/intel/ is dead

Isn't Intel supposed to release 10 nm in 2H/2017? Or was it 2018?

>>58913421

i do remember reading an article where intel themselves said changing to 10nm was going to be a huge challange, as the traces will be only a few atoms wide.

maybe its harder to do than anyone thought. or they are just holding off to suck more money out of hapless uninformed consumers.

It's about god damn time we had a bit of stagnation. This means coders will start to fucking optimize again instead of "oh, lets use this inefficient but convenient bullshit and just let time take care of it cuz CPUs will get faster"

>>58913039
I don't get it. Cannonlake is still set to be released as of new yesterday. Cannonlake is still 10nm. Coffee lake, a 14nm chip, has been pushed up. Coffee lake went from 2018 to a release date that looks to be simultaneous with cannonlake.
https://s21.q4cdn.com/600692695/files/doc_presentations/2017/2017_Intel_Investor_Meeting_Renduchintala.pdf
This is what he cites. It doesn't even mention the chip sizes outside of a few graphs which put 10nm at 2017 like we've all known for 2 years now

How many years will it be until we get a CPU with 2xSIngleThread performance of the current 7700k?

>>58913421
10 nm will be mobile and low-end only, codenamed Cannonlake. Desktop parts will remain 14 nm, codenamed Coffee Lake.

>>58913600
Forever

>>58913039
THANK YOU AMD

>>58913584
What is there to not get? 10nm yields are shit

>>58913603
10nm is going to be physical limit of Moore's Law

>Born just in time to watch to watch computers stagnate

Quantum computing is magic and elves

>>58913921
Quantum computers do not replace CPU's for traditional computing. If they were ever to beomce mainstream, they would be like a 3rd processor(CPU/GPU/QPU).

Of course they are not going to prove their mastermind's law (moore or whatever) was wrong

Will Jewish people (Israel and Intel) going to lose to Asian? This is an interesting match.

>>58914024
Like the Math Processors of the olden days.

>>58913600
30

>>58913506
This soooooo much.

File: 1442277093472.jpg (48KB, 400x400px) Image search: [Google]

48KB, 400x400px

>>58913039
>tfw still running a 2600K

>>58913506
I appreciate your optimism. But we all know it won't happen.

>>58913452
>as the traces will be only a few atoms wide
>>>>>>
1nm = 1.000.000fm = 1.000.000 Atoms
10nm = 10.000.000 Atoms wide
>few atoms

[spoiler]roughly, contain your autism please[/spoiler]

Isn't the Snapdragon 835 10nm?

>>58914387
dude atom sizes are within the regime of Angstrom (10^-10 m = 100,000 fm = 0.1 nm).
You're 5 orders of magnitude off.

>>58914401
Yes

>>58915506
Does Intel just suck then?

>>58915566
The SD 835 is a few mm^2 big, while Intel's chips are generally 150-250mm^2

>>58914387
I have no idea what I'm talking about but maybe if I act obnoxiously enough someone will think I do: The post.

I thought AMD already has plans for 7nm? Or was that Qualcomm?

>>58914387
A silicon atom is 1.46 angstroms, or 1.46*10^-10 m wide.

10 nm would be 1*10^-8.

10 nm fabrication would be a mere 68.5 silicon atoms across.

>>58915799
I figured it was a physicist as well.

>>58916048
>>58915799
how to be triggered when someone wants to contribute and I see that others disprove him in a topic I truly have no fucking idea about: the shitposting

>>58915884
Its AMD's token foundry: Global Foundries. 7nm SOI FinFET process developed by IBM.
Intel's 7nm fab, Fab42, is still underconstruction.
TSMC and Samsung both have 7nm nodes in the works as well.

>>58916120
That post wasn't contributing. It was a shitpost coming from a false sense of superiority. It was annoying and retarded and should have been shit on even harder

>>58916220
like all your posts

>>58914387
You did your math wrong

Atoms can be massive as wide as 0.125nm

>>58916120
Jokes aside, it's nice you meant well at least. No hard feelings.