Haswell Core Count

[Lonbjerg]

No we don't. The GTX 680 is pathetic at general-purpose computing.

Games are severely limited by the lack of throughput computing power offered by today's CPUs, and clearly even the latest GPU is rubbish for GPGPU.

Hence AVX2 will revolutionize gaming. It brings us SPMD throughput computing, without the limitations the GPU runs into. The future is homogeneous computing, not heterogeneous computing.

GK104 has been neutered in GPGPU capabilities...it's common knowlegede....wait for Big K before pointing the finger at GPGPU...using the wrong...GPU :whiste:

 

[greenhawk]

990x still selling for 1400 USD.
(Ironicly, 3960x is like 1325 USD ish - yea we're messed up).

not that messed up. It is not uncommon for the old hardware to keep it's price regardless of performance as even at equal prices, it is a cheaper upgrade to just replace the processor for people with compatable hardware already. Buying new on the other hand is a no brainer.

Intel has been doing this process for as long as I can remember.

Once you buy into a socket you are a captive audience without a full system upgrade.

 

[Lonbjerg]

possibly, but they are thinking in the now.

I was one of the people saying dual was better than single back when dual required two cpu's, the rest came around to my thinking.

That being said, I have not seen a huge number of things that benifited my daily tasks to make quad massivly better over dual until recently (still running a q6600, so it took a long time).

It seems the best thing since the q6600 has been turbo mode which disables unneeded cores. Not sure what feature will be next but I do not have the gut feeling it will be going to 6 cores for the masses (like the local wall mart / ect will find it any easier to get mom and pop to but a quad over a dual).

I don't live in the future, I live in the present.
4 cores/8 threads is more than fine for mainstream.

When programs will require more...THEN people can upgrade...and get better hardware for the same amount of money...and better peformance.

You buy for the future...then I'll will buy for today.

 

[HexiumVII]

Hey at least you get true quad cores vs bulldozer ;p

 

[Lonbjerg]

While I would generally agree with the bubble statement, I find it sillier still that if there is such a large price difference between local and US prices, why not mail order? been doing that for years. Far easier than going off about local prices.

That would end up costing more.
The parcel would be stopped in customs, VAT and import taxes would be added...plus an EXTRA fee...for the trouble the customs have opening hte parcel...I kid you not.

In DK allmost all parcels are looked over by customs....very hard to get around VAT and import taxes.

 

[khon]

Kinda disappointing that Haswell mainstream will only be quadcores. Quadcores have been around for more than 5 years already, and by the time Haswell comes out that will be at least 6 years.

 

[T_Yamamoto]

Kinda disappointing that Haswell mainstream will only be quadcores. Quadcores have been around for more than 5 years already, and by the time Haswell comes out that will be at least 6 years.

the thing is, not everything utilizes 4 cores yet.

 

[Mr. Pedantic]

There is a need for performance, and intel deliver options for it.

Not really. There is always going to be a demand for more performance. Simulating weather; a human brain; nuclear reactions; molecular dynamics; rendering movies. Intel is not even close to what is 'needed'. Nobody is.

Just because there is no competition at that level means intel can charge what ever they want for it.

I know.

might as well be going off at AMD for being bad competition, but still be one of those people that do not buy there products to help fund then to be a good competition against intel.

Oh, I do. AMD really dropped the ball with Bulldozer. And I don't buy their products because the purpose of my wallet is not to subsidize bad decision making. I already have taxes for that.

Nope, but the desire can head in the direction of being labeled the same as the people that have not grown up who complain, or the ones that go down the path of "it is too expencive to buy, I'll just steal it instead".

I don't have to buy processors, because I don't work in an industry where CPU power affects how much I make. But for someone whose livelihood does depend on this, paying $4000 for a 10-core Westmere Xeon would be very, very disheartening.

Whether Intel's mainstream lineup is value for money may be a contentious point. But saying Intel's high-end line is value for money is just silly, because it's not supposed to be.

the thing is, not everything utilizes 4 cores yet.

Irrelevant to the point. If I write a basic calculator, it doesn't have to be multithreaded because nothing it will ever do will stress more than a fraction of 1 second of core time. Simple programs don't have to be multithreaded. The programs that really need multithreading, by and large have already implemented it.

Only major exception I can think of is Microsoft Office.

 

[ShintaiDK]

Not really. There is always going to be a demand for more performance. Simulating weather; a human brain; nuclear reactions; molecular dynamics; rendering movies. Intel is not even close to what is 'needed'. Nobody is.

Thats where you buy Intels MIC product.

 

[Mr. Pedantic]

Thats where you buy Intels MIC product.

Lol. 1TFLOPS. To model a human brain, you'd probably need around 25 million of them. As I said. Nowhere near close.

 

[AtenRa]

Just a food for thought,

SB is 216mm2
SB-E 8-core die is 430mm2

SB-E at 22nm could be close to half the size, that means close to SB's 200-230mm2.
I believe Intel could very well have a 6-core IV-E at $299.00 and make a huge profit from it. 8-core 22nm IV could start at $499 and above.

The only problem i see is that IV-E will first be designed as a Server CPU and i expect them to raise the amount of L3 cash. That will raise the size of the die and any hopes for cheaper 6-cores.

 

[MisterMac]

Many Replies..


To the people arguing "We don't need more, in general mainstream - nothing supports it".
Of course it "beep" doesn't.

Intel is pulling the handbrake like crazy.
BD was designed with a modular threaded world in mind - sadly we aren't there.

Anyone providing a solid arguemnt, if we had 8 or 12 nehalem core chips now (instead of say, SB) - that we wouldn't see anything from windows to browsers being completely rewritten with threaded scalable core in mind - ill buy you a beer.

And i'm personally disappointed by haswell regarding this - because while im a novelty user i could use it.
And EVERYONE could use it - had the visions AMD saw when they created BD came true.

Hence my problem.
Knowing how my own programming abilities can much easier get more perf out of "half arsed" threading implementations - than spending god knows how many hours looking at my compiled assembly and knowing my IDE to optimize my code.

I'll wager 90% of consumer programs would benefit more from adding simple threading than having AVX2 plus being the easiest solution for the software suppliers both now and in future aspect.

 

[khon]

the thing is, not everything utilizes 4 cores yet.

Almost nothing used 4 cores when the first quad cores came out, but once the hardware is available the software will follow.

 

[pelov]

Almost nothing used 4 cores when the first quad cores came out, but once the hardware is available the software will follow.

It becomes progressively more difficult as you go up in thread count. There's definitely wiggle room but how much depends on the actual relevant code. Consider, too, the fact that we can't passively thread software yet (efficiently) so there has to be a ground up commitment by developers to see that happen by putting it into practice with additional work (legacy) or from the ground up when creating newer programs/code.

There's also die size, IPC and clock speed to consider as all of those are inversely proportional to an ever increasing core count as far as microarchitecture on x86 goes. Adding cores and threads is something that has to be weighed with the adverse effects it may have on code that isn't properly threaded (and optimized, really) which is still the majority of what we run on the desktop (again, we can't passively thread new and legacy code yet).

Rather than addressing a problem that doesn't exist, Intel is attacking their crappy GPU performance because that's something nearly everybody can benefit from, regardless of thread count. Some of us may not like it, sure, but it's a smart move nonetheless.

 

[Infraction Jack]

I am hoping for a 3 core I3.

 

[IntelUser2000]

Not really. There is always going to be a demand for more performance. Simulating weather; a human brain; nuclear reactions; molecular dynamics; rendering movies. Intel is not even close to what is 'needed'. Nobody is.

Yea but the people who does such big jobs can easily afford those processors. CPUs are probably the least of their costs anyway. But individuals don't run those sorts of stuff.

-Tasks being accelerated by dedicated and low power units like QuickSync and/or being offloaded to GPUs.
-Tablets becoming the potential biggest threat to PC.
-Multi-threading being extremely limited in the consumer space
http://www.anandtech.com/bench/Product/443?vs=287

You see applications like Adobe CS4/WME/3DSMax/Blender already benefitting minimally from 2 additional cores. The gains are about the same as what Hyperthreading offered. Hell the 3770K is beating the 3960X! I could point out few more that will benefit absolutely nothing going to 8 cores but still benefit moving from 4 to 6 cores.

-Low power increasing in importance

So why should they make a dedicated chip for the few that wants those extra few cores in the consumer space? Why not just keep making derivatives(3960X like chips) of the Xeon, which the customers really need the extra performance?

 

[dma0991]

While it is possible that Intel is able to give a 6 core mainstream CPU but working within its own TDP, something has to give. It isn't likely we'll see a 6 core IB that still fits within the 95W TDP and have the same stock clockspeed as the current 4 core offerings. Intel could go with a hypothetical 6 core IB that does 2.5GHz max and ST performance will suffer compared to its predecessor.

Mainstream SB/IB is meant to scale down from desktops to laptops and the highest priority is given to improve efficiency and reduce power consumption. I wouldn't be surprised if a mobile IB is in fact a desktop part that has few of its cores disabled to meet the necessary specifications. For Intel it is a win as they would not have to spend the extra on creating a design that is meant for mobile alone.

Unlike SB-E/IB-E which are meant to scale up, they are not constrained by the limits that are imposed on mainstream products.

 

[IntelUser2000]

Mainstream SB/IB is meant to scale down from desktops to laptops and the highest priority is given to improve efficiency and reduce power consumption. I wouldn't be surprised if a mobile IB is in fact a desktop part that has few of its cores disabled to meet the necessary specifications.

Uh, no. Intel creates the mobile chips first then scales UP. That has been true for quite some time. In the future they say they'll go even lower(from current 35W) and make the 17W base and go up from there.

Also Sandy Bridge has 3 seperate dies and Ivy Bridge, 4. Desktop chips are merely scaled up versions.

Sandy Bridge: 4 core + GT2, 2 core + GT2, 2 core + GT1
Ivy Bridge: 4 core + GT2, 4 core + GT1, 2 core + GT2, 2 core + GT1

That's not counting server chips with own dies and extreme enthusiast chips like Sandy Bridge E.

 

[beginner99]

3930k is 500 USD.
+ a cheap mobo - 250/300 USD.

800 USD.

Now let's try that overseas here in EU:

3930k = 900 USD
Granted there are cheap mobo's for 350 USD in my territory, but most are 400 USD.

That equals 1250/1300 USD averagely.

no idea where you live but I live in europe too and in a country not really known for its low prices. However a 3930k is about $615. Motherboards range from $200 up to over $500.

Or said otherwise prices aren't that much higher than US.

 

[dma0991]

Uh, no. Intel creates the mobile chips first then scales UP. That has been true for quite some time. In the future they say they'll go even lower(from current 35W) and make the 17W base and go up from there.

So if it is scaled up, they should be capable of fitting in a 6 core part as well right? Or is it not possible due to the fact that mainstream products have an IGP and scaling up from mainstream to enthusiast/server level would cost more than to create a design of its own from the ground up.

 

[pelov]

So if it is scaled up, they should be capable of fitting in a 6 core part as well right? Or is it not possible due to the fact that mainstream products have an IGP and scaling up from mainstream to enthusiast/server level would cost more than to create a design of its own from the ground up.

That's limited by the size of the die and the on-die GPU, yea. Only the server chips have dies large enough for 6 cores and the SB-E server class chips are huge, 2.27B transistors and 435mm2.

They could make anything they wanted but it would mean much higher prices and less viable chips-per-wafer which doesn't make any sense for laptops. Who the hell needs 6 cores on a laptop? Even on the desktop the cost isn't warranted but for a few people, thus creating a 6-core IB + on-die GPU means an even bigger die and more transistors and even more expensive. It doesn't make any sense.

http://www.anandtech.com/show/5771/the-intel-ivy-bridge-core-i7-3770k-review/3

 

[IntelUser2000]

So if it is scaled up, they should be capable of fitting in a 6 core part as well right? Or is it not possible due to the fact that mainstream products have an IGP and scaling up from mainstream to enthusiast/server level would cost more than to create a design of its own from the ground up.

Scaling up in this case means increasing voltage and frequency, plus binning. The desktop and mobile dies are same. But I assume when they decide to design a CPU, its based off mobile needs first.

Whether they decided to create a dedicated enthusiast CPU, or a derivative from the Xeon parts, it'll require different layout. Making 3 and 4 different dies in SNB/IVB merely requires them cutting out parts. That's not the case for devices with more cores and/or caches. The easiest way is if Intel has been keeping secret from us and there's yet another die with 6 cores plus integrated graphics. So they can cut their way down. But that does not seem to be the case.

I don't think it would cost a lot more to create a dedicated die. It's just that its not worth it from a volume point of view. The regular desktop/mobile dies sell in very high volume, and server/workstation chips ship in decent volume and sell at higher costs. A dedicated enthusiast part fills neither. Come to think of it, I think they always made enthusiast parts based on server dies.

 

[MisterMac]

no idea where you live but I live in europe too and in a country not really known for its low prices. However a 3930k is about $615. Motherboards range from $200 up to over $500.

Or said otherwise prices aren't that much higher than US.

Welcome to Scandinavia.
Where it's cheaper to actually send a package from GERMANY to Denmark than from denmark to denmark.

hi5.

 

[ninaholic37]

Dual-core is more than enough for most of the things I do (I think quad core and dual-core/dual-threaded is still overkill and just adds unnecessary complications for most things, seems more like a marketing gimmick than anything to me). I'd rather they focus on increasing performance in other areas (more SSDs, faster GPUs, etc.), as some others have said.

 

[Edrick]

SB-E at 22nm could be close to half the size, that means close to SB's 200-230mm2. I believe Intel could very well have a 6-core IV-E at $299.00 and make a huge profit from it. 8-core 22nm IV could start at $499 and above.

Yea, and Intel could also keep selling their 6 core at $500 and make an even bigger profit. Without AMD stepping up, Intel can do what they want.