Not on desktops, of course. But at the price they get for mobile quads, they certainly could afford to make it standard across the board. I was not aware that Coffee Lake U will be 4+3e. Does anyone really know for sure though, and will it be a mainstream part or simply for high end 1500 dollar ultrabooks/surface tablets?
In any case, this still illustrates, IMO, the way Intel is shooting themselves in the foot with market segmentation. Why not make edram standard on mobile 45w quads as well, which could much better utilize it with their higher TDP, than on a U chip that probably will be highly thermally limited anyway in any application that utilizes heavily the cpu and gpu at the same time? An 800 dollar 45 watt quad core laptop with iris pro would be a very nice overall performance, light gaming system.
How are they shooting them selves in the foot with market segmentation? Their PC biz continues to rake in obscene amounts of cash even as the PC market struggles in no small part due to market segmentation.
If you're gonna put a big GPU and an eDRAM cache on a processor you really should sell it to customers willing to pay. Don't just give it away.
There is another factor you fail to consider: Iris Pro simply flopped. Intel may give consumers extra, but for OEM it's way more profitable to get a cheaper eDRAM-less chip, slap a cheap ass dGPU and call it a day.Shintai, is that you??? Heaven forbid intel should actually give the customer more value when they dont have to. In any case, from a business standpoint, yes, I see your point, at least in the short term. But long term, especially if AMD makes a strong comeback with zen, and in a flat/declining market, I think segmenting to the nth degree to wring every possible dollar from the consumer will eventually come back to bite you. I mean, would it really kill them to try to give consumers a bit extra, at least in a mid-end laptop or above, to drive customer good will and slow the rate of sales decline?
On that Skylake E5 2699 v5 ES, I should point out that the v4 model is 22 cores at 140 W. Maybe it's a typo since you would have to think that the TDP would have to be much higher if it's 32 cores at 2.1 Ghz.
Leaks point to 205W TDP parts.
Seems about right since those are basically 16 SKL-Us.Leaks point to 205W TDP parts.
Yea, but again though, I dont think consumer who buys a prebuilt and doesnt want to change the PSU is going to want to or be comfortable with cloning the hard drive and installing a SSD.
Are you saying you think that i3 line is about to be canned?Also looks like death of 2+2 on desktop. Quad is now the new baseline.
Are you saying you think that i3 line is about to be canned?
Okay, I can kind of see how this could work, it will be interesting to see what they do on clockspeed and features to distinguish between the i3 & i5 line.Wouldn't be surprised if Intel's 2018 desktop lineup looked like this:
Core i7 = 6C/12T
Core i5 = 4C/8T
Core i3 = 4C/4T
Pentium/Celeron = 4C/4T, GeminiLake based.
Okay, I can kind of see how this could work, it will be interesting to see what they do on clockspeed and features to distinguish between the i3 & i5 line.
But long term, especially if AMD makes a strong comeback with zen, and in a flat/declining market, I think segmenting to the nth degree to wring every possible dollar from the consumer will eventually come back to bite you.
Ah, good old competition.Wouldn't be surprised if Intel's 2018 desktop lineup looked like this:
Core i7 = 6C/12T
Core i5 = 4C/8T
Core i3 = 4C/4T
Pentium/Celeron = 4C/4T, GeminiLake based.
Whatever they do the point is they can respond to zen in even a handfull of months if they wanted to. Labelling their portfolio and setting price so it gives a perfect response. Using the same dies.Ah, good old competition.
Core i5 would (still) be good value proposition with OC.
So there's no distinguishable benefit other than optane.
That's good to know, I'll stick with my z170 given I'm not bothered about optane or using more than one graphics card.
Whatever they do the point is they can respond to zen in even a handfull of months if they wanted to. Labelling their portfolio and setting price so it gives a perfect response. Using the same dies.
I am not so sure they actually have to do something as radical as i3 4c. Zen is bound to be lower performant but also quite small. 4c8t in 80mm2 is possible. But when an high freq i3 can beat it soundly in st and is adequately small using a big costly 4c die to fight zen is not attractive nor needed if price is right.
It also depends if Intel gets some serious mobile wins that take up capacity. But anyway they can play the cards fast and precise. Talk about perfect maneuverability and control.
Ok If thats the case then the numbers of 4c zen will be next to nothing if any at all so then its hardly relevant to discuss it. That is if they can fuse off to get functional 6c dies instead of 8c ?4C/8T Zen will almost certainly be a disabled 8C/16T Zeppelin die. The 126mm^2 quad Coffee Lake die will not have trouble on the cost side of things.
Ok If thats the case then the numbers of 4c zen will be next to nothing if any at all so then its hardly relevant to discuss it. That is if they can fuse off to get functional 6c dies instead of 8c ?
Chipset lanes are wired through DMI link, so m.2 ssd RAID on Z170 and Z270 should be pretty much exact same.And 24 lanes instead of 20. Z270 would be able to do m.2 ssd RAID without sacrificing gpu pcie lanes.
Wouldn't be surprised if Intel's 2018 desktop lineup looked like this:
Core i7 = 6C/12T
Core i5 = 4C/8T
Core i3 = 4C/4T
Pentium/Celeron = 4C/4T, GeminiLake based.
Thats extremely unlikely, intel would never do that unless amd becomes extremely competitive with zen...they just launched a pentium with 2c4t...not a single chance on earth they will make it a 4c by 2018My bet is
I7 6/12
I5 6/6
I3 4/8
Pentium 4/4