Intel Demonstrates 65W Broadwell-K Socketed Processors at GDC 2015

[DrMrLordX]

Intel are not stupid - if they price Broadwell-K at a significantly higher price than the 4790k, it will not sell well, since people will just wait for Skylake.

I think the price will be reasonable, though we have to see it's performance first before we can make any conclusion.

Depends on how many chips they put out, how many they expect to sell, and whom they expect to buy the things.

If Intel only intends to put Broadwell Iris Pro into the hands of people who bought the i7-4770R, then it may wind up being fairly expensive. Expect a launch price equal to i7-4770R launch MSRP - typical cost of a Z97 motherboard. There probably won't be all that many produced (possible artificial scarcity) and the SFF/HTPC crowd will snap them up greedily.

If Intel is throwing the Broadwell Iris Pro into the same arena as the 4790k, then it will be priced around the 4790k based on its stock clocks and turbo, plus a premium on top of that thanks to the iGPU, at least to start. Unless Intel can show some benefit to having the iGPU present in a system that's probably already got a dGPU, Broadwell Iris Pro may find itself being the next Kaveri . . . until Intel slashes prices on it and suffers a reduction in margin.

Personally, I am very interested in the chip, just as I am interested in seeing what kind of GPGPU software stack Intel releases to accommodate Gen8/Gen9 graphics. HSA is really hard to set up, and the full software stack isn't even available on Windows. In contrast, Intel really has nothing other than OpenCL/DirectCompute support, but that fails to distinguish Gen 8 Iris Pro from a dGPU.

 

[tenks]

Ugh, no. I don't pay attention to that pile of garbage. As NTMBK pointed out, I was talking about Haswell-R..

Ah gotcha. What I was saying was Wccftech are the ones (THE ONLY ONES) calling this an i7/Broadwell-K where as Anandtech was saying it was something else. It was unlabeled and no information was given, and Im sure Wccftech, like all those sites do, interpret it their own way for headlines and views.

 

[erunion]

Intel are not stupid - if they price Broadwell-K at a significantly higher price than the 4790k, it will not sell well, since people will just wait for Skylake.

I think the price will be reasonable, though we have to see it's performance first before we can make any conclusion.

Broadwell K isn't competing with skylake.

 

[crashtech]

Was it really necessary to create a straw man?
I obviously never said that.

I am well aware of the limitations of anecdotal evidence.

Anyways, I'm genuinely interested in hearing about why you might want to swap broadwell into your board.

I like staying current with my hardware. It's a hobby. If Broadwell-K has better IPC and can clock to at least 4.6GHz without resorting to drastic measures, I might buy one, sell my Haswell, and stay up to date while spending less than a hundred bucks. For me that is cheap fun, and that's all the reason I need.

However, if Broadwell-K turns out to be as disappointing as some seem to think it will be, then I will be stuck with Haswell for a while, since either Skylake or Haswell-E both require a whole platform upgrade.

 

[Dave2150]

Depends on how many chips they put out, how many they expect to sell, and whom they expect to buy the things.

If Intel only intends to put Broadwell Iris Pro into the hands of people who bought the i7-4770R, then it may wind up being fairly expensive. Expect a launch price equal to i7-4770R launch MSRP - typical cost of a Z97 motherboard. There probably won't be all that many produced (possible artificial scarcity) and the SFF/HTPC crowd will snap them up greedily.

If Intel is throwing the Broadwell Iris Pro into the same arena as the 4790k, then it will be priced around the 4790k based on its stock clocks and turbo, plus a premium on top of that thanks to the iGPU, at least to start. Unless Intel can show some benefit to having the iGPU present in a system that's probably already got a dGPU, Broadwell Iris Pro may find itself being the next Kaveri . . . until Intel slashes prices on it and suffers a reduction in margin.

Personally, I am very interested in the chip, just as I am interested in seeing what kind of GPGPU software stack Intel releases to accommodate Gen8/Gen9 graphics. HSA is really hard to set up, and the full software stack isn't even available on Windows. In contrast, Intel really has nothing other than OpenCL/DirectCompute support, but that fails to distinguish Gen 8 Iris Pro from a dGPU.

There are clear CPU performance benefits from having Iris Pro's 128MB l4 cache, the Anand article by anand himself demonstrated this a few years ago.

 

[Dave2150]

Broadwell K isn't competing with skylake.

Of course it's not, I didn't say it was.

It will be released before Skylake - then once Skylake launches, it will still be available for a time, in order to cater to the Z97 owners who may want to upgrade their CPU (people who are currently using an I3, or I5 on Z97, for example).

Once Skylake lauches, it will obviously completely replace Broadwell-k for those buying a new CPU/motherboard.

 

[CakeMonster]

Not much point in that... With Z97 being so recent. If it could work with Z87 on the other hand then the upgrade path would make a tiny bit of sense from 1st gen Haswell.

 

[ehume]

Oh. Now we're calling the 4770 a 1st gen Haswell. Cute.

 

[erunion]

Of course it's not, I didn't say it was.

It will be released before Skylake - then once Skylake launches, it will still be available for a time, in order to cater to the Z97 owners who may want to upgrade their CPU (people who are currently using an I3, or I5 on Z97, for example).

Once Skylake lauches, it will obviously completely replace Broadwell-k for those buying a new CPU/motherboard.

I don't think so.

Broadwell-K is going to be sold along side Skylake for a long time, the obvious explanation is that Skylake won't launch with an Iris Pro option.

 

[DrMrLordX]

There are clear CPU performance benefits from having Iris Pro's 128MB l4 cache, the Anand article by anand himself demonstrated this a few years ago.

Correct, and Broadwell itself should have maybe a ~5% increase in IPC on average (with fluctuations from one workload the next). The question is: what will be the top clockspeed for the chip? Will people with 4.5-4.6ghz Haswell quads want to jump on the Broadwell chip if the top clocks they can get on good air/AiO will be 300-400 mhz less?

Or more to the point, if the stock Broadwell Iris Pro clock is 3.2 ghz and the turbo 3.9 ghz, will Intel be able to price that at $320-$360 while keeping a straight face? People with a 4790k don't even have to overclock the thing to get excellent performance. That means no buying an aftermarket heatsink, no picking out a board based on VRMs/VRM cooling/BIOS options, etc.

The i7-4770R was walled off in its own little market where it didn't have to compete with anything in its class. It was a very powerful chip for the BGA-buying crowd, and it was priced accordingly. Broadwell Iris Pro might have to compete with the 4790k with its stock clockspeed of 4.1 ghz and 4.4 ghz turbo. It also has to contend with the 5820K whenever the buyer has the option of LGA2011 instead of LGA1150.

So Intel can't necessarily price the chip at $400+ and get away with it, especially if it proves to be slower than the 4790k when both are benched at stock. Unless, of course, they choke off supply and aim it at the same people who wanted an i7-4770R. For them, it would be a big step up at a price they've already paid before.

 

[Eric1987]

Whats the point of an HTPC CPU? Why not just pick up your run of the mill modern CPU for any tasks?

 

[witeken]

So Intel can't necessarily price the chip at $400+ and get away with it, especially if it proves to be slower than the 4790k when both are benched at stock. Unless, of course, they choke off supply and aim it at the same people who wanted an i7-4770R. For them, it would be a big step up at a price they've already paid before.

It will not have a clock speed of 3.2GHz. It will be a worthy successor of DC at a lower TDP with greatly improved graphics performance with L4 cache, priced not at $400+ but as the successor of DC (~$300).

 

[DrMrLordX]

If that's the case, then great. We'll find out more when it's actually available. And really $300 would be great compared to what you had to pay to get an i7-4770R + board.

 

[Fjodor2001]

It will not have a clock speed of 3.2GHz. It will be a worthy successor of DC at a lower TDP with greatly improved graphics performance with L4 cache, priced not at $400+ but as the successor of DC (~$300).

Any source confirming that, or are you just speculating?

 

[crashtech]

I don't particularly appreciate speculation being phrased as if it is a statement of fact.

 

[frozentundra123456]

It will not have a clock speed of 3.2GHz. It will be a worthy successor of DC at a lower TDP with greatly improved graphics performance with L4 cache, priced not at $400+ but as the successor of DC (~$300).

Seriously? I assume you think the clock speed will be above 3.2ghz. Personally, I think we will be lucky if it is not less than that. If it is a competitor to DC, then what is the point of almost immediately bringing out Skylake, especially Skylake K?

I see this as a very niche product. It should have the best igpu gaming performance, but I have the same problem with that as with AMDs apus on the desktop. Except for niche SFF or Brix like applications, you are much better off with a 200 dollar haswell quad core and a 150.00 discrete gpu. Now if (and it is a *very* big if), they can match the cpu performance of say 4690k with 50 to 100% better gpu performance and sell it for 250.00, that would be a coup(basically a free increase in igpu performance). Personally I dont see any of those happening. I am expecting a 400 dollar chip, lower cpu performance than DC, and igpu performance somewhat better than Kaveri, but still in the limbo of not being really a robust solution for1080p gaming with the most demanding titles.

I also think it is concerning that we are only a few months from the supposed BWK release, and intel gave no indications of the performance. I remember when Iris Pro was demoed, they were running a game, and touting that the performance was close to GT650m. All we got from BWK was just basically showing that it does in fact exist. No indications of cpu or gpu performance.

Really, what I would like Intel to do would be to put iris pro into all their quad core mobile parts, and make the mainstream quad core laptops (say 750.00 or less) a decent solution for low/middle level 1080p gaming. It is just my personal preference, but I think they could easily do this with 14nm, and I would much prefer this to throwing huge amounts of money at smartphones or trying to get another 45 minutes of battery life out of a super expensive ultrabook that has less than half the performance.

Edit: I want to buy my grandson a laptop for his high school graduation, and would jump on such a laptop instantly if it were available. But really no laptop without a dgpu offers the kind of gaming performance I want, so I will either buy something like the XPS13 for battery life or a 1000.00 gaming laptop with a Maxwell gpu.

 

[Ajay]

Whats the point of an HTPC CPU? Why not just pick up your run of the mill modern CPU for any tasks?

Lower TDP for a small form factor case (SFF), plus a better iGPU for limited gaming.

 

[witeken]

You might be right that BDW-K won't replace DC because of SKL-K if it's released shortly after, but I disagree on price. It will be a ~200mm² chip, so it's hardly necessary to ask a premium price. I also don't see what's the target audience for such a $400 chip.

 

[jpiniero]

Seriously? I assume you think the clock speed will be above 3.2ghz. Personally, I think we will be lucky if it is not less than that. If it is a competitor to DC, then what is the point of almost immediately bringing out Skylake, especially Skylake K?

To push DDR4 adoption with OEMs. There is no guarantee that Intel will release an unlocked model initially when Skylake-S gets released BTW.

 

[frozentundra123456]

To push DDR4 adoption with OEMs. There is no guarantee that Intel will release an unlocked model initially when Skylake-S gets released BTW.

I know the timing of SKL-K is uncertain. In fact, with the Broadwell delays, I still am skeptical of the entire skylake roadmap and dont really expect volume availability until the end of the year or early 2016. I am "hopeful" of earlier availability at best.

 

[frozentundra123456]

You might be right that BDW-K won't replace DC because of SKL-K if it's released shortly after, but I disagree on price. It will be a ~200mm² chip, so it's hardly necessary to ask a premium price. I also don't see what's the target audience for such a $400 chip.

When has die size ever been a determining factor for intel chip prices? I am not knocking intel's pricing for mainstream products, but they pretty much ask whatever they feel like and think the market will bear for specialized parts like this.

As for the market for this chip, I see it as the same market whether it is 300 or 400 dollars. It is those who want to build a very small form factor device with the best cpu and igpu performance without a discrete card, and are willing to pay extra for the small size and low power usage. It certainly will not be those who want the best gaming performance per dollar, because they will buy a cheaper i5 and add a discrete card.

It is basically the same market as for the 4770R, but at least now you are not locked in to buying a cpu/mb combo.

 

[JM Popaleetus]

There are clear CPU performance benefits from having Iris Pro's 128MB l4 cache, the Anand article by anand himself demonstrated this a few years ago.

Would any of these benefits also apply to those with discrete GPUs?

 

[Dave2150]

Would any of these benefits also apply to those with discrete GPUs?

Yes, we're talking of a double digit performance % increase in CPU tasks, even when using a dedicated GPU. Some tasks/applications will benefit more than others though.

 

[DrMrLordX]

You might be right that BDW-K won't replace DC because of SKL-K if it's released shortly after, but I disagree on price. It will be a ~200mm² chip, so it's hardly necessary to ask a premium price. I also don't see what's the target audience for such a $400 chip.

Target audience is former i7-4770R buyers. They paid out the yin-yang for that chip, on BGA with locked multipliers to boot. For them, $400 for Broadwell Iris Pro would still be a better deal than what they got.

Don't you still have to pay $599 for a Gigabyte Brix unit with 4770R barebones, no RAM no hdds no nothin? Okay you get a weak PSU, woop woop.

 

[Ajay]

I know the timing of SKL-K is uncertain. In fact, with the Broadwell delays, I still am skeptical of the entire skylake roadmap and dont really expect volume availability until the end of the year or early 2016. I am "hopeful" of earlier availability at best.

I also feel that there is too much uncertainty this late after 14 nm started going into production. With the rumored Skylake delay, I keep wondering if it's for marketing purposes, of if a the high performance quad's required a respin to meet yield and throughput requirements (or some other reason).

I suppose the real bottom line is that I don't like the new Intel model where the enthusiast crowd gets much less information in advance than b/4.
Change isn't always for the better - at least for our purposes. Instead of discussing the pros and cons of upcoming features or architectures - we end up in circular debates about speculative information from questionable sources.