• Hey there Guest! This holiday season you have a chance to win some cool AnandTech swag! Check out “The Holiday Hotlist” giveaway, which you can find here.

14nm 6th Time Over: Intel Readies 10-core "Comet Lake" Die to Preempt "Zen 2" AM4

cortexa99

Junior Member
Jul 2, 2018
20
2
41
#1
https://www.techpowerup.com/249956/...s-10-core-comet-lake-die-to-preempt-zen-2-am4

If Intel's now-defunct "tick-tock" product development cadence held its ground, the 14 nm silicon fabrication node should have seen just two micro-architectures, "Broadwell" and "Skylake," with "Broadwell" being an incrementally improved optical shrink of 22 nm "Haswell," and "Skylake" being a newer micro-architecture built on a then more matured 14 nm node. Intel's silicon fabrication node advancement went off the rails in 2015-16, and 14 nm would go on to be the base for three more "generations," including the 7th generation "Kaby Lake," the 8th generation "Coffee Lake," and 9th generation "Coffee Lake Refresh." The latter two saw Intel increase core-counts after AMD broke its slumber. It turns out that Intel won't let the 8-core "Coffee Lake Refresh" die pull the weight of Intel's competitiveness and prestige through 2019, and is planning yet another stopgap, codenamed "Comet Lake."

Intel's next silicon fabrication node, 10 nm, takes off only toward the end of 2019, and AMD is expected to launch its 7 nm "Zen 2" architecture much sooner than that (debuts in December 2018). Intel probably fears AMD could launch client-segment "Zen 2" processors before Intel's first 10 nm client-segment products, to cash in on its competitive edge. Intel is looking to blunt that with "Comet Lake." Designed for the LGA115x mainstream-desktop platform, "Comet Lake" is a 10-core processor die built on 14 nm, and could be the foundation of the 10th generation Core processor family. It's unlikely that the underlying core design is changed from "Skylake" (circa 2016). It could retain the same cache hierarchy, with 256 KB per core L2 cache, and 20 MB shared L3 cache. All is not rosy in the AMD camp. The first AMD 7 nm processors will target the enterprise segment and not client, and CEO Lisa Su in her quarterly financial results calls has been evasive about when the first 7 nm client-segment products could come out. There was some chatter in September of a "Zen+" based 10-core socket AM4 product leading up to them.
14nm++++?:eek:
 
Aug 25, 2001
42,095
106
126
#2
10-core RingBus? Latency numbers should be interesting... I wonder at what point, RingBus stops making sense for Intel? They've already introduced Mesh topology for the core/ cache communications, in their newest HEDT / Server CPUs.

How soon is this supposed to happen? Many people just bought 8th-gen 6-core, or 9th-gen 8-core Intel CPUs recently. Man, are they gonna be pissed, if Intel comes out with a 10-core CPU in 3 months. (Speculation on the timing.)
 

Qwertilot

Golden Member
Nov 28, 2013
1,311
7
106
#3
Also power consumption of course. Their top clocked 8 cores are already rather crazy.
 

NTMBK

Diamond Member
Nov 14, 2011
8,050
28
126
#4
Did Intel seriously never start backporting Icelake to 14nm, even after so many years of 10nm delays? We've had the same core since 2015, and it's going to be in use until at least 2020? Wow. I think every gamer would take an 6 or 8 core with higher IPC Icelake core over a 10 core Skylake.
 

TheGiant

Senior member
Jun 12, 2017
251
8
76
#5
WTF is really happening in Intel. Did they keep only the senior managers in 2015 and the engineering stuff went away?
 

coercitiv

Platinum Member
Jan 24, 2014
2,906
106
136
#6
10-core RingBus? Latency numbers should be interesting... I wonder at what point, RingBus stops making sense for Intel
Judging by their previous LCC/MCC/HCC designs from Broadwell and other older gens, the upper limit seems to be around 10-12 cores per ring (that's where they split to multiple busses). Therefore 10 cores should be just fine.

They should've called it Meteor Lake though: it's gonna get very hot near the impact area.
 
Last edited:

Kaloi48

Junior Member
Jun 2, 2016
16
4
51
#7

JoeRambo

Senior member
Jun 13, 2013
627
11
136
#8
Intel needs to find every guy in upper management, who voted against backporting IceLake to 14nm and fire them immediately. Exposing Intel to 5 years with no new micro architecture should not go lightly with any board.
 

.vodka

Golden Member
Dec 5, 2014
1,013
9
136
#9
Intel needs to find every guy in upper management, who voted against backporting IceLake to 14nm and fire them immediately. Exposing Intel to 5 years with no new micro architecture should not go lightly with any board.
I think you mean manglement, there's no other way to describe how they've mismanaged things.

This is approaching Hector Ruiz levels of failure.
 
Mar 10, 2004
27,788
45
126
#10
This doesn't seem possible unless Intel has a new trick up it's sleeve. The 9900K is already on the ragged edge. I think there is some big info missing from this leak.
 
Mar 10, 2004
27,788
45
126
#11
This says the rumors have it using a dual ring bus / dual die.
Didn't we see some "5 core" leaks somewhere?

Perhaps the most interesting detail offered is that the 10-core CPU will be built with a dual-ring bus interconnect. A dual-ring bus architecture tends to introduce some latency as the separate core rings have to flow through a buffered switch in order to communicate with each other. This is seen in Intel Xeon E5 v4 HCC (high-core count) processors.

The need for a dual-ring bus design is peculiar; a single ring design should be able to handle 10 cores without much trouble. However, a dual-die design would help with thermal dissipation, and a dual-ring bus would be necessary to connect the dies. Intel has been criticized for the poor thermal performance of the Core i9-9900K 8-core CPU. Overheating problems with the 9900K obviously limit overclocking potential and require beefy hardware cooling to prevent. A dual-die design may mitigate some of these issues.
https://www.notebookcheck.net/Intel...0-core-Comet-Lake-S-desktop-CPU.368371.0.html
 

coercitiv

Platinum Member
Jan 24, 2014
2,906
106
136
#12
This says the rumors have it using a dual ring bus / dual die.
Oh dear god, say it ain't so. I take my Meteor Lake joke back, this is clearly Frankenstein Lake: a zombie built with Skylake limbs, glued and reordered in a Zen-like shape and form.

So much for the advantages of monolithic design.
 

Atari2600

Senior member
Nov 22, 2016
656
29
106
#13
Image removed.

First of all, we do not allow images with no
commentary. Secondly, this was trolling.


If you want to discuss what Intel previously
said about AMD server CPUs, do so. However,
you can't just drop a pic of super glue and run.


AT Mod Usandthem
 
Last edited by a moderator:

Kaloi48

Junior Member
Jun 2, 2016
16
4
51
#14
Mar 10, 2004
27,788
45
126
#15
I think there is some misunderstanding or poor translation. The original source of that article didn't say that the 10-core Comet Lake will be utilize a dual ring bus design.
I think there's been a lot of misunderstandings with these rumors...but that's what we get for treating them as if they are reliable. Many times they are not.
 

scannall

Golden Member
Jan 1, 2012
1,249
44
136
#16
I think there's been a lot of misunderstandings with these rumors...but that's what we get for treating them as if they are reliable. Many times they are not.
Like all rumors, they can be interesting. Even amusing. But it's usually a mistake to assume it's entirely truthful.
 

coercitiv

Platinum Member
Jan 24, 2014
2,906
106
136
#18
I think there's been a lot of misunderstandings with these rumors...but that's what we get for treating them as if they are reliable
Like all rumors, they can be interesting. Even amusing. But it's usually a mistake to assume it's entirely truthful.
I'm having a hard time identifying who exactly in this thread treated this rumor as "reliable" and "entirely truthful".

If anything, the source of the rumor itself warns that Intel roadmaps are currently being updated quarterly: we're more likely to predict weather rather than Intel's CPU lineup in 2019.
 

Kaloi48

Junior Member
Jun 2, 2016
16
4
51
#19
OK I find the source(it seems to be):
https://www.ptt.cc/bbs/PC_Shopping/M.1541681106.A.F30.html

the author 'chrisdar' said at Nov.8 that Comet Lake-S 10C was confirmed at a 'meeting'. But so-call double ring is just a speculation by others.

He also said B365 chipset is coming which is look like H310C rebrand.
Yeah, other posters already talked about the hypothetical "double ring" design before Comet Lake-S 10C was mentioned by "chrisdar".

 
Mar 10, 2004
27,788
45
126
#20
I'm having a hard time identifying who exactly in this thread treated this rumor as "reliable" and "entirely truthful".

If anything, the source of the rumor itself warns that Intel roadmaps are currently being updated quarterly: we're more likely to predict weather rather than Intel's CPU lineup in 2019.
You don't think some of the replies indicate a belief in the rumor? I do.
 

coercitiv

Platinum Member
Jan 24, 2014
2,906
106
136
#21
You don't think some of the replies indicate a belief in the rumor? I do.
That would imply that at the time of your first replies on this thread, you were 100% convinced the next Intel mainstream SKU would be a multiple die solution.

Bringing the dual die rumor in discussion and then insinuating people in this thread were genuinely convinced after just 2 other replies is a very weird approach to say the least.
 

maddie

Platinum Member
Jul 18, 2010
2,198
63
136
#22
You don't think some of the replies indicate a belief in the rumor? I do.
So what? Big deal.

I for one, in the light of recent actions, find this quite plausible. Do we forget so soon the recent shenanigans? Does the 5GHz 28 core monster in this quarter ring a bell?

Is it some heresy to make fun of Intel and their present state. I often feel as if some posters are defending the faith, and trying to ridicule the other. These reactions by some are purely political and apparently are devised to keep the faithful engaged.
 
Mar 10, 2004
27,788
45
126
#24
That would imply that at the time of your first replies on this thread, you were 100% convinced the next Intel mainstream SKU would be a multiple die solution.

Bringing the dual die rumor in discussion and then insinuating people in this thread were genuinely convinced after just 2 other replies is a very weird approach to say the least.
Nope. I mentioned the article with the dual die comment when I saw it. Prior to that, I would have had no idea to make such a suggestion. I didn't even know Intel had used a dual die solution any time recently.

However, I do see "Skylake Z" on the Lenovo sheet. Never heard of that before. What is it?
 
Mar 10, 2004
27,788
45
126
#25
So what? Big deal.

I for one, in the light of recent actions, find this quite plausible. Do we forget so soon the recent shenanigans? Does the 5GHz 28 core monster in this quarter ring a bell?

Is it some heresy to make fun of Intel and their present state. I often feel as if some posters are defending the faith, and trying to ridicule the other. These reactions by some are purely political and apparently are devised to keep the faithful engaged.
I will actually be surprised if 9550U is Comet Lake 14nm...
 

ASK THE COMMUNITY