Broadwell-E 10 Core Costs $1723

[2blzd]

Well, *one* hex core is priced reasonably. The more expensive hex core is not. All it does is fill the spot the 8 core should fill.

agreed 100%

Like many, when the 10c leak first appeared and surprised people (at the time official intel slides still had 6c/6c/8c), I assumed the 8 core would replace the high-end hex's sku spot at around $550-$700 and the 10core would be the new $999 chip...

Still bummed...i cannot wait for skylake-e though, I need something by the end of this summer

 

[frozentundra123456]

Some posters (well one certain one anyway) will rationalize/defend whatever intel does to the last breath. I consider myself an Intel supporter basically, but I totally dont understand that extent of defensiveness toward any company.

I am not sure I buy that argument, but even if one does, who dictates Xeon prices??

In any case Intel could easily have priced the 10 core 1200.00 or so, and lowered the price of the 8 core by a hundred or two.

 

[moonbogg]

Intel probably saw how Nvidia priced their GPUs and figured they might as well charge a truck load as well because people will just buy it anyway. $1,700 won't stop most of the PC top end buyers from buying this chip.
They realize people have money to spend, and with $700 mid range GPU's, $1,000+ high end GPU's, and an already $1,000+ high end CPU, why the hell would a few extra hundred stop someone?

 

[Torn Mind]

I guess I didn't find the comment that Intel's R&D is mostly poached from US Universities, that are subsidized with gov't funding, "derogatory", nor "inflammatory". I took it as a factual statement.

Intel's CPUs aren't produced solely with Intel's R&D. Plenty of University papers are used in the design of a modern CPU.

In fact, wasn't there some University suing Intel over poached research, that they claimed that Intel had to license from them, but didn't? Was it Berkeley?

Edit: Yep, several, in fact!
www.xbitlabs.com/news/cpu/display/20080211211942_University_Sues_Intel_for_Patent_Infringement.html
http://www.eetimes.com/document.asp?doc_id=1257902
http://www.cnet.com/news/wisconsin-madison-sues-intel-for-patent-infringement/#!

Intel (allegedly) STEALS RESEARCH from Universities, funded by the US Gov't.

Edit: In hindsight, any taxpayer-funded research should be available for free to any US Citizen. Including corporations. But I'm guessing that those Universities, while receiving gov't funding, are not SOLELY funded by the gov't, so their private donators, etc., have a stake in those inventions too, so they want a cut.

Intel run as it is now is better than becoming a gargantuan Post Office 2.0. When you way something is nationalized, then you must swallow everything that comes with it. Pension plans and other generous benefits. And acceptance that the "worst" can happen, which is that Intel is in the red but remains immortal as long as taxpayers and buying debt fund it.

 

[beginner99]

Intel just declared broadwell-e as failure. Who in their right mind will buy this? It's not even out yet and end of 2016 you get kaby lake, which will be 2 generations ahead of BW-E. And since z170 already has much more Pcie lanes compared to z97, makes the cheap HEDT hex-core pretty much useless.

 

[RussianSensation]

Intel probably saw how Nvidia priced their GPUs and figured they might as well charge a truck load as well because people will just buy it anyway. $1,700 won't stop most of the PC top end buyers from buying this chip.
They realize people have money to spend, and with $700 mid range GPU's, $1,000+ high end GPU's, and an already $1,000+ high end CPU, why the hell would a few extra hundred stop someone?

That's a valid point. 1080/1080 SLI and $1700 CPU go well with the type of customer who pays $700 for next gen mid-range chips. If we figure i7 6700K is Intel's upper echelon chip, Intel must make at least 60% profit margin on it to have > 60% average. That means for Intel, every new customer who is building a 1080/1080 SLI rig is either getting an i7 6700/K or i7 5820-5960X/6800-6950X. They don't care what you buy since they are getting at least 60% gross margins from this customer. Could have charged $699 for this chip; it'll still lose in gaming to a $309.99 USD MicroCenter 6700K. That's the saddest part of it. These chips are truly meant for workstation/productivity use. If someone is YouTube streamer/content creator, $1723 is peanuts if this makes you $4000-10000 USD a month. What's surprising is that Intel raised the price too much so fast. I bet the market would have been more receptive if they did $1299 this gen, then $1599 with 12-core SKL-E. I think Intel had the right idea but the execution was awful.

BTW, SKL-E isn't going to be any better. Let's say it launches in June 2017, in about 1 year or slightly more, Intel will have its next generation Icelake. That means for top end gaming performance, SKL-E will again lose. Worse, Skylake architecture will be almost 2 years old by the time it launches. Intel, just like NV, put us in a position where we cannot win. I suppose sooner or later something has to give. Either IPC gains for new architectures will be even worse than Haswell --> Skylake, or Intel releases a 6-core mainstream chip. Either of these scenarios tilt the buying decision towards one of the platforms.

 

[2blzd]

Intel just declared broadwell-e as failure. Who in their right mind will buy this? It's not even out yet and end of 2016 you get kaby lake, which will be 2 generations ahead of BW-E. And since z170 already has much more Pcie lanes compared to z97, makes the cheap HEDT hex-core pretty much useless.

I am going to buy it, thanks. And it most definately will not be useless. Thanks though.

spoiler alert: HEDT chips will always be behind the DT cores, your revelation isn't news

 

[ZaphodBbrox]

I expected Intel to charge more for the 10C chip, but not this much! Oh well, I just bought a 3.0GHz 24C/48T Xeon v4 workstation with 256GB RAM for my work.

 

[Aristotelian]

Well, I expected roughly EUR 1499 for the 6950X and I just found a place in the EU where I can pre-order one for EUR 1799. If it's true that the 6950X is going to be USD 1723, then the EUR 1799 price is actually a steal...and therefore I'm thinking that the USD 1723 price is actually wrong. Can that still be possible? I've never seen EU pricing so close to the USA pricing, and the EUR price includes VAT.

 

[strategyfreak]

I don't buy the argument that the 6950x should be priced higher because it offers higher productivity.

First, if your task scales well enough to take full advantage of the 10 core 6950x, then chances are it will benefit from going to 12+ cores of the Xeon setups. So even though its overclockable, the 6950x performance will lose out to a higher core count for parallelized tasks. Furthermore, instability from overclocking too high will produce unreliable results, which may be unacceptable in certain professional scenarios.

If your task just needs single thread performance, then clearly the 6950x isn't the way to go. Either a 6700k or Broadwell is the way to go, and if time is really important then you can take the risk of overclocking.

Really the only case that you might choose this chip for productivity reasons is where you have a task that scales only to 8-10 cores and you need higher performance from overclocking.

If you have a multithreaded task that needs lots of cores and you're on a tight budget: There are plenty of used older Xeons (Ivy Bridge-EP or Haswell-EP) that can be had for cheap that offers many more cores for a lot less than a brand new 6950x. Instead of spending $1800 on a 6950x, you could get a Xeon that's clocked lower but has many more cores. Not to mention the opportunity to use 2P/4P boards, which the 6950x cannot. Many professionals might balk at used equipment, but then again if you're worried about that then you're probably not going to be getting a 6950x for overclocking - both introduce unreliability. They'd just go for a new Xeon system instead.

If you're a professional with the money to spend, then you'd go for a Xeon workstation or server instead of getting a 6950x. The 2687 v4 offers 12 cores @ 3.0 ghz, and the option for a multisocket board, and costs only about $200-300 more.

The sole purpose of the 6950x and its exorbitant price is to milk enthusiasts dry. The only relevant question is whether the ego boost of having the "most powerful" processor is worth the cost. Intel is guessing yes.

 

[dark zero]

Agreed. I feel that even if AMD Zen is released at 599 dollars on their Octo Core, people will stay at the Intel one since is a Deca Core.
/Human logic.

 

[RussianSensation]

I am going to buy it, thanks. And it most definately will not be useless. Thanks though.

spoiler alert: HEDT chips will always be behind the DT cores, your revelation isn't news

Alright so tell us then how useful will it be for productivity against 2 systems with used 5960X PCs, and how it's acceptable for a $1723 chip to lose to a $310 i7 6700K in games. The worst workstation chip on X58 platform was the $284 i7 920. It was better than the best Lynnfield chip that launched one year later once both were overclocked. After X58, the entire workstation platform is always an architecture behind the mainstream one. By the time SKl-E comes out in 2017, the SKL architecture will be 2 years old and just 1 year away from Icelake. The only thing that would stop the SKL-E workstation platform from being driven into total irrelevance for 98% of PC users would be the lack of a 6-core Icelake part.

 

[Arachnotronic]

Intel just declared broadwell-e as failure. Who in their right mind will buy this? It's not even out yet and end of 2016 you get kaby lake, which will be 2 generations ahead of BW-E. And since z170 already has much more Pcie lanes compared to z97, makes the cheap HEDT hex-core pretty much useless.

I'm going to buy one as soon as they become available tonight. Already have my motherboard + new quad channel DDR4 kit in. I'm pumped. I believe that I am in my right mind

 

[Arachnotronic]

Alright so tell us then how useful will it be for productivity against 2 systems with used 5960X PCs, and how it's acceptable for a $1723 chip to lose to a $310 i7 6700K in games. The worst workstation chip on X58 platform was the $284 i7 920. It was better than the best Lynnfield chip that launched one year later once both were overclocked. After X58, the entire workstation platform is always an architecture behind the mainstream one. By the time SKl-E comes out in 2017, the SKL architecture will be 2 years old and just 1 year away from Icelake. The only thing that would stop the SKL-E workstation platform from being driven into total irrelevance for 98% of PC users would be the lack of a 6-core Icelake part.

IMHO, does not buy a 6950X because it is a "good value." One buys this chip when one wants to own the absolute fastest multi-threaded monster consumer chip on the planet. 6950X will appeal to those customers.

And my gosh, that box is slick. MrTeal may have been sarcastic when he was talking about a $500 box, but that is definitely a box I would keep around after the build is finished.

 

[Madpacket]

Wow I was looking forward in replacing to my 5930K with a new 8 core but not at these prices. I bet Intel got wind of Zen performance and priced their chips accordingly.

 

[Arachnotronic]

Wow I was looking forward in replacing to my 5930K with a new 8 core but not at these prices. I bet Intel got wind of Zen performance and priced their chips accordingly.

I hope that the reviews show a ~5% IPC boost and no regression in OC capability. Fingers crossed.

 

[Heatshiver]

Thinking I am awaiting Zen to see where that lands and if 8C+ chips get a little cheaper...

This, definitely this!

 

[Jovec]

If you need 10 cores it's fairly cheap.

Sorry, I want cheap 6/8/10 cores like all of you do for hobby/home use, but as a productivity/money-making tool $2k isn't that bad.

 

[Madpacket]

If you need 10 cores it's fairly cheap.

Sorry, I want cheap 6/8/10 cores like all of you do for hobby/home use, but as a productivity/money-making tool $2k isn't that bad.

This is backwards thinking and why we're all still stuck on 2-4 cores for the mainstream. When did the 2500K come out? There's no excuse other than lack of competition to not give us more cores at affordable prices in 2016.

Intel could easily release affordable 6 - 8 core CPU's and we would all buy it up in droves.

This would lay the groundwork necessary for game developers to justify the R&D costs to really take advantage of multiple fast cores allowing for more capable and diverse games. Most PC games still scale poorly with multiple cores forcing IPC as the primary factor to obtain high frame rates so we can enjoy cool things like 144Hz monitors.

Ironically the gaming consoles have been the primary driver of multi-core development even with dog slow jaguar cores but we PC gamers could have so much more by now if not for insane prices for a few more cores.

And people wonder why the PC desktop industry is in decline.

 

[Arachnotronic]

This is backwards thinking and why we're all still stuck on 2-4 cores for the mainstream. When did the 2500K come out? There's no excuse other than lack of competition to not give us more cores at affordable prices in 2016.

Actually, I don't think this is true. The parts that Intel sells with more cores tend to have lower frequency per core as well as higher power consumption. Also, the mainstream parts are derived from mobile parts, and there would be very little benefit to a 6-8 core in a laptop, for example.

Intel could easily release affordable 6 - 8 core CPU's and we would all buy it up in droves.

The 5820K, and soon the 6800K, will not be much more expensive than the top quad core on the mainstream platform. Could you please give your definition of affordable?

This would lay the groundwork necessary for game developers to justify the R&D costs to really take advantage of multiple fast cores allowing for more capable and diverse games. Most PC games still scale poorly with multiple cores forcing IPC as the primary factor to obtain high frame rates so we can enjoy cool things like 144Hz monitors.

IMHO, the groundwork has been laid with those 8-core AMD APUs that are inside the two major consoles. However, taking advantage of a lot of cores is simply tough.

Ironically the gaming consoles have been the primary driver of multi-core development even with dog slow jaguar cores but we PC gamers could have so much more by now if not for insane prices for a few more cores.

And people wonder why the PC desktop industry is in decline.

The enthusiast PC industry is actually booming, most people seem to not care about the lack of cores. It's only the portion of the PC industry in which people don't actually care about performance that's doing poorly. Just food for thought.

 

[Madpacket]

I hope that the reviews show a ~5% IPC boost and no regression in OC capability. Fingers crossed.

As long as they keep using properly fastened caps this will likely remain true. I just want lower power consumption and moar cores it'll be interesting if the cheapest hex core model can replace the i7-6700K as the preferred gaming chip. Not sure if a 5% IPC will be enough but with the right RAM the gap is already within 10% with Haswell clock for clock. The problem with existing 59xx chips is they top out a few hundred MHz lower than Skylake so yeah the OC capabilities will be of interest.

 

[Arachnotronic]

As long as they keep using properly fastened caps this will likely remain true. I just want lower power consumption and moar cores it'll be interesting if the cheapest hex core model can replace the i7-6700K as the preferred gaming chip. Not sure if a 5% IPC will be enough but with the right RAM the gap is already within 10% with Haswell clock for clock. The problem with existing 59xx chips is they top out a few hundred MHz lower than Skylake so yeah the OC capabilities will be of interest.

The problem with X99 IMO is that the PCH isn't as nice as the Z170, and it'll only be worse against Z270. The mobo makers can make up for this with auxiliary chips, and of course the CPU itself has a ton of lanes, so that helps.

I would say that the 6800K can become the preferred chip if games start using 6+ beefy cores well. As long as games don't benefit from more than 4 cores, the 6700K (and eventually KBL) will be the go to chip.

 

[24601]

The problem with X99 IMO is that the PCH isn't as nice as the Z170, and it'll only be worse against Z270. The mobo makers can make up for this with auxiliary chips, and of course the CPU itself has a ton of lanes, so that helps.

I would say that the 6800K can become the preferred chip if games start using 6+ beefy cores well. As long as games don't benefit from more than 4 cores, the 6700K (and eventually KBL) will be the go to chip.

Monopoly pricing dictates that Intel decides when consumers make the transition from 4 core for mainstream to 6/8/10/12/14 etc. etc.

As the setter of monopoly prices on the now commoditized (Since at least Nehalem) Intel desktop chip landscape, Intel is the market maker in this situation. When Intel decides, the market will have it. If Intel decides the market shall not have it, it shall not.

So decrees Intel, all hail.

 

[Madpacket]

Actually, I don't think this is true. The parts that Intel sells with more cores tend to have lower frequency per core as well as higher power consumption. Also, the mainstream parts are derived from mobile parts, and there would be very little benefit to a 6-8 core in a laptop, for example.



The 5820K, and soon the 6800K, will not be much more expensive than the top quad core on the mainstream platform. Could you please give your definition of affordable?



IMHO, the groundwork has been laid with those 8-core AMD APUs that are inside the two major consoles. However, taking advantage of a lot of cores is simply tough.



The enthusiast PC industry is actually booming, most people seem to not care about the lack of cores. It's only the portion of the PC industry in which people don't actually care about performance that's doing poorly. Just food for thought.

Sorry on mobile device so can't multi quote your reply easily.

"The 5820K, and soon the 6800K, will not be much more expensive than the top quad core on the mainstream platform. Could you please give your definition of affordable?
"

Right the 5820K was actually cheaper than the i7-6700K for a while and only recently did MSRP prices for Skylake take hold. Its total platform cost though of HEDT that drives up the prices. Quad channel memory (yes I know this isn't absolutely necessary) and motherboards that cost a good deal more money, especially compared to decent entry level Z170 boards. That's what I mean by affordable.

Intel could easily toss 6 cores in socket 1151 chips but this would make product segmentation more difficult to implement?

They have the transistor budget but do silly things like spend huge portions of the die adding sub par GPU's. I understand needing a GPU to do basic things like video decoding / encoding but beyond basic desktop needs they waste die space that otherwise could be used for extra cores, more cache etc.

"Also, the mainstream parts are derived from mobile parts, and there would be very little benefit to a 6-8 core in a laptop, for example. "

Yeah this is one of the problems. Who really games on a laptop besides college kids or people with lack of space or perhaps frequat travelers? It's mostly niche. Gaming on a laptop is super uncomfortable for extended periods of time. Intel shouldn't be forcing laptop parts on the mainstream where the optimal TDP is around 35 Watts. They should be giving us a properly optimized desktop gaming processors but we have high TDP notebook parts or gimped server parts to choose from.

If this market is really booming like you stated then we should be getting more than 5-10% performance bumps every tick or tock (or whatever Intel has switched to).

Anyways I understand the average gamer doesn't care about the number of cores but developers and publishers do, especially the good ones that can separate their games from others by taking full advantage of the hardware to gain a competitive edge. That's what gamers really care about, better games - and more hardware at the developers disposal allows for this to (eventually) take shape.

 

[strategyfreak]

If you need 10 cores it's fairly cheap.

Sorry, I want cheap 6/8/10 cores like all of you do for hobby/home use, but as a productivity/money-making tool $2k isn't that bad.

No professional trying to maximize productivity would buy an unlocked 10 core processor with a consumer-oriented X99 mobo. They’d be buying Xeon workstations and servers (which offer even more cores), and they’d be willing to pay extra for all the features associated with those platforms, such as multisocket support and ECC RAM support. At $1700, the 6950x isn’t even a discount compared to the pricier broadwell-ep lineup: http://www.anandtech.com/show/10158/the-intel-xeon-e5-v4-review/6

So it's not even cheap compared to the professional alternative.

If you’re on a budget and need cores, you’re much better off looking at the used Xeon market, where you can get even more cores for much less, and probably get equivalent performance at a lower power consumption, especially if the 6950x is overclocked.

A professional or company would only use an overclocked 6950x if it offers performance that absolutely found nowhere else, especially with the unreliability that comes with overclocking. It has a very small niche of multithreaded applications that scale very poorly beyond 10-12 cores. In all other cases, you can find a superior alternative elsewhere, as I discussed before.

This processor is directed only towards those enthusiasts whose ego won’t let them have anything less than the “largest” processor money can buy. That’s the market Intel is going after – just take a look at the leaked slides.