years ... maybe ... if everyone else changes their things to work optimally with it ... and a two year old chip that treads water right now is still up to the task then... ".

At least the Athlon64 was ahead of it's time *and* had kickass performance in non-speculative enivironments.
 
Well that was a dissapointing read. Much like the rest it was hard to get past the marketspeak, I'm just sad I wasted my time reading it.

Honestly I know why they did it but I would probably buy the FX 8130/50 this christmas in my new right if they manned up and said:

"we made mistakes, but this is our architecture of the future and we have faith in it"

instead we got

"Our product is better then our last product and competes against Intel, look to windows 8 for a (maybe) 10% increase in scheduler performance."
 
This part made me laugh out loud..."The new CPU core delivers higher frequency while maintaining IPC,"

They think it's a good thing by increasing MHz but not the IPC...
 
And the saddest answer award goes to:
7. Why would I buy a $275 Bulldozer cpu when the $170 1090t seems to equal its performance or actually do better at every benchmark and game we've seen?

Adam Kozak, Product Marketing Manager, AMD - We understand our customers make purchase decisions based on how they use their PCs, and in many cases our AMD Phenom™ II processors are a great (purchase).

For those ready for a more modern architecture, who want a desktop for high resolution gaming and to tackle time intensive tasks with newer multi-threaded applications, the AMD FX processor is a great upgrade.
Buy our new over-priced, under-performing processor because it's a more modern architecture!

The kicker is if someone would have included why buy either when I can go to a Microcenter and get a 2500k for $170 which stomps over both.
 
10 questions that could have made into one

Why does Bulldozer suck?
 
In their defense, someone has to open new markets for PC programming or the market stagnates. Giving programmers a way to differentiate themselves from the competition is great for the industry I'd think.

Game A is multithreaded and runs 15% faster on newer chips but just matches Game B on older chips. Game B is single threaded and struggles to meet the 60 fps standard consistently. Both are fun but, Game B sometimes has frame issues so multiplayer contestants are penalized whenever it happens. Which game would you purchase? Which developer would you start paying attention to?

For the above scenario to happen, someone has to put the technology out there to be utilized. I seriously doubt a company would invest resources into new technology without something to showcase it on. It's definately not perfect technology as anyone that owns BD chips can attest. But it's not a bad technology as I see the point in it.

Well I like shiny things.

:)
 
Wow after reading this I've lost a LOT of faith in AMD. I think perf/$ radeons still kick ass but after reading this I don't know if I can buy another AMD proc for a while. That was the biggest pile of marketing BS answers I've ever seen. Unless piledriver kicks total ass and drops right into my 990fx sabertooth I may be getting an iny bridge board and proc next go around.

To make things more clear I haven't built an Intel rig in like 7 years. So yeah way to go and kill my faith in you AMD.
 
Dammit Dark you beat me to it, that was easily the worst non-answer I've read in a while (and I read a lot of student bullshit). The biggest problem is that clock for clock the old AMD stuff is pretty much just as good in an enthusiast system as their new stuff, while saving a bunch of money. And yes that's not even taking into account getting stomped all over by intel. From what I read there, they didn't even really address the root of the question, which is "why should we as customers pay more for equal (or less) performance?"

The way BD performs I may have been less disappointed if they had just OC'd an X6, slapped a "new and improved hardware" sticker on it, and raised the price $50.
 
Considering they had people in marketing answering questions about performance, it goes to show that AMD has no answer. It would of been refreshing for them to just admit they didn't hit the performance target they were shooting for. There's no excuse for previous generation x6 to outperform their 'latest and greatest' x8.
 
And the saddest answer award goes to:

Buy our new over-priced, under-performing processor because it's a more modern architecture!

The kicker is if someone would have included why buy either when I can go to a Microcenter and get a 2500k for $170 which stomps over both.

That's a BINGO!!!
 
I get the jist of their answers as saying wait for a few years before judging the performance of this CPU because thats when the applications we built it for will be here. Thats all fine and dandy but they seem to have forgotten we live in the present and not a few years into the future. They should have implemented them changes when those apps are actually here.
 
I have the same feeling as the others. They took an opportunity to be completely honest and turned it into another marketing opportunity to deflect or promote their product(s).
 
"It is also important to note that the "Bulldozer" architecture is configured and optimized for server throughput. The two integer execution cores present in Bulldozer are designed to deliver area- and power-efficient multi-threaded throughput."

Thank you and good night. At least they came out and said it. It would certainly explain some of the benchmarks we've seen. I really hope they start to focus a lot more on the single-thread aspect of computing in addition to multithreaded software/hardware. I *hope* they've realized that general use in the desktop encounters a whole lot of single-threaded, poorly threaded and legacy applications, but judging by these answers I'm not very hopeful and can see Piledriver being another disappointment.

I think their other answers pretty much said if you want single-threaded performance go buy Intel.
 
Lol, when I saw the benchmarks for BD I just grabbed a bigger cooler for my 1100T and kept right on trucking, I am still tempted to upgrade to a 990FX mobo but the CPU seems like a waste of time until it scales past say 4.4ghz stock (may seem anal prestidigitation on that number, but based on their comments and a shitty wild ass guess, plus its 5-10% more than I can get out of this 1100T).
 
I have the same feeling as the others. They took an opportunity to be completely honest and turned it into another marketing opportunity to deflect or promote their product(s).

This pretty much sums it up, after being pulled along by marketing for many months that IPC is faster, and so on... it just another kick your customers and tell them how stupid they are day @ AMD for not bowing down to the BD.


:rolleyes:
 
In their defense, someone has to open new markets for PC programming or the market stagnates. Giving programmers a way to differentiate themselves from the competition is great for the industry I'd think.

Game A is multithreaded and runs 15% faster on newer chips but just matches Game B on older chips. Game B is single threaded and struggles to meet the 60 fps standard consistently. Both are fun but, Game B sometimes has frame issues so multiplayer contestants are penalized whenever it happens. Which game would you purchase? Which developer would you start paying attention to?

For the above scenario to happen, someone has to put the technology out there to be utilized. I seriously doubt a company would invest resources into new technology without something to showcase it on. It's definately not perfect technology as anyone that owns BD chips can attest. But it's not a bad technology as I see the point in it.

Well I like shiny things.

:)

Your argument is valid but the implementation was not. They left IPC and single-threaded performance behind in an attempt to usher in something revolutionary and 64bit-esque. But where the amd64s offered fantastic performance in both 32bit and 64 bit applications the bulldozer lacks in the single-thread department where it falls behind the Phenom II's, lest we mention intel's offerings. As a result it's been underwhelming as a choice for a desktop CPU and fanboy-ism aside, you'd have little reason to purchase one outside of running a handful of applications that very few people really use. (These applications also tend to be desktop workstation apps and they'd benefit more from a 2011 platform than bulldozer, unless, of course, price was an issue. And even there the bulldozer often falls flat on its face when compared to the 2600k price-performance wise [so too do the sandy-bridge E chips])

Let's put it this way: if the amd64 came out and offered poor performance on 32bit but great performance on 64bit, would it have been such a success? I sure as hell wouldn't have paid 230+ for an AMD64 3000+ and the Netburst fiasco would have been thought of completely differently (Good on 32bit applications AND it heats my home? SWEET). That's what the bulldozer and its "improvements" are akin to. It's always cool introducing something new and revolutionary and that shouldn't be understated, but unless you've got the software and performance increases to back it up then it's unnecessary.

All of my bitching aside, if they priced the 8150 at $160-170, I'd buy it. At that price point the only real issue I'd have with it would be the power consumption.
 
This pretty much sums it up, after being pulled along by marketing for many months that IPC is faster, and so on... it just another kick your customers and tell them how stupid they are day @ AMD for not bowing down to the BD.


:rolleyes:

I don't really see it as such, I take their comments of "it is optimized for servers/supercomputers" at face value, drop a pound of salt on the "IPC decisions are based on looking forward to newer processes (I.E. it shines at higher clocks)", and say wait and see.

I have an 1100T, most people I know have SB, neither of us has any reason to waste so much as a second saying "what if" about a BD build, its a server chip they released on the consumer market, and basically said unless you overclock the shit out of it don't waste your time to us gamers.

At least its not RAGE, we didn't all pre-order some pile of shit, we had the opportunity to see the benchmarks, collectively go ":meh", and back to playing BF3.
 
I have the same feeling as the others. They took an opportunity to be completely honest and turned it into another marketing opportunity to deflect or promote their product(s).

Honestly, what do they gain by admitting a fault?
Nothing positive to anyone other than maybe a neutral feeling by a very small market share.
 
AMD said:
For example, on parallel DirectX 11 gaming titles like Civilization V and Metro 2033, the AMD FX-8150 outperforms the Core i7-2600 (both with a AMD Radeon™ HD 6970 graphics card) by up to 18 percent and 8 percent, respectively.

This seems to contradict this.
 
If Bulldozer smashed 4 and 6 core CPUs in multithreaded benchmarks, I would be OK. The biggest problem is, even in heavily threaded scenarios, Bulldozer lags with/behind its 4 and 6 core competition, both from AMD and Intel.
 
Same answer techs gave to management that got them in this mess in the first place.
No amount of PR will fix the watts/performance ratio.

Higher frequency than intel for same performance or lower. Even when windows 8 comes out or a new
Process shrink to match or better performance, Intel will be 2 step ahead.

Management needs to shake things up. Maybe take inspiration from their graphics department that seems to be doing better.
 
Seriously, what did you expect them to say in their answers?

Personally, I just don't get the purpose of trying to sell "forward looking" designs? Why do you develop and market a design that won't be relevant for several years?

They did the same thing before, but at least designing a 64bit chip did not cripple it's 32bit capabilities.

But the answer that get's me most, is claiming a higher IPC per watt....that's a good one.
 
Zarathustra[H];1038081408 said:
This seems to contradict this.
notice AMD said "up to"

That means they found the perfect storm of resolution, details settings, game map, and weather conditions to demonstrate that performance improvement
 
Wow, all of that marketing fluff made my head hurt. BD was a flop. End of story.
 
It seems AMD marketers are on the counter spin offensive big time. If AMD would put as much effort into engineering as their marketers are doing to try to resurrect what they have now things would be different.
 
Last edited:
Back
Top