So.....what are you really expecting from Zambezi cpus?

[dma0991]

what if that translated to less then what a high clocked sandy could do even with 8ghz out of the box.

Then I'll get a SB instead but then again SB-E is not within the amount of money I'm willing to spend.

If I knew that BD is not going to be as spectacular as it was hyped I would have bought a SB rig like 9 months ago. Made a bad decision and too late in the game to be considering SB. Ivy Bridge perhaps.

 

[aigomorla]

You sure they had fun on final bios and final retail stepping and silicon just like JFAMD said?

well lets say this..

They broke the WR for clocks... but not superpi, or any computational benchmark?



usually one would go with the other unless the GHZ scale was off no?

Dont get me wrong... if ur a new system shopper looking to have fun in overclocking, they will be fun.

But speedwise... well, you can see the coming numbers as they speak for themselves.

 

[Idontcare]

My friends said, yes they clock like mad... but whats the point when the speed dial is the only thing moving, while the odometer is at a slow crawl?

LOL, so you're saying they are red-lining the engine, taking it to 11, but the transmission is limited to 1st gear only, so they are still only moving along at about 35mph tops

Why AMD!? Why!?

First you bad-mouthed Intel's "glued together chips" only to then do MCM yourselves. Now you've decided to revisit the P4 D:

WTF is next? Rambus ram?

 

[aigomorla]

in all honesty, id hope the current X6 prices fell due to BD.

And if i needed a cheap system, id get a PH X6 for core whorage..

I think that is what im going to do for my friend who needs a cheap system with a lot of working threads.

 

[grimpr]

Then something is borked, deliberatly or not, and it will be fixed.

 

[gmaster456]

I expect it to be between sandy bridge and Phenom ll. And given the price point, that sounds reasonable.

 

[aigomorla]

Then something is borked, deliberatly or not, and it will be fixed.

i am hoping this which is why they have another delay.

 

[RussianSensation]

well lets say this..

They broke the WR for clocks... but not superpi, or any computational benchmark?



usually one would go with the other unless the GHZ scale was off no?

In 2011, AMD doesn't need to show benchmarks. They simple state it as a fact.

http://www.marketwatch.com/story/correction-amd-showcases-worlds-fastest-cpu-2011-09-13

"The record-breaking processor speed that resides in the AMD FX CPU clearly demonstrates performance gains for the new AMD 'Bulldozer' multi-core architecture, which will provide x86 computing power for this CPU and future AMD Accelerated Processing Units," said Chris Cloran, corporate vice president and general manager, Client Group at AMD. "Along with world-record frequencies, the AMD FX processor will enable an unrivaled enthusiast PC experience for the money -- extreme multi-display gaming, mega-tasking and HD content creation."

 

[BlueBlazer]

http://www.marketwatch.com/story/correction-amd-showcases-worlds-fastest-cpu-2011-09-13

"The record-breaking processor speed that resides in the AMD FX CPU clearly demonstrates performance gains for the new AMD 'Bulldozer' multi-core architecture, which will provide x86 computing power for this CPU and future AMD Accelerated Processing Units," said Chris Cloran, corporate vice president and general manager, Client Group at AMD. "Along with world-record frequencies, the AMD FX processor will enable an unrivaled enthusiast PC experience for the money -- extreme multi-display gaming, mega-tasking and HD content creation."

I call that "smoke and mirrors"....

 

[Arkaign]

Heh, all I want is a chip that starts at $150 that can OC to ~2500k stock gaming performance. That'd make a great option for single-GPU gamers. Anything less than that will be a failure imho.

Nobody with a brain I think expects BD to hang with 2500k and 2600k at commonly used clock speeds of 4ghz+ in anything outside of very heavily multithreaded scenarios.

 

[Idontcare]

In 2011, AMD doesn't need to show benchmarks. They simple state it as a fact.

http://www.marketwatch.com/story/correction-amd-showcases-worlds-fastest-cpu-2011-09-13

"The record-breaking processor speed that resides in the AMD FX CPU clearly demonstrates performance gains for the new AMD 'Bulldozer' multi-core architecture, which will provide x86 computing power for this CPU and future AMD Accelerated Processing Units," said Chris Cloran, corporate vice president and general manager, Client Group at AMD. "Along with world-record frequencies, the AMD FX processor will enable an unrivaled enthusiast PC experience for the money -- extreme multi-display gaming, mega-tasking and HD content creation."

mega-tasking? really now the marketing is just getting silly, when did I miss the era of kilo-tasking?

 

[Accord99]

mega-tasking? really now the marketing is just getting silly, when did I miss the era of kilo-tasking?

Mega-tasking!?

Haven't heard that term since...

http://www.amd.com/us/press-releases/Pages/Press_Release_114483.aspx

 

[LOL_Wut_Axel]

Like I said before, anywhere from 5% slower IPC to Llano to 5% higher. Don't expect anything higher than that; it's wishful thinking. AMD is pricing Bulldozer so they can sell in high quantities, not for high performance.

Here's the facts:

• The top-end FX-8150 will be priced at $260.
• It has a 3.6GHz base clock speed and eight cores.
• Bulldozer's architecture is made in a way that you lose performance when a second thread is being run on the module, by way of getting 180% speedup in comparison to 200% with two cores, according to AMD.
• Two Anandtech editors have already said they expect IPC comparable or even worse than Llano.
• John Fruehe has downplayed the importance of IPC for desktop workloads.

What this gets us at: Bulldozer was made to be efficient in many-core systems like servers, and it can only get acceptable performance in desktops by having extremely high clock speeds, which is why AMD is hyping its high frequency capabilities.

Bulldozer loses to Sandy Bridge in single-threaded due to low IPC. Given that, AMD made many revisions to make sure it can reach high clock speeds without being a huge power hog. High clock speeds is the only way they can save face in single-threaded applications, even if there it's still significantly slower than Sandy Bridge. Likewise, many cores is the only way to save face for the fact that a module has a 180% speed in comparison to 200% from two normal cores (according to AMD). Since a Bulldozer module is much slower than two Sandy Bridge cores, AMD had to put 2x as many there. Even then, it'll still only get close in multi-threaded. AMD knows that no smart person will buy it knowing it's slower overall, so they priced it lower while throwing in overclocking capabilities as icing on the cake.

 

[grimpr]

An overclocking fu**toy, what a shame, still i'm not convinced that its a flop, something goes beyond total crap, transcends to fixable and soon enough.

 

[BlueBlazer]

mega-tasking? really now the marketing is just getting silly, when did I miss the era of kilo-tasking?

That's the previous Quad FX (aka "QuadFather") marketing "catch-phrase".

 

[hooflung]

I am expecting a performance for integer operations to be somewhere on the level between Phenom II and Sandy Bridge. What I also expect is idle power consumption to go down and better power features. With updates to the SIMD extensions and VM extentions I hope we see more software that can actually take advantage of said machine to demonstrate its strengths.

That is the problem I see with most people who 'just want gaming.' I can do what I want with my current system and see no real game that can put me out in the near future. So what I'd like to see and expect is something that can make workloads in virtual machines smoother and vector as well as encryption engine instructions to go faster. With less power use at idle. Did I mention less power use at idle.

I also like AMD based mobo's better. I'd also like to see Nvidia make a new chipset for the AM3 platform... or hell... bring back the 980a upgraded! My 780a is still quality. But that is for another discussion.

 

[Avalon]

I think the IPC of BD is going to be 5% better than Phenom II, so the clocks and core count will make it interesting.

 

[trollolo]

i herp and derp all over the place every time i see 300$ sandy bridge chips compared to 150$ amd chips and so on and so forth. i can understand if you need a top of the line CPU for your job, but if it's just recreational, AMD can do what you need for much less

 

[exar333]

i herp and derp all over the place every time i see 300$ sandy bridge chips compared to 150$ amd chips and so on and so forth. i can understand if you need a top of the line CPU for your job, but if it's just recreational, AMD can do what you need for much less

It's your fault if you don't read or look at prices. I can buy a 2500k for <$200, which is only slightly more than AMD's top X4 or middle X6. The 2500k wins hands-down.

Stop with the 'AMD is cheaper' crap. Yes, AMD is better <$100, is competitive a little above this, and is a waste of $$$ >$150.

 

[VirtualLarry]

Let's put it this way - is it an easier route to get to the "5Ghz club", by using a BD, rather than SB? Could that not be a drawing point to get customers, enthusiasts that like to OC?

Even after seeing some of the pre-retail chip benchmarks (disappointing though they appear to be), I'm still moderately interested in BD. Maybe not the top-end chip, but something lower. Still debating whether the 4-core (2-module), or the 8-core (4-module) is the better buy, if they all have unlocked multipliers.

I still don't know why I want one, possibly bragging rights. I almost hopped on a Phenom II X6, as the first affordable consumer hex-core chip, but I skipped that one, and I'm fairly glad I did.

The thing is, now, I don't do anything remotely computationally-expensive (well, possibly F@H), so I would just as easily be served by a Zacate, or possibly a mini-ITX G530 rig.

 

[infoiltrator]

The most marketble chip AMD can put together, probably will do some things better than SB and some worse. Probably not the top performance leader but more capable across many functions/uses.

 

[inf64]

Guys,by the look of things right now,5&#37; IPC increase over Llano would be enormously positive news. I'm afraid this won't happen.Not even in single thread mode when the one core has all the shared resources to itself. I don't know what they borked inside this novel(on paper) design,but it won't be that good I'm afraid. I think most reviews will label is out right bad due to abysmal integer and floating point performance.Just wait and see.

PS 8C ,or 4 module,whatever you prefer @ 3.6Ghz barely beats 3Ghz X4 Deneb in 3dmark06 CPU subtest and beats X6 1100T in C11.5 by 1%. Yes you read it right,1%. Both tests are multithreaded and should show good points of BD design(but they don't). Not good at all.

 

[LOL_Wut_Axel]

The most marketble chip AMD can put together, probably will do some things better than SB and some worse. Probably not the top performance leader but more capable across many functions/uses.

That's actually the opposite of what it should be, at least on the things done better or which it'd be capable of. What we're basically looking at is a CPU that compared to the Core i5 will have good multi-threaded performance but will be a dog when it comes to single-threaded.

 

[podspi]

inf64, if that was true, wouldn't Cray refuse shipment?

Edit: Wow, Vesku said almost the exact same thing in another thread... Great minds...

At least I know I'm not the only person who finds it hard to believe BD is such a hard fail that AMD is slipping backwards (and that their partners don't care while AMD outright lies about it).

 

[SlowSpyder]

My guess is IPC better than Phenom II but still somewhat behind Nehalem (and obviously Sandy Bridge then). 4.6GHz+ overclocks and good pricing.

Just a stab in the dark.