AMD chief say BD will offer only 35% not 50% more performance than previous gen

[Nemesis 1]

That's where you went wrong. The base in a math equation isn't always the lowest number, it's the point of reference that you start at. If A uses 87% more than B, B does NOT use 87% less than A because in the former, A is the base, while in the latter, B is the base. Let me explain it to you with a simple example:

If CPU A consumes 50 watts and CPU B consumes 150 watts, you can say that CPU B consumes 200% MORE power than CPU A, which is NOT the same as saying CPU A consumes 200% LESS power than CPU B. I hope you realize that it is IMPOSSIBLE to consume 200% LESS power.

In this example, you can say CPU B consumes 200% MORE power than CPU A or you can say CPU A consumes 66.7% LESS power than CPU B.

Yep and I demonstraited that plainly with the core counts. But when you talk power usage the base is and always be the best result or best effiency it has to be done this way .

Hear is the proof . AMD 300 watts intel 150 watts . now give me both sets of numbers. that we can reach . After you look at them you will see why it has to be done 1 way only.

The only thing that can change is the wording . But that not possiable in this example . Because the best result has to be used as the base the worse result can never be used in the case of best power usage . even tho its 6cores Vs. 4 . Now If we wanted to break down each cores efficiency than the other number is correct maybe I didn't do the math . . But thats not what were doing . We are comparring 6 cores to 4 cores . In other words were talking DIES. If we figured cores than you have to figure work completed / Time / power consumption = efficiency/Horsepower and torque

 

[Nemesis 1]

If CPU A consumes 50 watts and CPU B consumes 150 watts, you can say that CPU B consumes 200% MORE power than CPU A, which is NOT the same as saying CPU A consumes 200% LESS power than CPU B. I hope you realize that it is IMPOSSIBLE to consume 200% LESS power.

In this example, you can say CPU B consumes 200% MORE power than CPU A or you can say CPU A consumes 66.7% LESS power than CPU B.
__________________

Think this thing threw guys. YES I know what your saying . But I spent to many years as a tech to let something like this slip threw . In your example . YOU are talking apple and grapes. Let me give ya a couple of min. here . Look at what ya wrote. relax. and you'll see it.

 

[AtenRa]

I see you refused to ans the question . Puts taill between legs and runs as he new he was going to get hammered. You relized AMD BD has shared resources and if you can only use one core kind screws the performance of shared resources now doesn't it.

Each core has 1 128bit FPU. kind bites doesn't it. So your sharred resources just won't go away and you can not compare single core results as performance would take a big hit . I seen ya used the thread word . But you used the single core only. Ya see lmpw matter what. You have to compare 1 module 2cores to intels 1 cores. You can't compare a single core SB to single core BD as the performance takes pretty big hit ; So no matter what We will allways comparred 2 intel cores to 1 AMD module . But we can do single thread apps. But single BD core against SB core will never work as 2 cores are required in resource sharring .

You really need to read more about Bulldozers Module,

BD module has two(2) Integer Execution Units and one shared FP execution unit (2x 128bit FMACs).
Edit: AMD counts as a core the Integer Scheduler + Integer Execution Unit + L1 DCache

A single BD core consist of ONE Integer Execution Unit and a shared FP execution unit (it can use ONE or BOTH 128bit FMACs).

When we have a single thread only ONE core (one Integer execution unit + FP execution unit) will be used within the Module.

So it doesn't matter if the front end or the FP unit is shared because at a single thread we dont have any kind of sharing inside the BD Module.

You have to understand that BD will not share a single thread between the two cores (Integer Execution Units) inside the Module, one thread will use one Integer unit and a second thread will use the second Integer unit.

So, even if the Module has two INT cores, if we only have a single thread then only one INT core will be used.

At the end of the day, yes we can compare a single BD core to a single SB core in a single thread

 

[Imouto]

Where does the 2600K fall in your performance scaling list? Or SB-E? Or IB?

Do you really think that HT adds a lot of processing power or something? Damn, Intel did their homework with the propaganda machinery.

http://www.elitebastards.com/index.php?option=com_content&task=view&id=705&Itemid=27&limitstart=8

Wait, wait. A mere 10% increase at max with 4 cores in when it matters at all? Damn dude, you got me. Let's add a 15% (lets being biased like you do) to the chart:

PhII = 1
Sandy Bridge = 1.5
Sandy Bridge + HT = 1.65
BD = 1.35
Sandy Bridge * 4 cores = 6
BD * 8 cores = 10.8
Sandy Bridge * 8 threads = 6.6 (2600K)

BTW, all those performance gains over PhII include HT, just helping with your daydream because no one and no review or any other shit will say that Sandy Bridge is 65% better overall than PhII.

You can find a lot of reviews about Hyper Threading performance, next time just read them.

 

[Riek]

I am embarresed for both you guys , Shameful display of math skills,

I not going to explain it to you .My 4 year old grandson has better math skills. here is the proper usage to get correct ans ,

Think this thing threw guys. YES I know what your saying . But I spent to many years as a tech to let something like this slip threw . In your example . YOU are talking apple and grapes. Let me give ya a couple of min. here . Look at what ya wrote. relax. and you'll see it.

No he doesn't. They are completely correct and you are completely wrong and off base. To be really frank, go get your 4year old grandson and let him explain things to you.

It is correct in every situation to use the highest or the lowest as the reference model. That is the basis of mathematics. Depending on how you want to compare you can use any of the 2 parts as reference material. It is perfectly allright to say some performs x% less or you perform y% higher. What isn't allright is to mix those numbers!!! which was done a few posts ago...

 

[dma0991]

Do you really think that HT adds a lot of processing power or something? Damn, Intel did their homework with the propaganda machinery.

You can find a lot of reviews about Hyper Threading performance, next time just read them.

You must be new here.

 

[Imouto]

Yeah, I'm new here but I really can't stand ppl who think that the SB's 50% advantage over PhII is increased a way more by using HT.

HT is enabled in every single review out there and the 50% advantage includes HT performance.

 

[Nemesis 1]

If you change the base on the 300 / 150 the results are 200% and 50%. In one example one cpu uses 200% thats the number you get 200% . Now look at the 2 numbers

we know that 300 is 2x 150 so 200% represents 2x150
We also know that 50% of 300 is 150. . so the 300watts and the 150 watts represents AMD / intel in this order. So whats going on .

How much power is used AMD 300watts . Intel 150 watts So intel uses 150 watts less or 50% less than AMD No arguments right.

The 50% represents how much less power intel uses . So now were talking 50% .. Your likeing this aren't ya . I am proving you right aren't I . Sure looks that way .

The 200% represents how much more power AMD is sucking . NO arguements right.

So i just proved myself and russian wrong didn't I . No arguments right .

Apples Intels 50% represents intels power from the wall . over amds power from the wall .

grapes AMD uses 300watts . thats 200% morepower or 300 watts

SO if we take .5 which is 50% 0f 1 . and multiply 300watts by .5 we should get 150 watts . Do you see ! In 1 instasnces were talking percentage differance of total power used . In the other were talking AMDs about the 150 watts more that AMD uses comparred to intel .

 

[Imouto]

You passed your elementary school for real?

 

[Nemesis 1]

You really need to read more about Bulldozers Module,

BD module has two(2) Integer Execution Units and one shared FP execution unit (2x 128bit FMACs).
Edit: AMD counts as a core the Integer Scheduler + Integer Execution Unit + L1 DCache

A single BD core consist of ONE Integer Execution Unit and a shared FP execution unit (it can use ONE or BOTH 128bit FMACs).

When we have a single thread only ONE core (one Integer execution unit + FP execution unit) will be used within the Module.

So it doesn't matter if the front end or the FP unit is shared because at a single thread we dont have any kind of sharing inside the BD Module.

You have to understand that BD will not share a single thread between the two cores (Integer Execution Units) inside the Module, one thread will use one Integer unit and a second thread will use the second Integer unit.

So, even if the Module has two INT cores, if we only have a single thread then only one INT core will be used.

At the end of the day, yes we can compare a single BD core to a single SB core in a single thread

I don't need to know anything about it . You best reread . It was said 1 core . What ya going to do with sharred resources that in multi thread cause 20% performance hit . Now all of the sudden were talking more than 1 core were talking sharred resources colored in purple . Cores are differant color . I not the one who started this, I said the shareed resources caused a performance hit . Than ya blew up and said 1core/ thread/. Good luck with that . NOW your going to beat me up over not allowing ya to use more than 1 core now you want to use the sharred resources that started this to beginn with . I thought when ya said that ya goofed. So in fun I thought a would run with it.

So when you guys talk cores size make sure ya include everthing you need to play a game in single thread mode. other than l3 caches , and controllers. THE FPU unit of 128 bits is core. or we going to play reality games

 

[Idontcare]

Do you really think that HT adds a lot of processing power or something? Damn, Intel did their homework with the propaganda machinery.

http://www.elitebastards.com/index.php?option=com_content&task=view&id=705&Itemid=27&limitstart=8

Wait, wait. A mere 10% increase at max with 4 cores in when it matters at all? Damn dude, you got me. Let's add a 15% (lets being biased like you do) to the chart:

PhII = 1
Sandy Bridge = 1.5
Sandy Bridge + HT = 1.65
BD = 1.35
Sandy Bridge * 4 cores = 6
BD * 8 cores = 10.8
Sandy Bridge * 8 threads = 6.6 (2600K)

BTW, all those performance gains over PhII include HT, just helping with your daydream because no one and no review or any other shit will say that Sandy Bridge is 65% better overall than PhII.

You can find a lot of reviews about Hyper Threading performance, next time just read them.

What's with the immature attitude? I asked a simple question, you fire back like a 13yr old who's dad works for AMD or something.

Check it at the door, you won't be long around here walking around with that chip on your shoulder.

 

[Idontcare]

Yeah, I'm new here but I really can't stand ppl who think that the SB's 50% advantage over PhII is increased a way more by using HT.

HT is enabled in every single review out there and the 50% advantage includes HT performance.

You are making this needlessly personal, directly against the posting guidelines.

Get over your ego or you will be shown the door, you may be new but you better figure out the posting guidelines quick if you plan to have personal issues with everyone who has an opinion differing from yours.

 

[Nemesis 1]

You passed your elementary school for real?

No actually they did infact Have to push me all the way threw elementry till we got to science and history and religion . The rest of the crap > didn't do . I just refused. But in the 3 I liked I was at the top . Ya I see what I did . Its ok been up for 3 days . Point is Intel has 2x better power usage . We well play this game again when we figure BD performance increases. You know I am going to grabb onto the smaller number . When ya complain I will bring this thread back and let the mods settle this. I getting a little punch drunk here. Man I can't even rember wgat I was going to ask . OH ya . Were did I get that 200% from . Sorry for that I rounded the numbers to 300 and 150 so rather than use the 87 % number . It would be 100% more watts. . Any way russian got sleep he fix this . I have to see how I side tracked myself with that 200%

 

[AtenRa]

I don't need to know anything about it . You best reread . It was said 1 core . What ya going to do with sharred resources that in multi thread cause 20% performance hit . Now all of the sudden were talking more than 1 core were talking sharred resources colored in purple . Cores are differant color . I not the one who started this, I said the shareed resources caused a performance hit . Than ya blew up and said 1core/ thread/. Good luck with that . NOW your going to beat me up over not allowing ya to use more than 1 core now you want to use the sharred resources that started this to beginn with . I thought when ya said that ya goofed. So in fun I thought a would run with it.

So when you guys talk cores size make sure ya include everthing you need to play a game in single thread mode. other than l3 caches , and controllers. THE FPU unit of 128 bits is core. or we going to play reality games

I cant understand what you want to say but i will try to show you with pictures in order for more people to understand it too.

When we have one thread, AMD Bulldozer module will only use the Front end (Fetch, Decode) + Core 1 + the entire FP unit if possible (one or both 128bit FMACs) + L2 + L3.

At that situation we dont have shared resources between Core 1 and Core 2 and the Single Thread can use all the resources of the Module, except it cannot use Core 2.

It's like as if we had a single Core

If we have two(2) threads in the same module then both threads will share the Front End(Fetch, Decode), they can share the FP or one of the threads can still use the entire FP unit and they share the L2 and L3 cache.

They never share Core 1 and Core 2, thread No1 will be executed in Core 1 and Thread No2 will be executed in Core 2.

 

[Nemesis 1]

No he doesn't. They are completely correct and you are completely wrong and off base. To be really frank, go get your 4year old grandson and let him explain things to you.

It is correct in every situation to use the highest or the lowest as the reference model. That is the basis of mathematics. Depending on how you want to compare you can use any of the 2 parts as reference material. It is perfectly allright to say some performs x% less or you perform y% higher. What isn't allright is to mix those numbers!!! which was done a few posts ago...

Yes Reik have it your way . Glad mods are reading this . Ya I did get off base . But I was likely dragged off . Just remember your words here Reik . When BD is released and I use the wrong base on purpose to use the smaller % gain that it doen't matter its a choice kind of thing. and I will bring friends to add to the party. And I said it out in the open . Not going to play stupid or play the mods. These guys have a pretty crappy job and I can be a real butt pain. But I don't not agree with your reply at all Russian had it right period. Now if you will excuse me . I have to find were I pulled that 200% out of my arse from . Talking 87% and suddenly I using 200% rather than the 100% I was trying for.

 

[Nemesis 1]

Ok found out were I got the 200% . From the dam math I just forgot to subtract intels 100% from 200% . Ding dang Wondering why I couldn't show why the base has to be the lower powered unit . Using 200% . I would slap myself but I already in to much pain .

 

[Nemesis 1]

I cant understand what you want to say but i will try to show you with pictures in order for more people to understand it too.

When we have one thread, AMD Bulldozer module will only use the Front end (Fetch, Decode) + Core 1 + the entire FP unit if possible (one or both 128bit FMACs) + L2 + L3.

At that situation we dont have shared resources between Core 1 and Core 2 and the Single Thread can use all the resources of the Module, except it cannot use Core 2.

It's like as if we had a single Core
Yes I got it along time ago I mean we all did I would guess . I told ya you goofed on your wording and I took advantage of it in fun .

You took defensive position because I brought up that in multithread apps there a 20% hit , So you said single core / thread. So I thought I would play with ya . I really don't want to make ya mad. Just it was early we were lonely and what the hay . Now lets kiss and make up . LOL. Bob will be back in the states tuesday and I know he is going to tell me everthing. So if I really good mood after tuesday. Be sad . If I grumpy Be happy.

 

[RussianSensation]

Nemesis, take off the blue class and you will see the correct math.

2500K has 46% LESS power consumption than PHII X6

PHII X6 has 87% MORE power consumption than 2500K

Like it or not, that's the correct expression

Yup, I saw that I worded it incorrectly. Thanks for correcting my math into English phrasing. But PHII X6 drawing 87% more power is still awful. Phenom II X6 draws more than 140 watts of power extra, despite still being unable to beat the 4 core overclocked 2500k.

It looks like I am the only one having such a hard time believing that an 8-core BD with an IPC of Nehalem and overclocking of SB, that also draws as little power as a SB, is impossible. I won't for a second believe that AMD was suddenly able to produce an 8 core processor with high clock speeds which draws half the power per core that a Sandy Bridge does. FX-8000 series will draw more power and have lower IPC than a 2500k/2600k in overclocked states. That's my view considering how far behind they are and have been all these years.

I'm sorry, but what are you rambling on about? The Core i5 2500K is faster in gaming than the Core i7 990X and yet I don't see you complaining one bit. There's more to a CPU than its gaming performance. If we're to follow your arguments, a Core i5 2500K is faster than the Core i7 990X in a real-world scenario, and is therefore faster and therefore the 990X "sucks", right?

You never asked me to comment on the 980/990X, so why would I voice my opinion on them?

Now that you asked, the 98/990X are a waste of $$ for most people (hence their $999 niche price). They target people who use professional / design / media applications at home or in the workplace. With Quick Sync in place, you can now transcode video very fast onto your smartphone or tablet. The slightly reduced video quality isn't that important for smaller devices.

With 2500k and 2600k, socket X58 became obsolete for 99% of people in January 2011. I never said 980X or 990X were good CPUs btw. Not sure why you even assumed I did. I would take an overclocked 4-core SB for $225 over often slower $999 6-core i7-990X any day.

AnandTech's review:

"....the Core i5-2500K is absolutely a steal at $216. You're getting nearly $999 worth of performance at roughly a quarter of the cost."

Not sure what you are on about. For most consumers, 2+HT / 4 fast cores > 6/8 slow cores. Notice what CPUs Apple puts in their systems? Ones with fastest performance / watt (and also the best IPC). Consumers don't really need slow 6 or 8 core processors, despite what AMD's marketing department wants them to believe (i.e., More cores is better). That's why LGA2011 is going to target workstation/niche market segment and "enthusiasts" who actually do need 6C/12T.

Quad-cores are only expected to reach mainstream by 2015 in the notebook space. Notebook sales now surpass desktop sales. You can purchase plenty of quad-core laptops today. So why is it going to take another 4 years before quad-cores are found in at least 50% of laptops? One of the reasons is that most consumers simply aren't willing to pay a premium for cores they won't use. They won't sacrifice smaller/thinner laptop form-factors at the cost of reduced battery life for a faster 4-core CPU when a 2-core + HT suffices. Now try selling those same people slower IPC 6- and 8-core processors (which are likely to have worse power consumption at load too) against a fast 4-core processor.

In the context of the new FX CPUs, if your main focus is gaming and gaming only, you should've slashed the FX-8000 series off your short list by now.

Do you know why Athlon 64 / X2 were so successful (and why we all respected them so much)? You know why so many of us bought and recommended A64/X2 processors? Because it was the best CPU for gaming and achieved amazing power consumption while at it. :wub:

"Note also that the 50W power consumption demonstrated by the Winchester core at 2.4GHz core clock is a very low value for contemporary processors. This fact gives us some reason to hope that the frequency potential of the Winchester core will turn out pretty significant. For a more illustrative comparison we also measured the power consumption of the Pentium 4 processors based on Northwood and Prescott cores and working at 3.4GHz core clock. The results turned out simply impressive: under maximum workload Pentium 4 processor on Northwood core consumed about 100W of power, while the Prescott based CPU (with a C0 core stepping) required about 132W.
" - Source

The primary reasons why Athlon 64 / X2 dominated Intel was because:

1) They had better overclocking headroom in % terms;
2) They had better IPC
3) They had better performance / watt.

For BD > SB, it would have to be superior in at least 2 of these 3.

Did you not notice what forum you are on? How many people here have GTX570/580/HD6950/5870/6970 GPUs and countless SLI/CF setups?

You know how many people play WOW and SC2 --> Strategy and MMOs are very CPU intensive. Strategy games accounted for almost 34% of all PC game sales last year. Gaming performance is important.

 

[RussianSensation]

PhII = 1
Sandy Bridge = 1.5
Sandy Bridge + HT = 1.65
BD = 1.35
Sandy Bridge * 4 cores = 6
BD * 8 cores = 10.8
Sandy Bridge * 8 threads = 6.6 (2600K)

BD = 1.35x IPC increase over Phenom II? Wishful thinking.

BD module design provides 80% of the performance of 2 separate cores.

If you were to take a 2-core Phenom II processor and magically redesign it into a Bulldozer 2-core module, then you would have 80% of the performance of the same 2 independent core Phenom II design.

Now to realize your BD module = 1.35x performance increase estimate over a 2-core Phenom II at the same clocks, you'd need a whopping 69% increase in IPC. (1.35x / 0.80x).

Impossible. Intel didn't even increase IPC that much from 65nm Yorkfield --> 45nm Penryn --> Nehalem --> Westmere ---> Sandy Bridge.

But you think AMD will do that all in 1 generational leap? And on top of that give us 4.2ghz Turbo and 8-cores?

These are my alternative calculations:

If Bulldozer is 25% faster in IPC over Phenom II, then 1 BD module = 2 Phenom II cores in IPC (i.e., 0.80x penalty * 1.25 increase in IPC = 1.0x 2-core Phenom II). My feeling is that BD will achieve most of its performance gains from clock speeds and more cores. Its average IPC increase over Phenom II won't be more than 20% at the same clock speeds.

 

[LOL_Wut_Axel]

First thing, I don't really believe Bulldozer is going to be slower than Atom.

The thing is, there is no data. It could be slower than Atom, as it could be 10x faster than Sandy Bridge. This is my point, as is Bridito's point.

There is some data, and no, it can't be slower than Atom. That makes no sense; for that to happen it'd have to be much slower than Llano. Let's just put an end to this would-be argument and focus on discussing what performance could be instead.

 

[LOL_Wut_Axel]

*Extremely long rambling*

Alright, so you say that the 990X is made for a niche market, but then fail to acknowledge that's what the Phenom II X6 is for, too. Like I already said, if you're not doing video encoding, rendering or content creation, you should remove the FX-8000 series from your list. If the lower-end CPUs have good single-threaded performance then the FX-8000 series would be good for people that want to do it all.

As for gaming performance, you've based 90% of your arguments regarding performance around it. Gaming is an important thing, but it's not 90% of what most people do. That, and if you want to talk about IPC, you shouldn't mention it anyway. If you want to compare architectures use CPU benchmarks, not CPU+GPU benchmarks.

 

[RussianSensation]

As for gaming performance, you've based 90% of your arguments regarding performance around it. Gaming is an important thing, but it's not 90% of what most people do.

Not even close. 2500k beats the X6 in almost everything.

Read the reviews I linked. X6 loses to 2500k in archiving, image editing, audio processing applications. Even in rendering and video encoding, the X6 is barely better (and once you add in Quick Sync, X6 isn't better for video transcoding).

On another note,

8GB DDR3-1866 = $79.99 (cheapest on the Egg right now)

8GB DDR3-1600 = $40-55 (Egg)

Reviewers should do testing for current AM3+ owners who plan to do a drop-in upgrade to see how much the memory bandwidth matters.

 

[LOL_Wut_Axel]

Not even close. 2500k beats the X6 in almost everything.

Read the reviews I linked. X6 loses to 2500k in archiving, image editing, audio processing applications. Even in rendering and video encoding, the X6 is barely better (and once you add in Quick Sync, X6 isn't better for video transcoding).

On another note,

8GB DDR3-1866 = $79.99 (cheapest on the Egg right now)

8GB DDR3-1600 = $40-55 (Egg)

Reviewers should do testing for current AM3+ owners who plan to do a drop-in upgrade to see how much the memory bandwidth matters.

Those other applications you mention are also some where the Core i7 970-990X lose as well. I don't know what you're rambling on about, really. I already told you the Phenom II X6 is for rendering, video encoding, and content creation. Whether QuickSync exists or not is irrelevant for most Sandy Bridge users as most have P67 motherboards. Another thing is that we're talking about CPU performance, not performance from the IGP or integrated video encoder.

And on your "other note", I don't get your links. Bulldozer doesn't require 1866MHz RAM, it just supports it officially. RAM prices for both are the same, and Sandy Bridge doesn't officially support 1600MHz.

 

[ed29a]

Not sure what you are on about. For most consumers, 2+HT / 4 fast cores > 6/8 slow cores.

For most consumers a dual core SB Celeron or low end dual core Athlon II is more than enough.

Notice what CPUs Apple puts in their systems?

I see what you did there. Apple was never known for performance. Never. They always put chips in their systems that were (1) Not expensive (2) Low power usage. Just how you think they get those insane margins on their products?

Consumers don't really need slow 6 or 8 core processors, despite what AMD's marketing department wants them to believe (i.e., More cores is better).

Read my initial comment. Also, on the flip side, Intel's GHZ GHZ GHZ GHZ marketing was ... honest? Let's stay out of strawman territory, shall we?

That's why LGA2011 is going to target workstation/niche market segment and "enthusiasts" who actually do need 6C/12T.

It's going to target that market because (1) It's way more expensive than other platforms (2) The vast majority of people don't need anything near that amount of CPU power. Thus it's a 'niche' platform.

So why is it going to take another 4 years before quad-cores are found in at least 50% of laptops? One of the reasons is that most consumers simply aren't willing to pay a premium for cores they won't use.

You are seriously being delusional if you believe the consumer knows what a core is. They look at the machine: price good? Sale. In some cases like Apple, they look at the brand. Your average consumer is dumb as a rock when it comes to computer hardware knowledge.

They won't sacrifice smaller/thinner laptop form-factors at the cost of reduced battery life for a faster 4-core CPU when a 2-core + HT suffices.

Read my previous comment.

Now try selling those same people slower IPC 6- and 8-core processors (which are likely to have worse power consumption at load too) against a fast 4-core processor.

Next time you are around a Best Buy or some other electronic store, go in and listen to salesmen. They could sell fridges to eskimos. The average consumers will go in, get sweetalked into buying the most expensive laptop in their budget. Of course, they won't use 10% of the computational power this machine offers but hey, they got talked into by some salesman who is 'an expert' and 'knows' this stuff.

The primary reasons why Athlon 64 / X2 dominated Intel was because:

1) They had better overclocking headroom in % terms;
2) They had better IPC
3) They had better performance / watt.

For BD > SB, it would have to be superior in at least 2 of these 3.

Wrong. The main reason AMD dominated Intel is because Intell fell asleep and let marketing dictate the orientation of products instead of engineers. You know, the Gigahertz uber alles plan.

Did you not notice what forum you are on? How many people here have GTX570/580/HD6950/5870/6970 GPUs and countless SLI/CF setups?

You know how many people play WOW and SC2 --> Strategy and MMOs are very CPU intensive. Strategy games accounted for almost 34% of all PC game sales last year. Gaming performance is important.

Yet you keep talking about consumers. Newsflash: the average consumer doesn't know what Anandtech is. The average consumer cares about price and buys Dell machines. Guess what, nothing SLI/CF on it. Also, 34% of PC gaming is a drop in a bucket compared to consoles. Let's put things into perspective. While PC gaming has a few anomalies like WoW and SC2, the vast majority of PC game sales numbers are eclipsed by consoles. If you think the PC gaming market is so important, I got a bridge for sale you might be interested in!

PS: In my raid guild on WoW, maybe 3-4 out of 50 people have 'enthusiast' level machines. Rest are chugging along fine on 4-5 year old rigs with a midrange videocard.

 

[Abwx]

Not even close. 2500k beats the X6 in almost everything.

Read the reviews I linked. X6 loses to 2500k in archiving, image editing, audio processing applications. Even in rendering and video encoding, the X6 is barely better (and once you add in Quick Sync, X6 isn't better for video transcoding).

Actually, in archiving , you took a single threaded apps as WINRAR
that favour the 2500K , not counting that this apps is Intel optimised.

Why not take a popular multithreaded soft instead ?...

Let s try with 7zip....

As you can see, in such a soft, a 2500K is not better than a X4....