It's likely going to be quite close to the 8700K frequencies. Probably the same single core turbo, and slightly lower all core turbo. Plus the extra cache. It should perform like the 8700K single core and pull nicely ahead of it multi core.
Are you for real? They can't even manufacture at high volume a tiny mobile SOC with 2 cores CPU, the GPU totally disabled and no improvements in efficiency compared to 14nm. Because yields are terrible now, imagine 1 year before.
Exactly. Intel cannot manufacture 10nm today even if its life depended on it. The problems are not simple. charlie has publicly tweeted there are 3 problems and david schor has agreed that he has heard the same too. This is a nightmare scenario for Intel. If Rome lands by mid-2019 its going to be a bloodbath in servers first followed by client PCs. Lisa Su stated at the recent JP Morgan conference that Zen 2 will launch in servers first and there will be significant volume in 2019. Even if we assume its H2 2019 Intel has no chance if they put 14++ based 28c CSL-SP against 7nm based 64c Rome.By mid 2020 AMD could very well hit its highest server share and be around 30% or even more and will only be limited by fab capacity in terms of how much share they can take. Lisa Su has stated that AMD will reveal Zen 2 manufacturer and performance (i am guessing atleast core count,cache and I/O specs) by year end. We will know by year end when Zen 2 will launch. If I had to guess its early Q3 2019. The question which remains to be answered is where is Rome manufactured - TSMC or GF ?
I think GF will be the manufacturer for Ryzen CPUs and APUs as AMD have certain wafer targets to meet. GF needs to up their game as if AMD take share from Intel they win too. Here is hoping GF can deliver for AMD as they will not get a better opportunity ever to take huge market share in a relatively short span of time.
Actually CNL-U is the first officially launched 10nm product, before CNL-Y.
How is it going to go to "8700k frequencies" when they can't rate the base within 800mhz of the 8700k?
It will probably turbo 1 core to 4.7 and otherwise perform substantially worse. Take a look at the overclocks people got on the other 8 core ring bus chips, they are substantially worse than the hex core chips and the clocks generally go down across the board.
Your optimistic outlook will not stand up to reality.
While that's true, Whiskeylake-U is the mass volume successor to current Kabylake-R.
You have to remember there's two distinct -U lines, 28W one called CFL-U, and 15W called KBL-R. KBL-R has a 14nm successor, which means that can't be the early one.
Since they have 10nm volume now(yes, I know its limited), I think its reasonable to expect another small-scale launch to happen H1 based on Icelake. 28W U fits the bill nicely because its rarely used over 15W parts.
Beemster:
The 2.6GHz 8C part is an Engineering Sample. It won't be something you'll get to buy. Final product will clock higher to make it acceptable as a replacement to the 6 core Coffeelake 8700K. We don't know the clocks, but you should know.
...yep...2.6GHz all core can't be right........but they did have to lower all core freq a lot when they went to 6C from 4C......why will it be different going to 8C?...........they both have the 14nm++ device.....Explain
You understand what I said right? Ignore 2.6GHz altogether, because Engineering Samples are *not* sold to you. It's a test chip.
They will have to lower the all-core Turbo when going from 6C to 8C, but they can keep the top Turbo clocks same. Running 8 cores require more power.
forget turbo.........I understood you PERFECTLY.....it is you who did not understand me..............Intel dropped all core freq. on 4 core from 4.2GHz to 3.7GHz on 6C and still raised the power a bit . Since 8C and 6C both have the 14nm++ device, why do you think they will not have to lower all core freq again? What 8C all core freq and power do you predict?.......a number please
You are quoting Base frequencies, not all core Turbo. You got it wrong. 4.2GHz on 7700K and 3.7GHz on 8700K is BASE.
7700K can run at 4.4GHz when all 4 cores are active and 8700K can run at 4.3GHz when all 6 cores are active.
I am very sorry, but your ....... are annoying as hell. Would you please stop it? Thank you!OK............got it........I was wrong then............so what do you predict for 8 core?....(freq and power).....why did they even bother then with a 6 core if 8 core all core turbo could go higher than 8C all core AMD 2700X?
OK............got it........I was wrong then............so what do you predict for 8 core?....(freq and power).....why did they even bother then with a 6 core if 8 core all core turbo could go higher than 8C all core AMD 2700X?
Beemster:
The 2.6GHz 8C part is an Engineering Sample. It won't be something you'll get to buy. Final product will clock higher to make it acceptable as a replacement to the 6 core Coffeelake 8700K. We don't know the clocks, but you should know.
Intel are in desperate need of a something that can compete with AMD's small multichip approach and Infinity Fabric. And that's why they needed Jim Keller.
It turned out to be enough.Only Intel knows.
They went with 6 core because they probably thought it was enough, and/or that was the best they could get out during the short time they had to plan.
It turned out to be enough.
The scores of the 8 core chip look pretty good even at the ES frequencies.
They do... that's what EMIB is, plus the theoretical Feveros/stacked die later. The question is of course how far and diverse the integration goes between all of Intel's products, and that's where Keller comes in.
You probably won't see a Core EMIB product until 2020 at the earliest however.
Yes, what I meant was the integration of all the components, not the Infinity Fabric solution itself (although EMIB is yet to be proven in anything besides GPU to memory communication. .
Will icelake be mesh or ring bus?.
For sure all core turbos will be significantly higher than that for retail, wouldn't be surprised at 3.0ghz base though.Genuine Intel(R) CPU 0000 @ 2.60GHz (8C 16T 2.59GHz, 8x 256kB L2, 16MB L3)
....2.6GHz all core????? when AMD is at 3.7GHz all core on 2700X with a likely 2800X in reserve??................can't be that bad....must be a chipset issue......dam them Core cores are power hungry........like I've been pointing out for some time..........the more cores....the harder it bites them........so of course no 10nm SERVER parts till God knows when............got to get to the multichip approach and a new core design
Do you mean alot better than skylake X or ring bus latency?EMIB is also being used by the FPGA for die<->transceiver. But if you want Core<->Core, Core<->GPU, Core<->FPGA, etc, Intel's not there yet.
Could go either way although if it is the mesh it will likely be a lot better on latency.