[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr / vt ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Due to resource constraints, /g/ and /tg/ will no longer be archived or available. Other archivers continue to archive these boards.Become a Patron!

/g/ - Technology

View post   

[ Toggle deleted replies ]
File: 319 KB, 1024x512, investing-v3d.jpg [View same] [iqdb] [saucenao] [google] [report]
58857417 No.58857417 [Reply] [Original] [archived.moe] [rbt]


>Intel Corporation today announced plans to invest more than $7 billion to complete Fab 42, which is expected to be the most advanced semiconductor factory in the world. The high-volume factory is in Chandler, Ariz., and is targeted to use the 7 nanometer (nm) manufacturing process. It will produce microprocessors to power data centers and hundreds of millions of smart and connected devices worldwide.


>> No.58857448


You can see the change in language already.

"Diversity" is out, nationalism is in. MAGA motherfuckers.

>> No.58857454

Weekly reminder Intel is working on a discrete GPU

>> No.58857474
File: 15 KB, 300x214, 1465171629979.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.58857482

>Intel us factory
>Samsung us factory
Companies better hurry up and keep immigrating before the 30% tariffs wreck the consumer PC market harder than Intel or Samsung ever could

>> No.58857491

>look at how much we "plan" to invest in a fab that already exists
>never mind the fact that we laid off half of our IC designers for pajeets

Literal virtue-signalling for blue-collar degenerates.

>> No.58857493

It will probably be for AI or enterprise stuff.

>> No.58857512

>blue-collar degenerates.
what did he mean by this

>> No.58857520

My dad said "GPU" and in the context of competing with AMDs GPU division
I don't know what it'll be used for but it will have mainstream competition when it's released

>> No.58857531

So Larrabee 2.0?

>> No.58857764

Its 2017.
Construction on Fab42 began in either 2011 or 2012 if I recall, its still not complete. The facility has to be completed, it has to go through a complete purge and filtration before humans can even step foot inside, before they can even start step 1 of running a test chip. They have years of work to do.
Cannonlake is 10nm.
Ice Lake is 10nm.
Tiger Lake is 10nm.
There will likely be another 10nm product line as well.

7nm chips will be reaching market in 2021 if everything goes well, if there are no delays, or major yield issues. 2022 seems far more realistic.

Meanwhile Global Foundries is starting work getting their advanced SOI based 7nm node online later this year. Risk production is slated for 2018, and production capacity could come in late 2018 to early 2019. Similarly TSMC's 7nm node is aimed at 2019, and will employ EUV.

>intel thinks they're advanced and they're already behind the curve
inb4 fanboys bitching about meaningless back end area scaling

>> No.58857842

Glo Fo literally lies about their manufacturing die size

>> No.58857869

>Being this uneducated faggot

GloFo's 7nm is equivalent to Intel's 10nm process

Just like Samshit's 14nm process is really 20nm with FINFET or TSMC's 16nm is really 20nm with FINFET

Stop being a dumb faggot

>> No.58857914
File: 144 KB, 1280x960, ed9de1d1998f3a266cb38f16eb1971ba_1377810701.jpg [View same] [iqdb] [saucenao] [google] [report]

Oh boy, I haven't heard this exact line of violent stupidity before. No foundry lies about anything regarding front end feature size. Not intel, not Samsung, not TSMC, not Global Foundries, not UMC, not any of the dinky small 90nm fabs still up and running. No one.
Your inability to understand a public facing marketing name is not their problem.

Stupid little fanboy bitching about meaningless back end area scaling right on cue.
Samsung and TSMC's 14/16nm nodes are their "transitional" FinFET nodes. They use a smaller front end on an existing back end to reduce the cost of bringing a smaller node online. The density of the back end does not impact the electrostatic characteristics of the transistors on the front end. All it impacts is area scaling, the number of candidates per wafer.

Global Foundries' 7nm process is an advanced SOI based FinFET node, radically different from the bulk silicon FinFETs used across the industry today. Intel is falling far behind the industry when even TSMC and Samsung will be using GAA devices before intel has a single 7nm test chip on hand.

>> No.58857947

Glofo 14nm hotter and less efficient than any other 14nm
Not to mention that it isn't actually 14nm, just finfet and close enough for glofo

>> No.58858019

Yeah, and Nvidia is working on a x86 SoC

>> No.58858027

Advanced Micro Devices, Inc. is grateful for your contributions and has deposited 5 shekels to your account.

>> No.58858057

>Glofo 14nm hotter and less efficient than any other 14nm
Its Samsung's process first of all, 14nm LPP.
The process has shown fantastic characteristics. The Ellesmere ASIC pulls 110w at 1266mhz, highest clocking GCN design to date. Half the power of the Radeon 7970 while clocking higher, having more CUs, and more transistors per CU, all while giving gratuitous voltage to keep yields as high as possible for cheap entry level cards.
Nvidia's 14nm GTX 1050 ti also shows extremely well on the process.
AMD's Ryzen CPUs are clocking as high as 3.6ghz/4ghz at launch, yet again extremely impressive. A 8c/16t CPU with a 3.6ghz base clock and 95w TDP.

>Not to mention that it isn't actually 14nm, just finfet and close enough for glofo
The fact that you don't even understand who designed the process, or why node names are what they are shows you're just regurgitating things you don't understand.

Trying to paint me as an AMD fanboy is just pathetic. Proves you're to dimwitted to handle facts, and conversely to the point you're trying to make, proves that you yourself are emotionally invested in defending intel.
Do you perhaps post on the Anandtech forums? Your type is really common there.

>> No.58858082

AMD doesn't have money to pay you, they're near bankruptcy, why do you continue defending them?

>> No.58858111


>> No.58858125

>most advanced semiconductor fab in the world
>still using silicon


>> No.58858127
File: 276 KB, 1280x720, image.jpg [View same] [iqdb] [saucenao] [google] [report]

>half the power of the Radeon 7970

>> No.58858130

Intel is the reason AMD struggled and closed fabs in my home country.


>> No.58858147

I'm posting simple facts, sad little fanboy.
It doesn't matter if its about TSMC, Samsung, Global Foundries, intel, or any other fab. I'm not blindly regurgitating talking points to defend Samsung and downplay TSMC. I'm not praising IBM's investment in SOI and denigrating the rest of the industry for sticking to bulk for so long. I don't have that childish fanboy mentality that you have. Intel is falling behind the curve, and for them to say that their 7nm facility will be the most advanced in the world is just marketing fluff.
Deal with it.

>> No.58858153

Intels 8c/16t CPU from two years ago goes well above 4GHz, "yet again extremely impressive"

>> No.58858221

Top end Haswell-E has a 3ghz base clock, and 3.5ghz turbo with a 130w TDP. They don't offer any 4ghz 8core/16 thread chips, no one is talking about whats possible when overclocking.
Stop thinking like a reactive emotionally invested child. Instead of behaving like a rational adult you're just shitposting like a tumblrina because I said something about your favorite company that you don't like.

>> No.58858257

D3D_13 based in Larrabee
Make DirectX great again

>> No.58858270

When ryzen is here Intel will be bankrupt in under two years. This time amd won't fuck things up. Note my words Intel Jews.

>> No.58858309

>nobody is talking about overclocking
>4GHz zen

>> No.58858318

Yeah DX12 is a failure with proprietary and patented Async™ Compute™, why other vendors (Intel, Nvidia, Qualcomm, Imagination) with dx12 capable gpus cannot use.

>> No.58858328

4ghz is the turbo frequency of the 3.6ghz base clock chip, anon.
Seriously, pull your head out of your ass. A mentally healthy adult shouldn't act so childishly.

>> No.58858362

Zen is 3.4GHz at best moron, everything else is overlcocking, if you knew anything about CPU architecture you'd know its pipeline is too short and GloFo sucks shit so its frequency is low

>> No.58858374

It was said to have a 3.9GHz boost
There is no proof or even evidence of a zen processor ever running at 4Ghz

>> No.58858403
File: 150 KB, 1360x1020, C1aelJ4UUAEG0za.jpg [View same] [iqdb] [saucenao] [google] [report]

F3 OPN chips on the floor at CES with 3.6ghz base clock/3.9ghz turbo.
F4 OPN chips are 4ghz turbo.

Stop shitposting, child.

Thats flat out a lie.

>> No.58858421


It's fine, you can cry now.

>> No.58858431

You see that 39/36N?
That processor does not go to 4Ghz and no zen CPU has been shown with the extra stepping for what is said to be a 4GHz boost

>> No.58858455

Is 7nm the limit?

>> No.58858461

Dahnald will cut H1B visas. They won't get away with this.

>> No.58858496

No he already did, and the order is already suspended
It was a shitshow for tech companies the last few weeks, people literally had to go home and renew their visas

>> No.58858513

What kind of hillbilly state judge can block a presidential order? Shouldn't only the senate or supreme court be able to do that?

>> No.58858554

Thats a screen cap of an F3 chip from the floor at CES, exactly as I just stated.
F4 OPNs are a fact, the top end 3.6ghz chip has a 4ghz turbo.

This is entirely aside from the point. I didn't make AMD's upcoming CPU lineup central to my argument, I only offered it as one of three examples. The issue at hand is that intel's foundry business no longer has the enormous technological process yield they've held in years passed. Saying now that their 7nm facility will be the most advanced in the world is PR nonsense, especially if they'll be years behind other fabs, numerous other fabs.

Grasp at straws elsewhere.

Not by a long shot. Gate topology will start getting much more diverse, 7nm nodes are going to require EUV machines which are still in a state of just barely working. The power they're outputting currently is still pretty low, and without them the alternative is incredibly complex quad patterning, unless intel manages to get immersion tech ready for volume production.
Though we might just transition to some self assembling devices, or heavily utilize self assembly in critical structures to reduce dependency on photolithography techniques all together. We'll be making progressively smaller devices for years to come.

>> No.58858571

We surely can't go 1nm, right?

>> No.58858614

Not that guy, but I think 3nm is the limit and then weird extreme physics take over.
Feel free to correct me, though.

>> No.58858624

That hasn't been used to design a CPU in like a decade

>> No.58858634

i live in az and i hate this. chandlers pajeet population has already exploded and this is just going to make it worse since all they're going to do is import more fucking pajeets.

>> No.58858644

Depends on how much voltage you're using, and how quality the conduction is
Supercooled devices could probably go well below the room-temperature path limit, but electromigration already happens on 7nm

>> No.58858658

There is absolutely nothing that technically limits a 1nm wide gate from existing. The the limitation that people talking about is sort of a theoretical model of log gate length, but there are tons of caveats to that. Silicon nanowires smaller than 5nm have been made, 10nm nanowires were first made quite a number of years ago even.
If silicon or enhanced materials like SiGe are no longer effective channel materials they'll simply be replaced with something else. 1nm wide carbon nanotube transistors were made in a lab, very strong proof of concept:

Ah, silly me. All those lith masks and light sources must just be for fun.

>> No.58858662

where do you think you are?
it's a well known fact that muh gigahurtz doesn't mean shit.

>> No.58858696

he won't. hes still a business man at the end of the day. he'll do some pajeet tier "reform" where american citizens have to higher priority but still littered with enough loopholes to not change a damn thing.

fucking shit entire east valley of the metro area is going to be little india.

>> No.58858856

you know, in the case of AZ, i can honestly say that intels excuse of not being able to find americans is actually true.

asu is a dogshit tier of a university. you either go there for liberal arts, business, accounting, and ironically, law. its engineering wing is decent, but not the pinnacle of asu. most people also go elsewhere for masters and higher so all you get is bachelors and companies like intel like masters and phd's. then asu has like a % graduation rate after four years. 45% after six. most people don't graduate and drop out by their second year. asu is number one for enrollment but that's because all you need is a 2.5gpa and any score on your sat's.

then all the liberals that graduate from the shitholes of likes of berkeley don't want to go to az because "ewww, they have guns and its hot there" "I mean, do they even have a wholefoods and organic gardens?"

so really, i can actually believe intel on this. one of the few exceptions.

>> No.58858868

They are.

Name (leave empty)
Comment (leave empty)
Password [?]Password used for file deletion.