[Tom's Hardware] Rumor Outlines Intel Comet Lake Features, Specifications, and Launch Window - Page 3 - Overclock.net - An Overclocking Community

Forum Jump: 

[Tom's Hardware] Rumor Outlines Intel Comet Lake Features, Specifications, and Launch Window

Reply
 
Thread Tools
post #21 of 48 (permalink) Old 07-18-2019, 10:28 AM
New to Overclock.net
 
sticks435's Avatar
 
Join Date: Jul 2011
Location: KCMO
Posts: 1,587
Rep: 64 (Unique: 52)
Quote: Originally Posted by AlphaC View Post
https://www.techradar.com/news/intel...ips-until-2021
Intel has explained how it struggled to bring 10-nanometer processors to market, while also admitting that we won’t see its 7nm chips until 2021.

However, as Swan notes, “The short story is we learned from it, we'll get our 10nm node out this year. Our 7nm node will be out in two years and it will be a 2.0X scaling so back to the historical Moore's Law curve.”

Original Source article with more fleshed out info and a single slide from Intel: https://www.pcgamer.com/uk/intel-say...chips-in-2021/

sticks435 is offline  
Sponsored Links
Advertisement
 
post #22 of 48 (permalink) Old 07-18-2019, 11:22 AM
New to Overclock.net
 
wreckless's Avatar
 
Join Date: Apr 2009
Posts: 297
Rep: 1 (Unique: 1)
intel 7nm PCIe5/6 along with (hopefully) oled 24-32" 120hz+ monitors in Q3 2021? the upgrade we deserved years ago lol

RustyShank
(9 items)
CPU
Intel Core i7 920
Motherboard
Asus P6T Deluxe
GPU
EVGA GTX 970 Sc
RAM
12GB Corsair Vengeance 1600
Hard Drive
500GB Evo SSD | 1TB WD Black
Power Supply
EVGA G3 750W
Case
Fractal Define C
Operating System
Win 10 Pro
Monitor
2 x Dell UltraSharp U2410 24"
▲ hide details ▲
wreckless is offline  
post #23 of 48 (permalink) Old 07-18-2019, 12:37 PM
New to Overclock.net
 
tyvar's Avatar
 
Join Date: Aug 2017
Posts: 114
Rep: 14 (Unique: 10)
These slides have been out for a while. And the PCGamer article doesn't mention that the planned 7nm parts in 2021 are GPU parts.

What this means is Intel has officially permanently lost most of the process tech advantage they had.

It wasn't just that they had slightly better nodes, its that they had them first by a decent margin.

TSMC is going to be doing 5nm SOC production next year, which means that AMD will have its Zen 4s on 5nm in 2021 also.

Now I know TSMCs "5nm" is in reality no such thing, but over all it is going to be roughly comparable to Intels 7nm.

Going forward the fight is solely going to be on architecture merits.
tyvar is offline  
Sponsored Links
Advertisement
 
post #24 of 48 (permalink) Old 07-18-2019, 02:50 PM
New to Overclock.net
 
Raghar's Avatar
 
Join Date: Jun 2012
Posts: 1,901
Rep: 38 (Unique: 32)
But there is second part of it. Are these 7 nm really better? AMD said it requires quite a lot of voltage to move electron trough transistor.

It looks like some companies somehow fixed up themselves on weird idea of higher density, when in reality it was about transistor efficiency and durability. 32 nm Intel planar had decent efficiency, and excellent durability. (And because it was planar, heat moved out easily into heatsink.) Higher density when industry was going from 65 nm to 45 nm meant higher energy efficiency, and better performance, and allowed more CPUs per wafer. But higher density when industry is at 14+++ nm could mean:
-higher costs per wafer/CPU
-MUCH harder to cool CPUs
-transistor that's harder to move
-electric/EM noise which can mangle data and reduce reliability
-vulnerability to lightning surges
-durability problems

Thus a result of innovation is a CPU with lower IPS, hotter, more expensive, less durable.
Raghar is offline  
post #25 of 48 (permalink) Old 07-18-2019, 04:51 PM
New to Overclock.net
 
tyvar's Avatar
 
Join Date: Aug 2017
Posts: 114
Rep: 14 (Unique: 10)
Quote: Originally Posted by Raghar View Post
But there is second part of it. Are these 7 nm really better? AMD said it requires quite a lot of voltage to move electron trough transistor.

It looks like some companies somehow fixed up themselves on weird idea of higher density, when in reality it was about transistor efficiency and durability. 32 nm Intel planar had decent efficiency, and excellent durability. (And because it was planar, heat moved out easily into heatsink.) Higher density when industry was going from 65 nm to 45 nm meant higher energy efficiency, and better performance, and allowed more CPUs per wafer. But higher density when industry is at 14+++ nm could mean:
-higher costs per wafer/CPU
-MUCH harder to cool CPUs
-transistor that's harder to move
-electric/EM noise which can mangle data and reduce reliability
-vulnerability to lightning surges
-durability problems

Thus a result of innovation is a CPU with lower IPS, hotter, more expensive, less durable.
Actually the shrinks are still large enough that even with a higher wafer cost, a design ported from 16/14/12 will be small enough that it its cheaper per chip.

take a theoretical 400mm2 chip. At equal yields, At 14nm your yields are going to be 93 good dies + 44 defects, some of which might be salvaged At 7nm that chip is 240mm2 even fudging a bit and assuming a 16mm by 16mm 256mm area, your looking at 173 good dies and 49 defects, again some of which could be salvaged. Its a 86% increase in chips made for what seems to be at this point a 60% increase in cost. Yes I know claims floating around claim 7nm is "twice as expensive" but some of the analysis and investment firms have good numbers and the actual price for TSMC 7nm is just under 10000 per wafer, and the price of 14/12 is just under 6000 per wafer.
tyvar is offline  
post #26 of 48 (permalink) Old 07-18-2019, 08:27 PM
New to Overclock.net
 
guttheslayer's Avatar
 
Join Date: Apr 2015
Posts: 3,872
Rep: 111 (Unique: 65)
Quote: Originally Posted by Raghar View Post
But there is second part of it. Are these 7 nm really better? AMD said it requires quite a lot of voltage to move electron trough transistor.

It looks like some companies somehow fixed up themselves on weird idea of higher density, when in reality it was about transistor efficiency and durability. 32 nm Intel planar had decent efficiency, and excellent durability. (And because it was planar, heat moved out easily into heatsink.) Higher density when industry was going from 65 nm to 45 nm meant higher energy efficiency, and better performance, and allowed more CPUs per wafer. But higher density when industry is at 14+++ nm could mean:
-higher costs per wafer/CPU
-MUCH harder to cool CPUs
-transistor that's harder to move
-electric/EM noise which can mangle data and reduce reliability
-vulnerability to lightning surges
-durability problems

Thus a result of innovation is a CPU with lower IPS, hotter, more expensive, less durable.
There is a reason why they are heading for 3D stacking. Its the only way forward that can last a while.

guttheslayer is offline  
post #27 of 48 (permalink) Old 07-19-2019, 07:57 PM
New to Overclock.net
 
tyvar's Avatar
 
Join Date: Aug 2017
Posts: 114
Rep: 14 (Unique: 10)
Quote: Originally Posted by guttheslayer View Post
There is a reason why they are heading for 3D stacking. Its the only way forward that can last a while.
The heat problems with 3d stacking are even worse then the problems with shrinks.

There is a reason why the only things that are currently stacked is memory which is a lot less sensitive to temps then a GPU/CPU die.

My gut feeling is the only thing that will end up being stacked frequently is ram like in HBM, but it will still be mounted to the side of a CPU or GPU die on a interposer in 2.5d stacking.
tyvar is offline  
post #28 of 48 (permalink) Old 07-20-2019, 04:05 AM
Madonion.com
 
Damage Inc's Avatar
 
Join Date: May 2015
Posts: 119
Rep: 3 (Unique: 3)
Quote: Originally Posted by TK421 View Post
they catched up with the bios modders who put 9900k in older mb

oy vey...






but on a more serious note, how can anyone believe this leak? highly doubt any of this info is even 1% accurate.
Oy vey! This needs to get shoah'd ASAP!

On a more serious note, I smell a fake. The disinfo is probably coming from Intel themselves. If Intel can't get their backdoors sorted out with their next lineup, I'm not buying their crap ever again.
Damage Inc is offline  
post #29 of 48 (permalink) Old 07-20-2019, 10:35 PM
User
 
Join Date: Oct 2011
Location: Canada
Posts: 143
Rep: 2 (Unique: 2)
Quote:
fabricated on a further enhanced 14nm process, up to 10 cores, new chipset and socket
I believe that rumor only based on these info because Intel is so boring nowadays. Tho if I see rumors of incredible IPC gain, I won't believe that one for sure, these "new" cpus will only exist as a stop gap and as a speed bump against AMD. Prices probably won't be much lower than now, IF they lower the price.
Warrax22 is offline  
post #30 of 48 (permalink) Old 07-22-2019, 01:40 AM
Waiting for 7nm EUV
 
tpi2007's Avatar
 
Join Date: Nov 2010
Posts: 11,486
Rep: 900 (Unique: 504)
This rumour in particular is pretty believable.

Max of 10 cores on a ring bus so that you get the best gaming performance? Check, we've known at least since Ivy Bridge that Intel's ring bus tops out at 10 cores.

No claims of IPC improvement? Check, it's just another iteration of Skylake, maxed out on the ring bus in terms of cores, with the predictable improvements going into getting the most out of single threaded performance (perhaps 5.1 or 5.2 Ghz on 1 or 2 cores) and then getting the highest clocks possible on 4 and 6 core loads in order to get the best real world performance in games.

125w TDP for the 10 core part(s)? Check, seems pretty believable. Well, at least as believable as the 9900K being a 95w TDP CPU, but that's another story.

DDR4 2666 support with possible 2933 support once further validation is complete? Seems believable, although actually underwhelming considering what their IMC's can do as of right now.

Further improvements in terms of iGPU codec support? Check, it's the usual yearly updates.

Still only 16 lanes of PCIe 3.0 from the CPU, i.e, no PCIe 4.0? Check.

New socket? Check, it doesn't get more 'Intel' than this.

Availability only in 2020? Check, The 9900KS will be arriving in Q4 of this year, so the timeline makes sense.


The price? The top of the line will be competing with the 3900X, winning in games by a bit more than the 9900K, but it will probably lose a lot of tests in multithreaded, unless they really crank up the clocks and throw efficiency out the door for good. They will have to price it the same as the 9900K at most. AMD has a better platform, a head start and most probably better efficiency considering that Intel will still make these on 14nm++(+). Also, AMD will have a 16C/32T CPU for the mainstream socket by then.



Last edited by tpi2007; 07-22-2019 at 04:05 AM.
tpi2007 is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off