Intel Skylake / Kaby Lake

Page 563 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

TheF34RChannel

Senior member
May 18, 2017
786
309
136
I don't think Icelake is going to be compatible even if they had planned for it at one point. I would think they would want to pair Icelake with Tiger.

No at the K level. I really think Core as an architecture is nearly tapped out clock speedwise, so while they will throw an extra 2 cores in there for Icelake you probably won't get much at ST or it will even be slower. And then Tiger would just improve clocks some more but not much if any at higher speeds.

First paragraph: that's my thinking as well as this would fit better with Intel's thinking.

Second paragraph: that'd be good, if the 8700K is top chip of the entire CFL line-up, because that's what I want :D I'd hate to learn a fictional 8760K releases after I made my purchase.
 

nvgpu

Senior member
Sep 12, 2014
629
202
81
Intel is only launching 6 Coffee Lake SKUs per the slides including the 2 overclockable SKUs, 8700K and 8600K and Z370 chipset this year, in January at CES 2018 Intel will launch the rest of the Coffee Lake SKUs including Pentiums and Celerons and the more future proof fully featured Z390 Cannonlake PCH chipset and consumer H3x0, business Q3x0, B3x0 and lower end H1xx chipsets, how anyone can get confused is quite perplexing. It's not rocket science, you know.
 

Timmah!

Golden Member
Jul 24, 2010
1,416
630
136
Apparently 7920x is coming out a week from now, on 12th August - if the information is true. Bit weird date given the fact thats Saturday. I wonder if there is any connection between the release date and the number of cores, i mean 12C allegedly released on the 12th day of August, 18C supposedly released on the 18th day of October... maybe 14C and 16C variants will come on 14th a 16th too.

Anyway, i demand more leaks! :p Why has nobody delidded the thing yet?
 
  • Like
Reactions: Ajay

kirbyrj

Member
Aug 5, 2017
122
27
61
Intel is only launching 6 Coffee Lake SKUs per the slides including the 2 overclockable SKUs, 8700K and 8600K and Z370 chipset this year, in January at CES 2018 Intel will launch the rest of the Coffee Lake SKUs including Pentiums and Celerons and the more future proof fully featured Z390 Cannonlake PCH chipset and consumer H3x0, business Q3x0, B3x0 and lower end H1xx chipsets, how anyone can get confused is quite perplexing. It's not rocket science, you know.

So Intel is going to allow CFL on the low end H1XX chipsets but not the Z170/Z270?
 

Ajay

Lifer
Jan 8, 2001
15,406
7,833
136
Maybe I missed it, but has Intel announced a release date for CFL-S. Looks like September from the graphic sweepr posted, but I'd like to know more precisely.
 
  • Like
Reactions: Pick2

TheF34RChannel

Senior member
May 18, 2017
786
309
136
Maybe I missed it, but has Intel announced a release date for CFL-S. Looks like September from the graphic sweepr posted, but I'd like to know more precisely.

Nah, the only thing that Intel announced was a forthcoming 8th generation CPU. The remainder comes from leaks, none of which show release schedules. The latest and greatest show a production schedule, for what that's worth... (and it's worth something, for sure)
 

wildhorse2k

Member
May 12, 2017
180
83
71
Apparently 7920x is coming out a week from now, on 12th August - if the information is true. Bit weird date given the fact thats Saturday. I wonder if there is any connection between the release date and the number of cores, i mean 12C allegedly released on the 12th day of August, 18C supposedly released on the 18th day of October... maybe 14C and 16C variants will come on 14th a 16th too.

Anyway, i demand more leaks! :p Why has nobody delidded the thing yet?

There is definitely something fishy going on with 12C+. 7940X is supposed to come like 2 months after 7920X. But its the same thing, except with 2 less cores deactivated. If they can manufacture 7920X, they can 7980XE as well as its the full version. I'm definitely waiting to see 7980XE before I buy anything.
 

jpiniero

Lifer
Oct 1, 2010
14,571
5,202
136
Performance = Clock speed * Perf/clock.

I'm not expecting much of an IPC improvement if any, and while I do also think that the ST clock speed shouldn't be too much lower than the 8700K it is possible that it could be enough that the Icelake 8C ends up slower.

There is definitely something fishy going on with 12C+. 7940X is supposed to come like 2 months after 7920X. But its the same thing, except with 2 less cores deactivated. If they can manufacture 7920X, they can 7980XE as well as its the full version. I'm definitely waiting to see 7980XE before I buy anything.

The decision to do the 14-18 core model was probally like literally days after Threadripper was seriously rumored, and it does take time to formally productize it, etc.
 

wildhorse2k

Member
May 12, 2017
180
83
71
The decision to do the 14-18 core model was probally like literally days after Threadripper was seriously rumored, and it does take time to formally productize it, etc.

Certainly it won't take 2 months to productize it as it already exists as Xeon. I'm expecting either Intel attempts to use solder or will introduce 2066V2.
 

Timmah!

Golden Member
Jul 24, 2010
1,416
630
136
There is definitely something fishy going on with 12C+. 7940X is supposed to come like 2 months after 7920X. But its the same thing, except with 2 less cores deactivated. If they can manufacture 7920X, they can 7980XE as well as its the full version. I'm definitely waiting to see 7980XE before I buy anything.

I guess it takes quite longer to produce sizable number of 7980XEs than 7920Xs, thus 2 month difference in release dates.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
z270 says that as well iirc correctly.

No it doesn't.

https://www.google.ca/search?q=Inte...gVQKHR3oABkQ_AUIDCgD&biw=1280&bih=915#imgrc=_

Z370/CNL-PCH is indeed Next Gen.

http://cdn.wccftech.com/wp-content/...ne-SSD-Roadmap-2016-2017-3D-XPoint-Memory.jpg

The current Optane Memory branded drives are named Stony Beach and based on the presentation originally supposed to be called Optane Memory 8000p.

The next gen is said to come with "CNL Platform" and called Carson Beach with x4 NVMe speed update.
 

coercitiv

Diamond Member
Jan 24, 2014
6,176
11,809
136
Last time they actually did that was . . . hmm.
Last time that technically happened was Broadwell. Sure we can argue it was only this corner case - high performance desktop - that saw ST regression, we can also argue that i7-5775C was a 65W TDP chip, but the truth is it was unlocked and could not clock high enough to match previous gen ST performance. On top of that it was late and had poor firmware support.
 
Mar 10, 2006
11,715
2,012
126
No it doesn't.

https://www.google.ca/search?q=Inte...gVQKHR3oABkQ_AUIDCgD&biw=1280&bih=915#imgrc=_

Z370/CNL-PCH is indeed Next Gen.

http://cdn.wccftech.com/wp-content/...ne-SSD-Roadmap-2016-2017-3D-XPoint-Memory.jpg

The current Optane Memory branded drives are named Stony Beach and based on the presentation originally supposed to be called Optane Memory 8000p.

The next gen is said to come with "CNL Platform" and called Carson Beach with x4 NVMe speed update.

Z370 isn't CNL PCH. It's KBL Refresh PCH. Z390 is the actual CNL PCH.
 
  • Like
Reactions: TheF34RChannel
Aug 11, 2008
10,451
642
126
Last time that technically happened was Broadwell. Sure we can argue it was only this corner case - high performance desktop - that saw ST regression, we can also argue that i7-5775C was a 65W TDP chip, but the truth is it was unlocked and could not clock high enough to match previous gen ST performance. On top of that it was late and had poor firmware support.
Actually, that depends on the workload. In some applications, even certain games, the extra cache made it equal to or faster than Haswell. In any case, the difference was very small. A reasonable overclock IIRC was about 4.2 plus 5% IPC over Haswell = 4.4 ish, while even the 4790k could usually do 4.7 or so, so less than 10% regression worst case.
 

nvgpu

Senior member
Sep 12, 2014
629
202
81
https://www.notebookcheck.net/fileadmin/Notebooks/Acer/Switch_5_SW512-52-5819/uhdupdateneu.jpg

uhdupdateneu.jpg
 
  • Like
Reactions: Sweepr

VirtualLarry

No Lifer
Aug 25, 2001
56,315
10,031
126
Will be kind of weird seeing Intel UHD graphics, does that mean that they will finally natively support HDMI2.0? If not, it's really a misnomer, kind of like how the "Coppermine" Pentium didn't use copper interconnects yet.
 
Mar 10, 2006
11,715
2,012
126
Last time that technically happened was Broadwell. Sure we can argue it was only this corner case - high performance desktop - that saw ST regression, we can also argue that i7-5775C was a 65W TDP chip, but the truth is it was unlocked and could not clock high enough to match previous gen ST performance. On top of that it was late and had poor firmware support.

5775C wasn't a proper successor to 4790K, and it was not positioned as such.
 

nvgpu

Senior member
Sep 12, 2014
629
202
81
KBL-R and CFL is still Gen 9.5 GPU and unlike Gemini Lake which has integrated native HDMI 2.0, neither of these have integrated native HDMI 2.0 support per the slides, still needing external LSPCON chip for HDMI 2.0 output support.

I wonder if Cannonlake Gen10 GPU has DisplayPort 1.4 support, because the slide mentioned Titan Ridge, next gen Thunderbolt 3.0 controller with DisplayPort 1.4 support.
 

VirtualLarry

No Lifer
Aug 25, 2001
56,315
10,031
126
So, let me get this straight? Intel "re-names" their iGPUs as "UHD", but still lacks the proper ports to actually drive an UHD display (at a proper frame-rate)? LOL. Typical smoke-and-mirrors Intel.
 

coercitiv

Diamond Member
Jan 24, 2014
6,176
11,809
136
Actually, that depends on the workload. In some applications, even certain games, the extra cache made it equal to or faster than Haswell. In any case, the difference was very small.
Even if it was a 1% performance regression, it was regression nonetheless.

5775C wasn't a proper successor to 4790K, and it was not positioned as such.
It was an unlocked part priced accordingly. Let's not be naive, it was positioned differently because it couldn't deliver, not the other way around. I really doubt Intel had clock targets for Broadwell lower than Haswell, but it happened.

Now, would I expect a future Intel CPU to drop in ST performance over it's predecessor? Certainly not. Could it be possible, based on previous experience? Yes, it could be a possibility.
 
  • Like
Reactions: Kuosimodo