AMD CPU speculation... and expert conjecture

[mayankleoboy1]

Its been fine that developers have been limited forcefully or by necessity to archaic legacy because Intel says so, but why not optimized coding for AMD systems? When will there ever be a change for something new and with potential except the same old excuse is thrown around. Coders for EA and other developer companies have said we are stuck in a limbo and HSA provides a way of expanding beyond that, Some have said we are still in the 80's all that has changed is things look prettier yet there is the same old "you cannot do it" which is funny when you see those who can do it, it ultimately comes down to who gives a damn and has the skills.

Coders of EA and others were hampered by the console they are targeting as their primary platform.

APU =/= HSA. On PC's, there is only the APU so far. Where is the promised HSA ? Hd79xx was supposed to have basic unified memory, but it never happened. AMD has a history of promising a lot, and not delivering. For all the hype AMD have built, where is the damned hardware ?
Even in the PS4 chip, the only part of HSA is the unified memory/PCIE lanes. (Which is a big deed in itself, but falls quite short of what AMD has bragged about.). The same Jaguar chip, on desktops, lacks the unified memory. I seriously doubt AMD's commitment to desktops.

 

[hcl123]

The i7 980x is faster then the fx8350 still and that was an early 2010 cpu. Also the 4770k wins on around over 80% of the benchmarks against an fx8350. 15% nothing really, oc the 4770k to 4.2ghz and you get that 15% easily. Intel barely have any competition hence why they hold most of the market as they are that far ahead. Intel architecture is so far ahead of amd now I reckon 2016 will be the year they get to sandy bridge levels and that will take a gigantic leap from where amd are now. I think steamroller will get close to sandy bridge level though.

Intel has already plenty of competition, why ppl fail to see it is what is incredible.

In a client/desktop 15% difference is nothing really, you wont feel or see a damn difference... not even with 30%... EXACTLY AS YOU SAID... "15% nothing really"

Depends a lot on the software, as past discussions is possible to achieve an order of magnitude better performance optimizing the software than optimizing the hardware... that is why HSA is mostly software not hardware, (the GPU=CPU is a path that only AMD will take, its not needed for HSA software to run well, but it seems much better)... that is why ARM CPUs don't have not even close the GIPS or the FLOPS of x86 counterparts, yet its software is as blazing fast as in a 980x or 4770 or 8350 (much less bloated, the ISA helps get slimmer binaries, and it has better memory handling).

So i beg everyone pardon, but sometimes this CPU discussions are incredible stupid... like kids on epennis measurement, always trying to stretch an ridicule the others...

Experiment all, and in TRUTH tell the differences and the feel to yourself (not me or anyone else), and in the end what is that you gain... epennies bench s**t contests are for kids, its an *entertainment* business, not a science endeavor. ... what does it represent for you actual software( next to slim, cause benches are run one at a time *only*, not at all representative of how ppl use their computers) and what does it represent for your productivity.(edit)

 

[8350rocks]

Its been fine that developers have been limited forcefully or by necessity to archaic legacy because Intel says so, but why not optimized coding for AMD systems? When will there ever be a change for something new and with potential except the same old excuse is thrown around. Coders for EA and other developer companies have said we are stuck in a limbo and HSA provides a way of expanding beyond that, Some have said we are still in the 80's all that has changed is things look prettier yet there is the same old "you cannot do it" which is funny when you see those who can do it, it ultimately comes down to who gives a damn and has the skills.

Coders of EA and others were hampered by the console they are targeting as their primary platform.

APU =/= HSA. On PC's, there is only the APU so far. Where is the promised HSA ? Hd79xx was supposed to have basic unified memory, but it never happened. AMD has a history of promising a lot, and not delivering. For all the hype AMD have built, where is the damned hardware ?
Even in the PS4 chip, the only part of HSA is the unified memory/PCIE lanes. (Which is a big deed in itself, but falls quite short of what AMD has bragged about.). The same Jaguar chip, on desktops, lacks the unified memory. I seriously doubt AMD's commitment to desktops.

HSA is already in Kabini and Temash, and Kaveri brings it to desktop.

 

[hcl123]

Im not trolling if amd get to sandy bridge levels by 2016 they will have done well from where they are now...

You even said 15% its nothing... well FX is already quite below that, compared with SNB, with the current benchmark software (which is not much representative at all)...

But lets imagine you are an engineer with decision making power... does all that 500%, and even more in other examples, heck! even OpenCL on intel can be 250% better than a CPU ALONE !.. tells you anything ? (edit)

Forget its AMD (it seems to trigger some nerve), if you were the developer of BD uarches, IN ALL HONESTY, why would you be concerned in having the better "Integer single-thread" performance than your competition ?? ... aren't you building 12 or 16 cores CPUs for the next years even for mainstream client/desktop ??... aren't you building kind of APUS to have the possibility of that 250% or 500% or more ??

Does 10, 15 or 20% Integer single-thread performance matter ???

 

[hcl123]

Its been fine that developers have been limited forcefully or by necessity to archaic legacy because Intel says so, but why not optimized coding for AMD systems? When will there ever be a change for something new and with potential except the same old excuse is thrown around. Coders for EA and other developer companies have said we are stuck in a limbo and HSA provides a way of expanding beyond that, Some have said we are still in the 80's all that has changed is things look prettier yet there is the same old "you cannot do it" which is funny when you see those who can do it, it ultimately comes down to who gives a damn and has the skills.

Coders of EA and others were hampered by the console they are targeting as their primary platform.

APU =/= HSA. On PC's, there is only the APU so far. Where is the promised HSA ? Hd79xx was supposed to have basic unified memory, but it never happened. AMD has a history of promising a lot, and not delivering. For all the hype AMD have built, where is the damned hardware ?
Even in the PS4 chip, the only part of HSA is the unified memory/PCIE lanes. (Which is a big deed in itself, but falls quite short of what AMD has bragged about.). The same Jaguar chip, on desktops, lacks the unified memory. I seriously doubt AMD's commitment to desktops.

HSA is already in Kabini and Temash, and Kaveri brings it to desktop.

HSA features have been on AMD APUs since Llano. It goes by phases, there are 5 phases i think.

HSA *software* is a totally different matter. When there will be a *runtime* to download and or install, then is when it is said, that platform is ready for HSA software.

Its possible that HSA foundation builds a runtime for intel platforms, which i doubt, but for ARM, possibly with Android, that tries to follow all kronos standards is quite to expect.

Who would have the best hardware support most probably will have the best performance, the HSA hardware definition, the minimal that is needed for HSA software, is hUMA and task scheduling queues with that integrated virtual memory awareness(visible from user space application software) (edit)... and very little else... apart from having a Parallel (GPU) or heterogeneous Processor present (could be a DSP), which is the fundamental condition as obvious, but that all already do, AMD, ARM and intel etc (edit)

So GPU = CPU i think only AMD will path, and its not necessary. Probably intel will path nothing of that, having similar but not really compatible... so why bother building a runtime for them ?

But even so, most of those minimal hardware requirements can be *emulated* in software, it will run notably slower but it will run all HSA software in intel platforms.

That is what gives intel the cold sweat nightmares... that is what is behind the wasty decision of having broadwell mostly BGA, and expansion in dead end, grab the most if not all while they can... then they could stop HSA.

But in the end i think incompatibility is unavoidable, there will be ARM + x86 on HSA vs intel... that is the future, and a good thing IMHO, all direct comparisons with intel will be moot, and then perhaps science can have a little better chance.

 

[Cazalan]

HSA is an ambitious standard but they are only at revision 0.95. These things take YEARS to come to fruition.

It's not like Intel is sitting still. They have similar Serial/Parallel ISA already available.
http://software.intel.com/en-us/intel-parallel-studio-xe

 

[hcl123]

Yes they could be incompatible... kind off... and besides x86 based like Phy X is clearly not the best approach.

About the years, if AMD alone it would take decades not years... but lets see what of most pertinent of the ARM armada (Samsung, Qualcomm, Texas Instruments, LG ... ARM itself,) can do to expedite things.

( On GPU, AMD radeon , Imagination of PowerVR, Vivanti... on DSPs TI, on video/imaging processors LG, Sony... and many more) (edit)

 

[mayankleoboy1]

samsung, for one, is eager to make itself independent of all other hardware and software partners. That is why Samsung is helping in the ARM port of Mozilla's next gen browser , which is a direct competitior to Google/Android.

 

[rmpumper]

Not sure why you call it trolling when the guy said "the i7 980x is faster then the fx8350 still and that was an early 2010 cpu". He is right, 9/10 times (maybe more) 980X wins (OC'ed included). How is this trolling?

 

[jdwii]

OMG Again not everything is so multithreaded when in are you guys going to realize this? Seriously when. OpenCL can't be used for every single thing.

 

[jdwii]

Also for the 10th time if Amd did not think single core performance mattered they would not be making a 4.8ghz CPU and trying everything to get their single core performance up with Steamroller. When you improve single core performance EVERYTHING is faster.

 

[jdwii]

Not sure why you call it trolling when the guy said "the i7 980x is faster then the fx8350 still and that was an early 2010 cpu". He is right, 9/10 times (maybe more) 980X wins (OC'ed included). How is this trolling?

He's using the Straw man fallacy however which discredits him somewhat.

 

[noob2222]

Im not trolling if amd get to sandy bridge levels by 2016 they will have done well from where they are now. Intel are releasing low end products in the market of what they are capable of for the last 3 years and amd still can't beat those low end cpus ie i7 2600k and i7 3770k.

I7 2600k & 3770k ... low end .... Is that why they are $120 more than anything AMD offers? I suppose the $1000+ I7 3970x is resonably priced as well ...

No wonder your trolling.

Here, ill help you out with your illusion. How can AMD reach SB by 2016? I thought AMD was still tring to catch the 486-dx4. AMD is soo slow that it can't even run windows 3.1 . AMD graphics are soo slow that it can't even run DOS. AMD motherboards are soo slow that they don't even boot up. AMD employees are soo slow they ride the short bus to work.

Does that help?

 

[8350rocks]

Its been fine that developers have been limited forcefully or by necessity to archaic legacy because Intel says so, but why not optimized coding for AMD systems? When will there ever be a change for something new and with potential except the same old excuse is thrown around. Coders for EA and other developer companies have said we are stuck in a limbo and HSA provides a way of expanding beyond that, Some have said we are still in the 80's all that has changed is things look prettier yet there is the same old "you cannot do it" which is funny when you see those who can do it, it ultimately comes down to who gives a damn and has the skills.

Coders of EA and others were hampered by the console they are targeting as their primary platform.

APU =/= HSA. On PC's, there is only the APU so far. Where is the promised HSA ? Hd79xx was supposed to have basic unified memory, but it never happened. AMD has a history of promising a lot, and not delivering. For all the hype AMD have built, where is the damned hardware ?
Even in the PS4 chip, the only part of HSA is the unified memory/PCIE lanes. (Which is a big deed in itself, but falls quite short of what AMD has bragged about.). The same Jaguar chip, on desktops, lacks the unified memory. I seriously doubt AMD's commitment to desktops.

HSA is already in Kabini and Temash, and Kaveri brings it to desktop.

HSA features have been on AMD APUs since Llano. It goes by phases, there are 5 phases i think.

HSA *software* is a totally different matter. When there will be a *runtime* to download and or install, then is when it is said, that platform is ready for HSA software.

Its possible that HSA foundation builds a runtime for intel platforms, which i doubt, but for ARM, possibly with Android, that tries to follow all kronos standards is quite to expect.

Who would have the best hardware support most probably will have the best performance, the HSA hardware definition, the minimal that is needed for HSA software, is hUMA and task scheduling queues with that integrated virtual memory awareness(visible from user space application software) (edit)... and very little else... apart from having a Parallel (GPU) or heterogeneous Processor present (could be a DSP), which is the fundamental condition as obvious, but that all already do, AMD, ARM and intel etc (edit)

So GPU = CPU i think only AMD will path, and its not necessary. Probably intel will path nothing of that, having similar but not really compatible... so why bother building a runtime for them ?

But even so, most of those minimal hardware requirements can be *emulated* in software, it will run notably slower but it will run all HSA software in intel platforms.

That is what gives intel the cold sweat nightmares... that is what is behind the wasty decision of having broadwell mostly BGA, and expansion in dead end, grab the most if not all while they can... then they could stop HSA.

But in the end i think incompatibility is unavoidable, there will be ARM + x86 on HSA vs intel... that is the future, and a good thing IMHO, all direct comparisons with intel will be moot, and then perhaps science can have a little better chance.

All true, let me reword...hUMA is in Kabini and Temash, and Kaveri will bring it to desktop.

 

[8350rocks]

HSA is an ambitious standard but they are only at revision 0.95. These things take YEARS to come to fruition.

It's not like Intel is sitting still. They have similar Serial/Parallel ISA already available.
http://software.intel.com/en-us/intel-parallel-studio-xe

Yes, proprietary software you have to have to run it on their hardware(and only their hardware is compatible with it), and they charge for it.

VS.

Open standard with easily obtainable, cost free, software development kits supported by 90% of the industry except Intel, and supporting hardware that's already in place in the next gen consoles.

Man...I am not psychic...but something tells me one of these is less likely to succeed than the other.

 

[juanrga]

Finally, I think you miss what HSA is. HSA is a brilliant attempt to optimize the resources in your computer. CPUs are good for serial workloads, GPUs are good at parallel workloads. HSA proposes a 'abstract' programming model where the best unit (CPU or GPU) is used at each time for a given workload in function of its nature: serial or parallel. If you can boost existent application by 5x by using HSA what is wrong with that?

This is already being offered by OpenCL.
Where HSA is different is that it aims to remove/diminish the difference between a CPU and GPU at the hardware level.

Also, i am much less impressed by developer claims that HSA increased some tasks perf to 5x in the PS4. It is easy to develop a specialized software if you know all the hardware internals, and are coding at the machine level. The real test of HSA is on a desktop, with general purpose software running on multitudes of HSA devices, each with a different configuration.

The only HSA we have yet seen on desktops is Richland, which is mostly a CPU+GPU on a single die. What is the timeframe for AMD to release a true HSA device for personal computers ? What is the chance that PC's will get a unified RAM+VRAM ? Will MS windows work properly in the HSA world ?

From the PC perspective, HSA is still only a vision shown by AMD.

OpenCL did not provide "an 'abstract' programming model where the best unit (CPU or GPU) is used at each time for a given workload in function of its nature: serial or parallel." OpenCL introduced an interface for using your GPU for non-graphic tasks. HSA fusions your CPU and GPU into a heterogeneous computing unit.

Where did you got the idea that 5x boost is for the PS4? Moreover, HSA has nothing to do with "coding at the machine level". They did not mention the console, but are discussing PC APUs in the link.

Richland only includes some aspects of HSA. The first fully enabled HSA APU for PCs will be Kaveri, which is coming this year, with hUMA (unified memory).

In the first place, AMD Piledriver CPU architecture is good enough to win the new Haswell CPU. There are benchmarks where the 'old' FX-8350 is faster than the new i7-4770k. Moreover, the new Centurion FX 9000 series will be about a 15% faster than 8000 series.

In the second place, AMD continues its plans with Steamroller and Excavator architectures. Steamroller CPUs will be about a 20% faster than Piledriver 8000, clock per clock, probably a 30% with a clock boost. If a 8000 or 9000 series can outperform a top-of-line 4770k, imagine what Steamroller series can do.

Finally, I think you miss what HSA is. HSA is a brilliant attempt to optimize the resources in your computer. CPUs are good for serial workloads, GPUs are good at parallel workloads. HSA proposes a 'abstract' programming model where the best unit (CPU or GPU) is used at each time for a given workload in function of its nature: serial or parallel. If you can boost existent application by 5x by using HSA what is wrong with that?

The i7 980x is faster then the fx8350 still and that was an early 2010 cpu. Also the 4770k wins on around over 80% of the benchmarks against an fx8350. 15% nothing really, oc the 4770k to 4.2ghz and you get that 15% easily. Intel barely have any competition hence why they hold most of the market as they are that far ahead. Intel architecture is so far ahead of amd now I reckon 2016 will be the year they get to sandy bridge levels and that will take a gigantic leap from where amd are now. I think steamroller will get close to sandy bridge level though.

As said the FX-8350 can outperform the 3770k and the 4770k, and the 980x as well. It can win the 5% or the 80% of the benchmarks, it depends of what benchmarks you select. If you select ancient useless single-threaded benchmarks that nobody uses or biased benchmarks with the Cripple-AMD function, then the things look better for Intel. If you select modern multi-threaded benchmarks and fair benchmarks, then the things look better for AMD.

You have been here for days trolling about how great is haswell/hasfail/hasbeen/failwell when the increase in performance of the 4770k over the 3770k is of about a 8% and now you pretend that an increase of 15% on AMD FX series is nothing? LOL

IB OC better than haswell, OC a 3770k and you will beat the 4770k chip in almost everything. Moreover, Centurion FX chip also OC.

 

[earl45]

Im not trolling if amd get to sandy bridge levels by 2016 they will have done well from where they are now. Intel are releasing low end products in the market of what they are capable of for the last 3 years and amd still can't beat those low end cpus ie i7 2600k and i7 3770k.

I7 2600k & 3770k ... low end .... Is that why they are $120 more than anything AMD offers? I suppose the $1000+ I7 3970x is resonably priced as well ...

No wonder your trolling.

Here, ill help you out with your illusion. How can AMD reach SB by 2016? I thought AMD was still tring to catch the 486-dx4. AMD is soo slow that it can't even run windows 3.1 . AMD graphics are soo slow that it can't even run DOS. AMD motherboards are soo slow that they don't even boot up. AMD employees are soo slow they ride the short bus to work.

Does that help?

AMD don't have the performance to charge a premium, when they did the FX cost a 1000.00
dollars.

 

[mayankleoboy1]

Richland only includes some aspects of HSA. The first fully enabled HSA APU for PCs will be Kaveri, which is coming this year, with hUMA (unified memory).

What does Richland have to do with HSA , apart from being a CPU+GPU on a die ?

The first semi-HSA device on PC would be Kaveri. And we dont know if it will have hUMA or not. And hUMA is not one thing, but a collection of many types of memory unification, such as register sharing, memory sharing, context switching etc.

 

[mayankleoboy1]

Im not trolling if amd get to sandy bridge levels by 2016 they will have done well from where they are now. Intel are releasing low end products in the market of what they are capable of for the last 3 years and amd still can't beat those low end cpus ie i7 2600k and i7 3770k.

I7 2600k & 3770k ... low end .... Is that why they are $120 more than anything AMD offers? I suppose the $1000+ I7 3970x is resonably priced as well ...

No wonder your trolling.

Here, ill help you out with your illusion. How can AMD reach SB by 2016? I thought AMD was still tring to catch the 486-dx4. AMD is soo slow that it can't even run windows 3.1 . AMD graphics are soo slow that it can't even run DOS. AMD motherboards are soo slow that they don't even boot up. AMD employees are soo slow they ride the short bus to work.

Does that help?

AMD don't have the performance to charge a premium, when they did the FX cost a 1000.00
dollars.

Shhh. Dont say that out loud. AMD fanbois like to believe that this phase of AMD never existed.

 

[mayankleoboy1]

edit.

 

[hcl123]

Also for the 10th time if Amd did not think single core performance mattered they would not be making a 4.8ghz CPU and trying everything to get their single core performance up with Steamroller. When you improve single core performance EVERYTHING is faster.

Yes in a sense, but you are dealing with a "SCALING DEAD" paradigma, its not because intel have not good engineers that hasfail, is hasfail... understand ?... some uarchs more than others, but when you stat relative low, usually you get better improvements, nevertheless in this view IF SR is 30% better, shows that AMD uarch is better than intel actual uarch.

Yet if nothing extraordinary passes, and it forcibly has to do with SOFTWARE... fails as hasfail like, will be the norm for all, simply because its x86 and it doesn't scale anymore... unless a breakthrough paradigma happens in software conjugated with CPU, heavy multithreading or something out of the box like HSA, where its "rumtime" can provide a "dataflow approach" (as example) with heterogenous extensions and even on-the-fly (its a runtime) soft developer kind of transparent speculative multithreading ( that is, spMT transform on-the-fly, "sequential" software in multithreading one -> the reverse Hyperthreading rumour)

It fits then better the GPU paradigma... GPU ALUS run at 1ghz they are 32bit wide relativly pretty "slim" compared with the SSE/AVX ones, yet for data parallel workloads thousands of these slim ALUs in a medium size chip can deliver a severe beating to even 32 or 64 fat CPU cores. (appreciate the difference ?)

So who keeps pushing the single-thread mantra its not AMD its the part of market centered most exclusively on the client/desktop sector(the server side wants MT, but they have plenty of money and machines with plenty of cores). AMD over estimated the chances of having generalized good multithreading software on the client/desktop by the time BD sorted... the problem is not the CPU is the SOFTWARE.

I don't know how much long i'm going to repeat this... the problem is SOFTWARE lol ...

 

[earl45]

Im not trolling if amd get to sandy bridge levels by 2016 they will have done well from where they are now. Intel are releasing low end products in the market of what they are capable of for the last 3 years and amd still can't beat those low end cpus ie i7 2600k and i7 3770k.

I7 2600k & 3770k ... low end .... Is that why they are $120 more than anything AMD offers? I suppose the $1000+ I7 3970x is resonably priced as well ...

No wonder your trolling.

Here, ill help you out with your illusion. How can AMD reach SB by 2016? I thought AMD was still tring to catch the 486-dx4. AMD is soo slow that it can't even run windows 3.1 . AMD graphics are soo slow that it can't even run DOS. AMD motherboards are soo slow that they don't even boot up. AMD employees are soo slow they ride the short bus to work.

Does that help?

AMD don't have the performance to charge a premium, when they did the FX cost a 1000.00
dollars.

Shhh. Dont say that out loud. AMD fanbois like to believe that this phase of AMD never existed.

Yes I know and it's sickening.

 

[blackkstar]

The i7 980x is faster then the fx8350 still and that was an early 2010 cpu. Also the 4770k wins on around over 80% of the benchmarks against an fx8350. 15% nothing really, oc the 4770k to 4.2ghz and you get that 15% easily.

Please do not be so naive. The maximum turbo speed of 4770k is 3.9ghz and it runs much better than base frequency nearly always (actually read some reviews instead of looking at graphs).

If 4770k is running at turbo speed, it needs a little under 4.5ghz to get a 15% increase, something that, according to Asus, you are going to struggle to reach.

Also, do not forget that 4770k has a nearly 12% clock speed increase when going maximum turbo while FX 8350 only has 5%. Thus, it's pretty unfair to compare base frequency of FX 8350 and base frequency of 4770k whilst the chips have turbos set at different levels.

As an example:

difference between FX 8350 and 4770k base clock speeds: 14.2%
difference between FX 8350 and 4770k max turbo speeds: 7.7%.

I grow rather weary of folks saying 8350 has ~15% higher clock speed and then they show me a review with turbo enabled with a single core benchmark where FX 8350 actually only has a ~7% clock speed advantage. It's just, to be quite honest, rather dumb and you're basically cutting the clock speed advantage for AMD in half while saying that it hasn't been cut in half.

I do not know if you're a troll, but I look forward to the folks reading this who will realize that "FX 8350 has 15% faster clocks and is 25% slower" isn't true in a large number of benchmarks, and it's much more like 7% faster clocks and 25% slower, which makes the IPC difference much smaller than people think.

When "enthusiasts" do nothing but masturbate to bar graphs, it's typical for them to take a year to pick up on it. Reading is too hard, pictures please!~

I forget if you were the guy I accused of being with Intel PIE or something but you really do sound like the kind of person who is parroting marketing points. If you are a genuine enthusiasts I would suggest you take a step back and take some time to learn some new things.

 

[8350rocks]

Im not trolling if amd get to sandy bridge levels by 2016 they will have done well from where they are now. Intel are releasing low end products in the market of what they are capable of for the last 3 years and amd still can't beat those low end cpus ie i7 2600k and i7 3770k.

I7 2600k & 3770k ... low end .... Is that why they are $120 more than anything AMD offers? I suppose the $1000+ I7 3970x is resonably priced as well ...

No wonder your trolling.

Here, ill help you out with your illusion. How can AMD reach SB by 2016? I thought AMD was still tring to catch the 486-dx4. AMD is soo slow that it can't even run windows 3.1 . AMD graphics are soo slow that it can't even run DOS. AMD motherboards are soo slow that they don't even boot up. AMD employees are soo slow they ride the short bus to work.

Does that help?

AMD don't have the performance to charge a premium, when they did the FX cost a 1000.00
dollars.

Shhh. Dont say that out loud. AMD fanbois like to believe that this phase of AMD never existed.

They just played Intel's game...you can't take the crown from the master of gouging...even if you have a better product.

 

[hcl123]

AMD don't have the performance to charge a premium, when they did the FX cost a 1000.00
dollars.

Not enough, not even when Athlon64 was average 50% better than P4... AMD would need an enormous propaganda and marketing machine as intel. There is no justification for $1000 CPUs on the client/desktop sector when 10X less expensive chips can beat them at some workloads (see chart of Expresso transcode http://www.guru3d.com/articles_pages/amd_a10_6800k_review_apu,14.html ) and when half the price CPUS are inside(less) the 30% difference.

So $1000 desktop CPUs MEANS SQUAT about Intel been better, more advanced, more anything... matter of fact G34 server CPUS can beat easily even the 3970x at MT loads, and about the same price(but its server).

$1000 desktop CPUs means very very VERY aggressive marketing... and all that aggressiveness, like time-sharing sales, needs *VICTIMS*... that by all means must continue thinking they have made great deals (but as usually the human subconsciousness mind is smarter, normally to compensate its ignorance and silence the subconsciousness, the only compensation the "vigil" conscience mind has is resort to violence (violence = ignorance).. the violence of insult...)