Intel 10nm

How did they fuck it up so bad?

Attached: FinFETTable.png (678x464, 27K)

Other urls found in this thread:

fool.com/investing/2018/07/07/is-intel-trying-to-hide-something.aspx
twitter.com/NSFWRedditImage

you want the short or long version

Both if possible. Long to BTFO /v/ manchildren

What does it mean for 5G?

Attached: 1505924784_10nm-intel-chip-process_story.jpg (760x435, 56K)

In Short:
Intel management wanted a massive shrink that would ensure their domination of the process market and attract customers to their fabs. The load combined with budget cuts and layoffs in the manufacturing group killed 10nm.


In Long:
There were three primary mistakes Intel made, Cobalt Metal Layers, Contact Over Active Gate (COAG), and plain ole hubris.

In 2013, Intel was late. The Self Aligned Dual Patterning (SADP) required by the feature size of 14nm had a bad learning curve, yields were very bad at first, to the point where Broadwell was mostly a paper launch in 2014, 2 quarters behind schedule. This was not a critical problem and it was fixed gradually, such that Skylake was not delayed. Behind the scenes though, the long ramp time created a problem. As Intel has only a single (large) process development team, not leapfrogging teams, the 14nm delay led to a delay of 10nm. The specifications it would shoot for were not set in stone until 2014.

Managers gave them a difficult task. To win mobile they had to be power efficient and dense. To win desktop they needed to be fast. To win servers they needed excellent yields. And above all they needed to be better than the competition to attract new customers.
Inorder to reach the goals set by management, the manufacturing group had to get creative. To that end a number of techniques never put into a production process before were adopted. COAG, SAQP, Cobalt, Ruthenium Liners, Tungsten contacts, single dummy gate, etc. This push is directly what led to the death of the process. Of those, only really COAG and Cobalt are causing the issues. I'll go into the specific problems next post.

If anyone is to blame, its the management, and their firing of the CEO with a bullshit reason shows the board will not accept responsibility for the companies failings. They will not come clean in the foreseeable future. Their foundries are virtually dead after all the firings and cost cutting.

>record profits
>still need to cut the corners so that they can hit a certain profit target in the short term which triggers massive bonus for C-suite execs

The idea with Contact Over Active Gate is that instead of extending a gate such that it connects up with a contact to the side (thus using space on the side), the Contact stretches directly from the metal layer to the gate, rather than laying ontop the substrate. This means there is NO room for error on manufacturing. The slightest misalignment leads to fucked contacts. Thermal expansion, microvibrations from people walking nearby, changes in air pressure, imagine a cause, and it'll affect yields. I bet you the bloody position of the Moon can affect it.
This kills the yields.

To hit the targets Intel set, a minimum metal pitch of 36nm was selected. When you have Copper wires on a process they need to have a liner around them, this prevents diffusion, electromigration, and other nasty electrical fun. But this liner needs to be a certain thickness, so when the overall size of the wire gets smaller, the liner takes up a larger portion of it. Below 40nm it was thought that Cobalt would have superior electrical properities, despite it having a higher bulk electrical resistance. Its far more resistant to electromigration and needs a miniscule barrier to prevent it, while its resistance decreases at a slower rate as the wire size gets smaller.
However, Intel overlooked two key problems: ductility/malleability, and thermal conductivity. Even at those tiny levels, Copper wires would be able to handle thermal expansion mechanical loads, bending and stretching ever so slightly as a processor made its rounds. And copper is Very good at transferring heat, letting the lower metal layers sink heat into the upper ones. Meanwhile Cobalt is hilariously brittle and has a sixth the thermal conductivity as Copper. On operation hot spots start to form, heat can't get away, brittle nature creates microfractures, and higher voltage to cross the fracture boundaries. Means the voltage/frequency curve is hilariously bad.
This kills the performance and power usage.

They literally are creating 10 nm micro housefires that burns their own chips.

So where does it leave us at?
10nm was meant to launch end of 2015, after 14nm this was pushed to 2016. It is now Q3 2018 and the only 10nm chip is a minuscule dual core made in a one-off batch of 100k units that took 6 months to assemble. Yields are sub 1%, the GPU doesn't function, and power usage is higher than 22nm.

TSMC and GloFlo are both ramping their 7nm processes, and while they're comparable to the 10nm in density, they actually fucking work. They both are using SAQP for the transistors, but the choice of 40nm metal pitch allowed for SADP for that particular layer. GloFlo is using Copper with a Cobalt/Tantalum liner while TSMC is using straight Copper/Tantalum. Neither fucked with COAG or Single Dummy, but GloFlo is using Cobalt contacts.

TSMC 7nm is in volume production now, GloFlo will enter it in a few months. Consumer products on both of them will enter the market in mid-2019. They're both going to outperform Intel's 10nm. Their manufacturing group failed at all levels. To make matters worse they didn't even bother backporting their improved cores to 14nm. Icelake has been inhouse since early 2017. The design is finalized, they just can't make the damn thing and didn't backport it. Management thought they couldn't fail.

Cobalt contacts (to replace tungsten ones)****

yeah basically

That was an interesting reading user. Mind if I make a screencap of it?
Can Intel unfuck the fabs soon?

It is in my personal opinion that their 10nm process will never be financially viable.

Cannonlake is virtually dead. Ice Lake-U and Y (sub 15w mobile dual cores) should come out still in mid-2019. I doubt Ice Lake-S (quads) will see anything beyond a paper launch ala Broadwell quads. Ice Lake-SP is a dead proposition, theres no way they're going to be able to make a server core based off that 10nm. Tiger Lake and Sapphire Rapids? Who the fuck knows.

Intels chance lies with 7nm, but it won't arrive until 2021 at the earliest.


btw, after Cascade Lake is Cooper Lake; Kaby Lake-U is being replaced by Whiskey Lake, which'll be replaced by Comet Lake to run alongside the Ice Lake-U/Y dual cores. Amber Lake will replace Kaby Lake-Y until Ice Lake-Y. All those new codenames are 14nm, and all are just tweaked Skylake.

Also the whole 14+++++ shit is meaningless. They're standard PDK updates. Everyone in the industry does them. It wasn't until recently that they started being branded.

>all these lakes
What the FUCK

What they will sell then?
They also appear to not improve their designs too much. For example, Kaby Lake didn't have a controller that supported LPDDR4 ram or at least lift the 16 Gb max of LPDDR3. Apple got shit literally because Intel.
I'm wondering if Apple isn't talking with AMD right now to license Zen Core design or something along these lines.

who cares, Apples primary business is their mobile lineup anyways

Whats wrong user? Won't you be just peachy trying to decide between a laptop with Coffee Lake-R, Whiskey Lake, Amber Lake, and Gemini Lake?

Attached: 1527222155321.jpg (728x665, 51K)

I at least care. Apple has the brand power to flip mindshares and crap like that.

Apple mobile lineup includes things like Intel made modems and shit, that are made in their fabs.
They already walked away.

Apple cancelling 5G Intel modems is only a rumor currently. What I posted above is concrete information.

ONE NOTHING WRONG WITH ME

Attached: 1492855909643.png (1228x1502, 944K)

Capitalism at it's finest. That was sarcastic.
In truth, one can see the vast majority of companies (especially American firms) taking short term decisions for short term gain without realizing the long term harm to their own business. I work at a small company and I see my boss (the CEO) doing this all the time.
>doesn't pay irreplaceable employees well
>doesn't have good PC hardware even though that's how we make money
>5 mbit internet. I know it's an ISDN but it's for 4 people.
>cant even have the same hardware for development the customers get.

INTEL IS FINE 14nm++++++++++++++++++++++++++++ will save us

Attached: 1528342425631.jpg (626x657, 81K)

>after Cascade Lake is Cooper Lake; Kaby Lake-U is being replaced by Whiskey Lake, which'll be replaced by Comet Lake to run alongside the Ice Lake-U/Y dual cores. Amber Lake will replace Kaby Lake-Y until Ice Lake-Y
WHAT THE FUCK WHAT'S THIS 14NM+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

I bet many Intel execs are already drowning in Whiskey lake

hi charlie

Shut Up and Buy (((Our))) Stock, Goyim.
fool.com/investing/2018/07/07/is-intel-trying-to-hide-something.aspx

>What they will sell then?
new chipset for each lake rebrand
ehehe

It means LG are in trouble

Attached: DELID.png (2518x1024, 468K)

Disgusting kikes

Cause they tried multiple things at the same time on the same node which shot everything to hell.

Will AMD really have 7nm cpus next year?

Attached: 1530502962013.png (381x353, 40K)

Nah they're going to have 7nm GPUs

>Ashraf Eassa owns shares of Advanced Micro Devices. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.

Yes. 7nm EPYC2 samples at the end of this year.
7nm Zen2 and EPYC2 will be in full availability next year.

7nm Epyc samping THIS year.
Q1 2019 7nm Ryzen most likely.

yes, 7nm gpus coming this year in a few months for ML as well

So 3800x then? By this time next year?

Or a 3990x

>64 TR coars
But why?

>2021
>10nm++
I TRIED SO HARD AND GOT SO FAR

why not

Attached: Haruhi6.jpg (757x910, 434K)

Because shitters are going to bench 28 core xeons and 32 core Zen processors for muh gaymen

who cares? muh gayme already runs at 300fps on a haswell i5

>EPYC2
Sounds like Apple could upgrade the Mac Pro for once.

Attached: 1531236094905.jpg (636x722, 70K)

What pisses me off about this is how they manage to survive of fucking loans and stuff when in a real free market they would go under in a second because of their shitty decisions.

Finance is the cancer that fucks modern capitalism.

Charlie is usually breddy gud.

> bench 28 core xeons and 32 core Zen processors

They won't bench them at all because the zen chips will have a 1ghz+ clockspeed advantage and no way will Tom's throw away its marketing shekels by making AMD look good.

Piece of shit jews are bribing everyone to shill their retarded garbage.

wait for intel 34 core hehe

>no way will Tom's throw away its marketing shekels by making AMD look good.

Attached: Capture.png (1060x769, 250K)

They do have some Mac Pro revamp slated for 2019 or 2020 so we'll see.

>10nm+
>10nm++
>Implying they can get first-gen 10nm to work at all

kek, poor Intel

Why Sakurako isnt jealous?

If their schedule isn't fluff they'll produce 7nm Vega20 (half-rate FP64 + microarch improvements + better HBM & memory subsystem) this year,
They'll start showing us 7nm Zen2 halfway through FQ4 (end of november-december) and release it late March or early April,
assuming Glofo is actually able to produce their 7nm at acceptable yields.

>screencap

Post it faggot.

Based info-user, how do you know so much about the manufacturing details? This is great stuff.

MAKE MOAR LAKES

Eh. Maybe. The product is very niche. I think the only truly unforgivable things "management" does is not having fucking hardware for testing/development and paying critical, irreplaceable employees shit wages.

A chipset change each year keeps the goyim in fear.

>Crush the market without question for 14 years
>Get complacent
>Don't push out 10nm the right way since you were focused on producing corporate-friendly SKUs
Can you blame them?

Attached: 1447161347847.jpg (1846x1923, 607K)

Gross mismanagement and physics catching up.

Intel through that their 10nm process would be the shit and overcome all of the incoming physical problems that have been plaguing semiconductors since 45nm.

Intel has fuck up so badly that TSMC and Samsung are going surpass them as the leaders of fabrication tech and Intel will most likely never regain its lead again.

I think radeon changed to tsmc

The processs looked so good in paper. But again, what wins in the end is what exists.

Heh, mainland chinks might actually beat Intel to 7nm

why does UMC suck so much?

Because they're tiny

UMC market share is similar to Global Foundries

They produce commodity shit though. They aren't producing high revenue parts like GloFo is.

No wonder amd is slow as shit to release b450 mobos. They have no competition in the desktop segment outside of elite gaymen fags who delid.

why hasn't intel stock crashed yet?

Kikes protect themselves.

AMD hasn't won the performance and IPC/watt metrics just yet...
Also their server platform lacks in absolute float operations, among other performance metrics, and has only just begun to be adopted. That area takes the most time to impact.

>AMD hasn't won the performance and IPC/watt metrics just yet...

Attached: 1458575698021.jpg (250x182, 6K)

intel is a blue chip stock. Major hedge funds invest in them because they're a historically insanely profitable company, and its a safe bet that they will continue to be insanely profitable in the future.
Minor issues within a company won't sway a blue chip very much. Intel could totally cancel their 10nm node, stick on 14nm++++++ for three more years, and they'd still be a blue chip.

The real reason is that investors typically dont know shit about the company they are investing in.

Brand name blue-chips stocks tend to hold very well, then drop like a rock when things become so obvious that even normans can realize they are underperforming. When Intel’s server marketshare drops to ~80% next year from the ~92% it currently holds, stock price will fall by half.

> t. Proud investor of AMD at $5

delid dis

Tomorrow will make a decent infographic with the relevant text as the only material.

Attached: Intel thread.png (1346x9412, 1.4M)

>Crush the market without question for 14 years
Underage please

>But Why
To make Intel Cry
ayyy lmao

>AMD hasn't won the performance and IPC/watt metrics just yet...
oh but they did

>socket 1331
>socket AM4
wait, these aren't the same?

They are. Whoever wrote the specs messed up.

SHUT UP GOY

Attached: 1527629778452.jpg (679x758, 54K)

Seeing we have some educated people here, i gotta ask. What happens after we inevitably reach 1 nanometer dyes? What scale do we use next? Is it quantum processors after that?

More corz
> quantum processors
Fucking kek. No those cannot be used like regular processors. Don't fall for the hype. They have specific uses.

they "invested" all their money in something else lmao

Attached: 1528231011159.png (1571x901, 1.09M)

Based Intel hiring Rajakesh Shekelstein feminists, truly the most progressive company to ever exist!!!

See also: General Electric stock crashing only last year and half despite warning signs of massive losses at its has power generation and oil well service divisions that first became known years ago

Intlel didn't give them enough shekels for 2018

What do is trying to reach more gigahurtz and cores with things like graphene or miniaturized vacuum tubes.
Yes. Vacuum tubes in nanometer escale. Those are a thing and can reach the speeds that Netburst only dreamed.
Image a 15 ghz netburst experimental rehash on miniaturized vacuum tubes. Perhaps would still a highly inefficient architecture, but at least their pipeline would work as intended.

Is it tomorrow already, user?

Yes. I'm working on it.

10.000 hours in Paint.

Attached: Why Intel is fucked beyond redemption.png (1400x1608, 399K)

>no yurus
Why do you do this?

Attached: 1530191217394.gif (439x335, 378K)

>node scaling is dead
>no new semiconductor materials ready for large-scale production

what will be the new innovation in semiconductor industry? I bet 100 rupees on inter-chip optical interconnects but we probably won't see that in a decade or more

I have to protect them from the disaster Intel is making.

We need more Chinatsu horny moments.

>wait for intel 34 core hehe
can't wait to sink in 20k on a single monolithic CPU that finally put out as much heat as a re-entering spacecraft.

The industry is already utilizing some III-V materials, and GAA will immediately succeed FinFETs in a couple years.