[ 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: 1.72 MB, 1510x1006, 14nm+++++++++++++++++++++++.jpg [View same] [iqdb] [saucenao] [google] [report]
67081063 No.67081063 [Reply] [Original] [archived.moe] [rbt]

Two more years before Intel's 10nm. AMD will be moving to 7nm in less than twelve months. How does this make you feel?

>> No.67081077

AMD will be moving to 7nm in less than 9 months anon

>> No.67081507

just wait for ZEN3 in 2020 to kill all those Intlel faggots 7nm with even lower power consuption and heat

>> No.67081539

Only a pleb would conflate a naming system with a process equivalency.

>> No.67081540

I dont give a fuck because im not a consumerost pig

>> No.67081596

It makes me feel fine because AMD will botch it without a doubt

>> No.67081611

protip: intel's 10nm was already less dense than glofo's/TSMC's 7nm before they had to gut it

>> No.67082316

In 2019 Intel will announce brand new Lake Lake on 14++++++
In 2023 Intell will give up and spin out its fabs

>> No.67082826

Only a pleb believes on Intel 10 nm working

>> No.67084382


>> No.67085150

I need 7nm everything RIGHT FUCKING NOW

>> No.67085206

How much do AMD shills get paid, the whole board is filled to the brim with useless consumerist threads like this one, neck yourself OP

>> No.67085471
File: 127 KB, 600x600, intlel.gif [View same] [iqdb] [saucenao] [google] [report]


>> No.67085482

Mama Lisa showing (((them))) who's boss

>> No.67085494

Fucking finally.
Intel does not move if you don't kick it forward.

>> No.67085507


Intel is run by sassy black women and pajeet diversity hires now. All the competent white men left for AMD, ARM, etc in 2015

>> No.67085514

I have to pay for electricity so I dont care

>> No.67085626

At this point I don't really care about node shrinks much, once you're in the ball park of 7 nm it gets harder and harder to make further improvements just due to the physical limitations of transistors and the atoms involved. I think overall Ryzen has an edge right now architecturally, but it would be interesting to see if Intel can leverage Altera for consumer grade CPUs and promote programmable logic as a mainstream performance enhancer for software rather than relegating it to their HPC product lineup.

>> No.67085685

Is ryzen 2400g made with 10nm process?

>> No.67085698

how many more generations will intel cpoos be vulnerable though?

no fucking way are they going to can the stuff they had planned for the next 1-2 years just to redesign the vulnerable speculative stuff

>> No.67085709

>doing anything right

>> No.67085751

14nm. The APU's are first gen Ryzen.

>> No.67085767


>> No.67086212

10nm is still vulnerable so maybe when they move onto 7.

>> No.67086239

>AMD will be moving to 7nm in less than twelve months
Winter of THIS year, for fuck's sakes. How many more times do I have to reapeat this to you, you dumb fa/g/s?

>> No.67086328

I guess you haven't heard, but it is 12nm now

>> No.67086339

not it's not, see>>67085751

>> No.67086393

>"AMD stealing our sales, how do we fix it? "
>"Add two more cores! "
>"AMD is ahead of our 10nm technology, how do we fix it? "
>"Add two more nanometers!"

>> No.67086428

Until they release a completely new uarch, which could be 2022-2023

>> No.67086669

i loved that intel made simple FMA calculations into a new instruction set and called them NEURAL
just for them to be not even 2% faster

>> No.67086824

It doesnt even matter because 10nm was slower than 14nm+++++++ even before it was gutted

>> No.67086848
File: 540 KB, 1100x1002, 1517943791961.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67086973

cooper lake? more like pooper lake hahaha

>> No.67087018


>> No.67087029

So when Intlel will finally be on 10nm in 2020 AMD will be on 5nm? lul

>> No.67087093

If you read the non-meme tech sites they all say that Intel's 10nm is the same as TSMC's 7nm in terms of density and performance.

>> No.67087144

There are 2 key differences.
First of all. TSMC 7nm works. Unlike dumpsterfire of Intel 10nm. You can have all the performance you want. If it doesn't work, it doesn't matter
Second. That was before Intel gave up on their 10nm jump and decided to scale down it a bit, to a point of semi-accurate calling it 12nm.
So to sum up. We have 7nm process that works and should products come out in a few quarters vs 10 nm that doesn't has to be scaled down and products come in 20xx.

>> No.67087176

IIRC The only known Intel 10nm chip has worse boost clock than that of 14+++. Which means that it is not in any shape, way or form "same" as jump from 14nm to 7nm

>> No.67087208

Wrong. Raven ridge is not summit ridge

>> No.67087274

Wouldnt that mean that intel will actually lose performance in a generation when switching to 10nm?

>> No.67087325

not exactly. If you look at the specs, 10nm is slightly worse then most 7 nm ones, especially in layer counts.

>> No.67087360

As of right now they would

>> No.67087470
File: 161 KB, 921x865, JUST.jpg [View same] [iqdb] [saucenao] [google] [report]

Intel have already stated themselves that they expect to see a performance regression in the first lineup of products using 10nm, so it may be another 12-18 months after that before they can ship a competitive product on it. The whole thing is a complete fucking disaster.

>> No.67087503

I can't believe just how accurate that tweet was.

>> No.67087737

That's what they want you to think. When AMD comes out with 10nm and 7nm intel will be right there, with CPUs that will run circles around poozen3. You will see.

>> No.67087786

Intel and Israel have been too busy working on creating chips that are easily manipulated and watched without user knowledge.

>> No.67087807

10 < 7

>> No.67087883

>just WAIT tm

>> No.67087884
File: 14 KB, 166x166, 440295919529.png [View same] [iqdb] [saucenao] [google] [report]

Massive erection to be honest

>> No.67088032
File: 6 KB, 350x314, 1502432194480.png [View same] [iqdb] [saucenao] [google] [report]

>Multi-core doesn't matter!
>Productivity doesn't matter!
>Price/performance doesn't matter!
>Performance per watt doesn't matter!
>Power usage doesn't matter!
>Temperatures don't matter!
>Soldered dies don't matter!
>Stutters don't matter!
>Streaming doesn't matter!
>Data centers don't matter!
>Locked CPUs don't matter!
>OEMs don't matter!
>Hyperscalers don't matter!
>Upgradeability doesn't matter!
>Anti-competitive business practices don't matter!
>Locked platform features don't matter!
>Synthetic loads don't matter!
>PCI-e lanes don't matter!
>Burnt pins don't matter!
>Heat doesn't matter!
>Server space doesn't matter!
>ECC support doesn't matter!
>*NEW* 7nm doesn't matter!
>*NEW* HEDT doesn't matter!
>*NEW* StoreMI doesn't matter!
>*NEW* Security doesn't matter!
>*NEW* Stock coolers don't matter!
>*NEW* Games don't always matter!
>*NEW* Hyperthreading doesn't matter!

>> No.67088527

5nm is as memey as 10nm, literally everyone perceives upcoming huge problems with it aside from just IBM who already uses it. AMD, highly likely, will be going 4nm instead of 5nm. However, there are 99.99% 3nm and lower won't exist at all as feasible option, since 4nm is currently being touted as the absolute wall. New materials/technologies needed. It's very understandable if everyone stops at 4nm and then moves to new materials/tech...or great crisis of stagnation happens, literal Great Depression in computer hardware means.

>> No.67088646

so intel clearly fucked up with 10nm but their 7nm(aka 5nm) process is according to all sources doing just fine and on track with fab 42 since early 2017. when you take this and their absurd r&d spending relative to amd aren't they very likely to just jump back ahead to the massive advantage they had before in ~2 years?

>> No.67088678

>(((heir))) """""7nm"""""" (aka 5nm)
See >>67088527

>> No.67088780

>when you take this and their absurd r&d spending relative to amd
AMD are fabless and don't have to spend money creating new nodes, which is why they've jumped ahead of Intel. TSMC and GlobalFoundries are doing all the work for them. Sure, AMD still have to pay for usage, but they have options and no risk, whilst Intel are tethered to their own failed node and completely unable to move forward at this point.

>likely to just jump back ahead to the massive advantage they had before in ~2 years?
Ha, good one. Their own roadmap says that 10nm won't be ready until 2020, let alone anything beyond that. Where are these "all sources" claiming that Intel will be jumping to 7nm in that timeframe? Why are they still clinging to the hope of getting their 10nm yields where they need to be if their 7nm process is almost ready? The answer is because that's complete bullshit.

Maybe five years down the line Intel will have clawed themselves out of the hole they've dug for themselves, but the competition isn't going to stand still and wait for them to do it. Everybody else will be moving to smaller nodes too, and almost certainly a long time before Intel does at this point.

>> No.67088838

How does this make sense?

>> No.67088944

I'd say it's more like that jumping to the 7nm node would mean that all the money went into 10nm is down the drain, along with a good chunk of their reputation. The jew might sell their soul, but they don't throw billions out of the window.

>> No.67089018

Jim Keller is going to do his magic thing again, but it's still going to take years. And even then, Intel's fabs are still fucked. I don't really see a chance for them until they spin their fabs off.

>> No.67089034
File: 91 KB, 1208x680, 21825d1529099842-slide5[1].jpg [View same] [iqdb] [saucenao] [google] [report]

desu ironically it seems for once intel is the one being the least jewish about their nanomemes and believing intels fuckup with this particular dye shrinkage to 10nm will completely erase the advantage they built up over the past 10+ years with 100 gazillion shekels in r&d is just amd redditors wishful thinking. amd having a competitive product for once is making you lose all sense of perspective, their latest chips are still not not even beating the supposedly ancient 14nm intel 10th or w.e iteration when it comes to IPC or gaming or high end products it's still muh price/performance and muh cores as always, correct me if im wrong. maybe they can improve their ~1.5% server market share somewhat to make up for crypto dying

the 7nm and 10nm processes at intel are completely seperate projects, they clearly still consider 10nm salvagable. also you don't seem to understand that intel 10nm=amd 7nm and intel 7nm=amd 5nm + you do realize amd is currently still on 14nm/12nm just like intel right? they haven't actually released anything more just yet . let's see if amd actually does beat intel to market with their 7nm(=10nm intel) zen3 cpu that is actually superior and for once delivers on their wild claims

>> No.67089094


>> No.67089102

>intel 10nm=amd 7nm and intel 7nm=amd 5nm
Lo and behold, ladies and gentlemen. Very clear example, the very epitome of what typical Intbecile consists of.

>> No.67089111

>7nm Zen 3 cpu
7nm is Zen 2, you dumb fuck.

>> No.67089205

>you do realize amd is currently still on 14nm/12nm just like intel right?
You do realize AMD is sending out 7nm EPYCs to select companies between December of THIS year and January of next year right? They aren't staying at 14nm for much longer.

>> No.67089211

TSMC and Samsung have more R&D money than 10 Intel companies combined

>> No.67089224

I think spinning off their fabs is the best move AMD's previous execs made and it actually worked out great for both AMD and GloFo too.

>> No.67089283
File: 162 KB, 633x900, tfw_intel.png [View same] [iqdb] [saucenao] [google] [report]



>> No.67089285

they don't have leapfrogging teams, if their 10nm got delayed, it means that everything after that gets delayed as well

>> No.67089435

The best part is that Ice Lake design is actually ready, they just can't fab it yet.

>> No.67089479

Did you mean to say ironically? Would make more sense if you'd said unironically.

>> No.67089604

Ryzen's 12nm is far bigger than Intels 14nm++, despite this, Intel cant even compete in anything except gaming. Once Globalfoundries and TSMC match Intels new (((10nm))) node with their 7nm node, Intel will lose the single advantage they had left.

>> No.67089758

Epyc 7nm is literally being sampled in 2H 2018

>> No.67089776 [DELETED] 
File: 54 KB, 679x758, 1527629778452.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67089783

Zen 2 will be on AM4+; current motherbords not compatibile



>> No.67089802

>no source whatsoever

>> No.67089805


>> No.67089829 [DELETED] 


>> No.67089836
File: 1.05 MB, 1196x676, 1528222249083.png [View same] [iqdb] [saucenao] [google] [report]

>intel shills spreading FUD again

>> No.67089852

>AMD is on "14nm/12nm"
Bullshit, if they were to use Intel's measurements, the "14nm/12nm" node AMD uses is actually 20nm. "7nm" is technically 16nm or 11nm. Intel remains ahead with their true 14nm node. Even their gutted 10nm node will still have higher density than TSMC's "7nm"

>> No.67089866

Makes me feel like the scale will finally balance for once. Competition ftw!

>> No.67089882

I look forward to seeing. It'll be funny.

>> No.67089893

>"7nm" is technically 16nm or 11nm
If you look at the actual measures, it's very close to intel's "10nm"
>true 14nm node
No such thing. It's a name made up by the marketing team, nothing else.

>> No.67089898

icd if it's 100nm, as long as it's faster...

>> No.67089911
File: 69 KB, 685x599, 7nm Comparisons.jpg [View same] [iqdb] [saucenao] [google] [report]

>being this delusional

>> No.67089932

>Intel shill tactic is to call AMD posts shill posts

Like clockwork

>> No.67089939

word on the street is intel will bring their process naming in line with the foundries and "skip 7nm altogether" aka rebranding their 7nm to 5nm, obviously for marketing reasons and the best is the other chipmakers won't be able to say shit about it because it matches their definitions for density and they'd only call attention to the fact that they frauded for years "matching" intels processes when they were playing catch-up

>> No.67089941

Intel's blue is literal eyecancer.

>> No.67090006

kek intlel shill is so anally devastated his piece of shit of a company is dying

>> No.67090028
File: 316 KB, 960x540, patton_7lp_fx-7_briefing_final_4.png [View same] [iqdb] [saucenao] [google] [report]

nice, more power means even higher clocks than before

>> No.67090039

>Intel is literal eyecancer.

>> No.67090070

>Intel's blue is literal eyecancer.
>Intel is literal eyecancer.
AMD stands for Age-related Macular Degeneration

>> No.67090076


>> No.67090086

So AMD is warning us about Intel?

>> No.67090090


>> No.67090101


>> No.67090329
File: 21 KB, 474x317, buyingamd.jpg [View same] [iqdb] [saucenao] [google] [report]

>honey, junior's new computer system did it again!

>> No.67090386
File: 162 KB, 2523x1241, p95-3.6ghz-1.15vid-locked.png [View same] [iqdb] [saucenao] [google] [report]

>says the intel shill

>> No.67090479


>> No.67090480


>> No.67090495

ice lake isn't a new arch, it's the same shit as Core but with some changes

>> No.67090498

amd big shit no buy
intel very good i buy and family
amd has bots to review
i longimte intel fan
intel foreva <3

>> No.67090499
File: 58 KB, 550x550, 1503401098944.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67090517

>AMD Shill mad at being called out so call others shill to muddy the waters on who's shilling what

You are a cancer.

>> No.67090548


>A bit further out, Intel plans to deliver a 10nm Xeon Scalable processor, dubbed Ice Lake, in 2020. No meaningful details about the chip were discussed, but there's apparently a new microarchitecture.


>> No.67090553

>intel shill STILL trying to pretend he's not a shill

>> No.67090574

>AMD shill calling others shill

What am I shilling here, faggot?

>> No.67090592
File: 21 KB, 221x250, 1527533367941.png [View same] [iqdb] [saucenao] [google] [report]

>I don't even know what I'm shilling anymore

>> No.67090613

As if that matters. AMD has never had anything to match Intels high end CPUs. There's like 10 AMD CPUs in the top 100, none of them in the top 10. It's all Intel. AMD is still doing fine but don't pretend like there's actual competition. Maybe in the low to mid range CPUs.

>> No.67090635

At least you admit what you are.

>> No.67090645

What fantasy world do you live in? Intel is barely even competitive in server market and once Threadripper2 is out, they won't have anything even remotely competitive in the HEDT market.

>> No.67090671



>> No.67090677 [DELETED] 
File: 39 KB, 600x600, 1487455639310.jpg [View same] [iqdb] [saucenao] [google] [report]

>AMD has never had anything to match Intels high end CPUs
I wonder (((who))) is behind this post


>> No.67090693

by that metric POWER9 and SPARC are actually releven't and you damn well know they fucking arent

>> No.67090716


>> No.67090722

In all seriousness , why couldn't Intlel at least spin a 12nm node serving as a riskless venture of the upcoming 10nm back around in 2016+ time frame? They would have stayed relevant at least in name, and ought to be easier to make compared to 10nm obviously. Otherwise this will mean that their 14nm node will run for another year for a total of fucking 5 years.

5 years, no progress. 2 years on literally no optimisations save for extra cores, which is a double edged sword causing house fires when OCing cuz they don't have a scalable multicore node like AMD does.

>> No.67090751

>28 cores at 2 ghz is somehow faster than 32 at 5 ghz
can you make it less obvious you are shilling

>> No.67090762

They ARE fucking relevant because they aren't fucking built to be massive paperweights that collect dust to play your pre-release fortnight in your off hours. They are processing immense amounts of data.

Especially when talking about Supercomputers clustering several thousand nodes. It's AMD who is largely irrelevant.

>> No.67090790
File: 604 KB, 800x523, 1488933859408.png [View same] [iqdb] [saucenao] [google] [report]

Kiketel shills who support a company who doesn't give a shit about fucking people in the ass as much as possible with locked, cripped shit and endless security flaws actually exist in this universe. Think about that for a second.

>> No.67090793
File: 32 KB, 496x499, mmlol.jpg [View same] [iqdb] [saucenao] [google] [report]

>Mobile poster

Checks out

>> No.67090810

intlel shills are clinically retarded, more at 11

>> No.67090823
File: 34 KB, 711x534, aHR0cDovL21lZGlhLmJlc3RvZm1pY3JvLmNvbS9YL0MvNzg5MTY4L29yaWdpbmFsL0ltYWdlMS5wbmc=.png [View same] [iqdb] [saucenao] [google] [report]


*Breathes in*

>> No.67090937
File: 32 KB, 650x354, UmZoXqsjv84tJyZgwpPtBF-650-80.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67090988

Is it Amdahl's law that AMD doesn't get much lead over Intel even with a 4 core advantage, or it is still the IPC advantage from Intel , or both?

Regardless, being better than your opponent at the same thin at a cheaper price in making AMD look good rn

>> No.67090989

you do know amd is literally using the node specs of ibm eh? thats why their smt implementation is literally the same allowing amd if they so with to do a consumer grade smt 2and even 4 depending on the l3 cache

>> No.67091018

yeah, no.

>> No.67091082

it's IPC and the fact that 4 cores don't make all that big of a difference when you're looking at cpus with 28 or 32 cores

>> No.67091099

This, that anon it's pulling bullshit out of his ass
If they had leapfrogging teams they wouldn't be in such horrible scenario, they would had scaled back their 10nm effort years ago and pushed an even more aggressive 7nm node,

>> No.67091149

>Intel's slide at their analyst conference confirmed 10nm won't land until 2020
>guise look at this slide from third party analysts, by 2020 Intel will be on 7nm!
How retarded are you?

>> No.67091169

Nice shilling, every other 7nm node beats Intel's 10nm, and that was before they gutted it

>> No.67091231

Upper management it's absolutely fucked, way worse than AMD was back on 2009
The entire board it's retarded

>> No.67091315
File: 93 KB, 1024x645, 1532772117016.jpg [View same] [iqdb] [saucenao] [google] [report]

>> No.67091672

Intel shills are getting desperate that they most irrelevant information to counter AMD's Ryzen 2. Tsk tsk

>> No.67091713
File: 213 KB, 1705x960, CT1YTkM.jpg [View same] [iqdb] [saucenao] [google] [report]

intel shills cant even lie anymore without getting caught

>> No.67091751

They don't have a 7nm process, give me whatever you're smoking.

>> No.67091797

>spreading FUD on plebbit then posting on 4chan
Go back.

>> No.67091988

What's even worse is that he actually browses that site.

>> No.67092005 [DELETED] 
File: 96 KB, 638x710, 1500260377912.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67092146
File: 96 KB, 638x710, 1500260377912.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67092304

Daily Reminder:
Intel 10 nm is the Hillary Clinton of semiconductor fabrication process.

>> No.67092365

7nm your weenie

>> No.67092547

you forgot
IPC doesnt matter!

>> No.67092587
File: 55 KB, 1260x892, intel_tech_lead.png [View same] [iqdb] [saucenao] [google] [report]


>> No.67092593


>> No.67092636


>> No.67092661

>similar to intel 14nm
Fucking lying faggot kikes.

>> No.67092722

If Intel had anything in the works close to that it would be disclosed for the sake of shareholders. Instead we just see 10nm failures and Intel quotes about trying not to give up more and more market share to AMDs lineup.

I get the old AMD hate but the people shitting on it now clearly don't understand what Keller did. Probably get news from the same analysts saying to sort AMD earlier around 11 or 12$

>> No.67092748

this, people don't understand that companies of this size don't hold stuff back if they're in a difficult position because shareholders will jump out

>> No.67092880

where's intel sekrit sauce

>> No.67092925

the interesting thing is that they had this team only working on 10nm, and it didn't work out as expected, so they gutted it back to 12nm-ish
but they don't have anything else down the line ready, so they'll either have to try and fix 10nm for a bunch of years, or keep releasing 12nm stuff until their 7nm is ready for mass production, which would also take multiple years

>> No.67092950
File: 132 KB, 1420x892, fixd.png [View same] [iqdb] [saucenao] [google] [report]

here's the updated version

>> No.67092976

Meanwhile I'm using a PC from 2009 with a 32nm node with the same per-clock performance as a Ryzen 1700X, twelve cores and 3.33Ghz (soon 4Ghz with OC hopefully).

BOTH Intel and AMD are pathetic at developing things.

>> No.67092987

now change the intel's 10nm to 12nm and add 7nm to the competitors

>> No.67093003

>AMD will be moving to 7nm in less than twelve months.


>> No.67093004

and what cpu is that?

>> No.67093015

>i've been in a cave for 10 years

>> No.67093048
File: 28 KB, 678x381, COMPUTEX 2018 - Lisa and David_24_678x452.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67093051

>Meanwhile I'm using a PC from 2009 with a 32nm node with the same per-clock performance as a Ryzen 1700X, twelve cores and 3.33Ghz (soon 4Ghz with OC hopefully).
No you aren't. Even Sandy Bridge has about 30% lower IPC than Ryzen.

>> No.67093069


>> No.67093129

They'll just copy zen architecture and add a wider pipe.

>> No.67093135


>> No.67093155

Indeed. We're seeing a repeat of Athlon64. Intel is Apollo Creed, AMD is Rocky. Spoiler: Creed dies.

>> No.67093198

will tejas and jayhawk finally get to see daylight at last?

>> No.67093249


>> No.67093361

This winter is less than twelve months away

>> No.67093397

anyone have the pic of the fella talking about COAG and the changing over to tungsten by intel

>> No.67093413
File: 399 KB, 1400x1608, 10nm fuckup.png [View same] [iqdb] [saucenao] [google] [report]

right here anon

>> No.67093546
File: 973 KB, 801x1500, 1476643449642.png [View same] [iqdb] [saucenao] [google] [report]

This is Pentium 4 all over again.

>> No.67093575
File: 57 KB, 472x411, liberal2017.jpg [View same] [iqdb] [saucenao] [google] [report]

AMD has always been the lesser performing company with worse processors and price/value points. It's now trying to swindle people into cheap 7nm meme while Intel is offering 10nm for the same money.

>> No.67093599

might be that x58 dicksucker that always shitposts about his 200w chips

>> No.67093692

I hope that 4nm is the wall and alternatives take at least a good while to develop.
Futurist and transhumanist fags scare me. If we aren't careful we'll end up in a dystopian future ruled by technology.

>> No.67093700

they didn't really have a choice. no money after getting shafted by intel

>> No.67093788


Dual X5680.

Tell that to benchmarking software

What's 200W when I'm burning 800W mining shitcoins to heat my flat? Also, not that guy.

>> No.67093814
File: 53 KB, 674x800, IPC.png [View same] [iqdb] [saucenao] [google] [report]

alright then, benchmark it like pic related and let's see the results

>> No.67093919

Not sure how you missed it, but this was pretty much /g/ reaction for the last 2 years. And Intel just keeps on giving.
In fact I hope that someone will update this meme again and include 5GHz "Re-entrancelake " CPU

>> No.67093923

>less dense than competition
Here's how processor lithography/fabrication happens.

Intel has their own fabs, they don't make their own fab machines.
One or two companies in the entire world make all the machinery for leading edge/bleeding edge litho.
All of the top fabs are using the same machines, capable of the same standard.
These Fab companies have their own design rules and processes.
Yields are essentially [Litho machine capability] times [complexity/density of chosen design rules],
Put another way a machine does 1x and the chip requirements take a slice of that 1x for each step in the process
Building a more dense circuit requires more precision and more steps
Thus yield is lower for more dense processes
Intel isn't yielding for their "10nm" process because they are and have always been the most aggressive with chip density (ignoring their monolithic die approach)

Simply put, Intel's "10nm" is closer to 8.5nm Gate Half-Pitch and GloFo/TSMC/Samsung's "7nm" is closer to 9.5nm
>less dense than
No, fool, it's much more dense and that's why they can't meet their target without a fully functioning EUV machine

>> No.67093936
File: 2.29 MB, 5098x1500, EBYN reaction.png [View same] [iqdb] [saucenao] [google] [report]


>> No.67093977

see >>67089911

>> No.67094107

>look at these unconfirmed numbers on paper some guy put together using nothing but imagination
Talk to me when all of these companies have real product for us to measure.

>> No.67094148

see you in 2020 then

>> No.67094463

>MCM is pretty good what do we do about it
>overclock our 28-core platinum sample to 5GHz on all cores

>> No.67094508

10nm will never be a real product, it's stillborn.

>> No.67094547

just remember that those were the old values that they speculated, before intel gutting it back to 12nm

>> No.67095079

I literally just come on here everyday just to hype myself up for 7nm and for people to tell me release dates

>> No.67095267
File: 54 KB, 679x758, 1527629778452.jpg [View same] [iqdb] [saucenao] [google] [report]

delid dis

>> No.67095299

Intlel processors have 0 value lul

>> No.67095326

I agree with your sentiment,
My long post was arguing strictly for Intel's density goals at 10nm compared to other foundry's 7nm

>> No.67095390

Quite fucking ironic considering that Threadripper 2 2990WX was revealed in CES by AMD, and intel put a server CPU with a chiller that performed worse in synthetics.

>> No.67095589

Oh look, he went quiet.

>> No.67095626

Honestly, I think they'll become a fabless company post 12/10nm and just use TSMC fabs. Really can't see a way forward otherwise.

>> No.67095641


>> No.67095668

but TSMC doesn't build stuff for clocks, they're specialized in efficiency, since they build gpus and mobile shit

>> No.67095852

>If we aren't careful we'll end up in a dystopian future ruled by technology.
>implying this hasn't happened already

>> No.67095853

Why do people act like Spectre only affects Intel?

>> No.67096113

Because Intel seems to have the hardest time mitigating it

>> No.67097546

[citation required]

Also, the 7nm will have very similar density to Intel's 10nm.

>> No.67097778

Spending your way to the top is still a valid strategy, and intel has the money for it, they just need the management back in place.
Maybe in 5 years or so intel will be a ruthless anti competitive dominating force again.
They'll probably keep their stranglehold on the laptop market even if AMD's APUs are better. Intel gives huge discounts to OEMs through supplier partnerships. Build a laptop with an intel CPU and you get a discounted display panel, cheap SSD, etc. That increases profit margins for the OEMs selling the systems, and thats why AMD laptops aren't more popular.

>> No.67098706

>10 years from now AMD will be Big Brother and Jewtel will be the underdog

>> No.67098789

Nah, I doubt that. I've met a bunch of AMD fellows and upper management back at the fusion developers summit. They're honest people.
Meet any intel execs at a booth anywhere and you get fake smiles, canned responses, and sanitized PR behavior. AMD is a company of real people, still mostly American.
I sincerely doubt they'd ever stoop to the practices that intel has in the past. At least not with current management.

>> No.67098797

Raw transistor density numbers don't matter as much as you think. Intel's fab division is known for pushing restrictive process design kits on their design teams. All those density advantages on paper disappear once you have to layout the design.
It's part of the reason why Intel's foundry business never really took off (besides the fact that they half-ass every entry into a new field), real-world designs on ostensibly-comparable nodes tend to be less dense than those on competitors like TSMC and Samsung.

>> No.67100199

Considering how much money TSMC and Samsung have Intel's fabs are a fucking joke compared to theirs.

>> No.67100210

>I sincerely doubt they'd ever stoop to the practices that intel has in the past. At least not with current management.
They never did and never will, Intel makes AMD look like saints in comparison.

>> No.67100220

Chipsets and architectures take years to develop

Intel's real answer won't be felt until 2-3 years from now

Right now they're just cannibalizing their Xeon line because they don't have anything real on the pipeline and this incoming 8-core chip is their real response to Gen 1 Ryzen but they panicked and rushed the 6-core chip just so they can sell a few more CPUs

>> No.67100227

>AMD be gud
(lol AMDrones

>> No.67100229

Truth triggers the shekelstein goldberg

>> No.67100238

AMD = real human beans
Intel = jewish corrupt reptilians

>> No.67100262

Someone photoshop reptilian intel jews mind controlling the filthy goyim.

>> No.67100270

AMD shills actually believe that AMD won't do the same shit Intel and Nvidia are doing right now if put on the same position.

They're shills and therefore their opinions are worth nothing.

>> No.67100291
File: 245 KB, 800x612, 1489160516428.png [View same] [iqdb] [saucenao] [google] [report]

Nice crystal globe, any more jewish reptilian magic tricks?

>> No.67100292


Why do people act like meltdown is spectre?

>> No.67100382

Intel will just market and lie like usual:

>new Intel(r) THICCTECC(tm) process give you DOUBLE the NM(tm) of competitors using proven manufacturing techniques
>THICCTECC(tm) delivers value providing you with up to double the silicon using 14nm(tm) while competitors are delivering 10nm or even now cutting down to 7nm
>with new Core(tm) i11 Intel(r) CPU you get the latest innovations in performance, including the all new SPEC-DOOR(tm) tuning tool which enables you to selectively disable redundant security checks and providing up to 30% boost in performance! (requires SPEC-DOOR subscription*)

>> No.67100546


Maybe they will, maybe they won't. But intel certainly does and they deserve to die.

>> No.67100720


>> No.67100726
File: 614 KB, 992x1043, 1532984775087.png [View same] [iqdb] [saucenao] [google] [report]


>> No.67100743

Well seeing as AMD fags are poor, they wont be able to afford 7nm chips and have to convert to Inyel fags.

>> No.67100769
File: 6 KB, 415x416, 1503253159417.png [View same] [iqdb] [saucenao] [google] [report]

Doesn't matter

There's currently a severe imbalance in favor of Intel. Blame gabidalism for not offering more competition

>> No.67100811

Are you sure? I thought all Ryzen 2xxx SKUs (including the 'first gen' APUs) are Zen+.

>> No.67101199
File: 1.03 MB, 268x274, 1480364714158.gif [View same] [iqdb] [saucenao] [google] [report]

>Two more years before Intel's 10nm.
Could have sworn I heard that 4 years ago

>> No.67101263

they will eventually, its their fiduciary responsibility to keep increasing shareholder value

>> No.67101283

>shilling for intel on your mobile phone
so they dont even pay you enough to buy a crappy laptop?
HAHAHAHA the current state of intel corporation shill team

>> No.67101382



AMD naysayers BTFO

>> No.67101407
File: 59 KB, 906x847, cwifd34ovbf11.jpg [View same] [iqdb] [saucenao] [google] [report]

Deludded Intlel shills

>> No.67101422

Who gives a fuck? Where are my cheap 4 core laptops with 64gb of ram?

>> No.67101436

Ask Intel because they are the one's who control most of the laptop market with overpriced junk through constant bribes and arm twisting.

>> No.67101470

>64gb of ram
>on a laptop
just why?

>> No.67101535

To run chrome

>> No.67101556

Looks like the jews finally gave up on trying to short AMD stocks

>> No.67101656

The majority of online traffic comes from mobile devices nowadays. Who actually uses a PC/laptop to casually browse the internet? How old are you? Kek

>> No.67101668

Intel trying to excuse themselves and talk up 10nm like it's better than AMD's 7nm.


>> No.67101737
File: 126 KB, 1024x721, 1528671260361.jpg [View same] [iqdb] [saucenao] [google] [report]

>ju-just wait anons, Intel will win in 2020, India Super Power

>> No.67101849

it's because you did, we were promised 10nm years ago, but the delays they had with 14nm made 10nm and 7nm be delayed as well, and now the 10nm delays will make 7nm and beyond be delayed too

>> No.67101864

did they tell goymen nexus that they had to gut that piece of crap so now it's basically 12nm and that it was only comparable to AMD's 7nm before that?

>> No.67101903
File: 255 KB, 719x720, 1530661076450.png [View same] [iqdb] [saucenao] [google] [report]

Delid right now

>> No.67101960
File: 130 KB, 1054x850, IMG0049517_1.png [View same] [iqdb] [saucenao] [google] [report]

>and my dad works for Nintendo
Intel bet on fully functional EUV arriving 2015/2016 and they have been delaying since Skylake. 14+++ is fin profile adjustments and 'hyperscaling' is their failing attempts at SAQP with 193i

This report from 2013 suggested that the lithography hardware wouldn't exist until 2016 and chips wouldn't be in hands until 2018. For everyone.
TSMC is the only company with shown 7nm product, the 32GB Vega64.

>> No.67102525

>For everyone.
>TSMC is the only company with shown 7nm product

>> No.67102675

lol you dumb fuck, intel is thinking of scrapping its fabs and just ordering cpus of TSMC, thats how fucked they are...

>> No.67102852

TSMC doesn't have the capacity to support Intel and I doubt that they will think it is worthwhile to build new plants just for Intel. There is also all that patent shit. Intel will bleed money if they can't un-JUST their fabs.

>> No.67102878

intel has literally no way out of this, both tsmc and samsung are shitting on intel's fabs, they have a fuck ton more money to spend on R&D

>> No.67102918

Intel is the best cpu maker out there. They have the highest quality chips who cares if theres 7nm chips the general public has no idea about this crap anyways.

>> No.67102928

They're also poorfags compared to Samsung and TSMC so there's that.

>> No.67103732
File: 54 KB, 711x457, 1533705388992.jpg [View same] [iqdb] [saucenao] [google] [report]

This one's better:

>> No.67103811



>> No.67103852

Leadership performance when it comes to making security holes

>> No.67103875

>smaller nodes don't matter! 14nm forever!

>> No.67103918

Holy shit, what a delusional fuck this David Kanter is. Disgustingly mistaken retard. It's very clear he has a big chunk in Intel's stock bought, so he's squirming by coming up with these bullshit excuses to minimize the damage that's Intel currently suffering. Fucking cuck.

>> No.67103940

>David Cuckter

>> No.67103992

>it's better goy we will never lie to you and bribe oems trust us

>> No.67104632

Corporations already own you, relax and let them install the botnet into your body. You've got nothing to hide anyways, right?

>> No.67105251
File: 2.56 MB, 3840x2160, DSC_0030.jpg [View same] [iqdb] [saucenao] [google] [report]

It makes me worry if I should buy the i9 if my CPU replacement still won't work or buy some old Intel or cheap Ryzen 1200 until Zen 2.

This is my 2700X build. it has no CPU and if I don't get a working replacement I will demand my money back.

>> No.67105404

I think you should get the 1200 and wait
absolutely worst case scenario the 3700X will be 5% slower than the 9900k

>> No.67105481

i doubt so, the 9900K will be 20% faster (singlecore) than the 2700x
the 3700x will also feature more cores
im not sure if i should keep the crosshair in that case, i dont know where the fault is

>> No.67105523

What case is this?

>> No.67105542

just look at it like that, if you spend 100 bucks on the 1200 right now, you can wait until next year and see what's up with zen2
if it doesn't look that great, it'll be cheaper to move to the 9900k compared to buying it on the launch date + z390 board

>> No.67105604

Corsair Crystal 570X
I own pic related, 2700X and Crosshair bought on release for almost 600 €
It just bluescreens and triple boots
5 RMAs now the 6th one where I RMA the CPU
I could get a 1200 and cheap B350 board for 100 €, I think this would even suffice for 4K60 gayming
I have 16 GB 4266 RAM

>> No.67105646

holy shit anon, I hope this CPU RMA works out for you

>> No.67105669

Less dense, but it'll probably clock higher.

AMD was pushing GloFo toward making their process more like TSMC's 7nm smartphone process as their server and notebook clients are more interested in the higher perf/watt than the higher raw per core performance.

I have a feeling that *if* we see 5GHz base clock Ryzen CPUs, they'll be expensive "Black Edition" on GloFo's IBM process, and that's *only* if server customers are also interested in a ~5Ghz+ part at a price premium.
But this is pretty baseless rumor. We still don't know what actual process Zen2 is on yet, afaik.

That sucks. I've had similar situations in the past.
It's difficult when two things are busted to figure out what's going on and it throws you for a loop.

>> No.67105718

I wonder if they're going with TSMC for parts that need more efficiency and/or bigger dies while using GloFo for the stuff that needs clocks
If they were smart about it, they could realistically design two different dies, one better suited for server stuff, with more cores or ccx's or whatever, and a more consumer oriented one, not as focused on core count, but on latency and clocks

>> No.67105724

>intel fags rule for years that that their cpu is the best amd is finished for years
> amd fags say ryzen soon trademark
>amd succeeds
>intel is shown to make insecure shit that reduces performance

>> No.67105790


Delete this thread! Intel will always be the best I will not stand for this!

>> No.67106016

Definitely not. They have a contract with TSMC so they have to pay a fee to GloFo every time they fab chips at TSMC.

So for mobile, they wouldn't be cost effective having TSMC fab them. AMD wants GloFo fabbing their mobile parts on a similar process to TSMC.

>> No.67106574
File: 1.93 MB, 3840x2160, DSC_0014.jpg [View same] [iqdb] [saucenao] [google] [report]

Yes I hope that too, the performance is good. >>67105669
I tested all, even other RAM, it can only be board or CPU. CPU defects are very unlikely. It still looks like brand new because it is, I could barely use it.

I now bought another 3 ML fans to do this so I was motivated enough to put the whole PC together again. It took me almost 3 hours. Now its missing only a CPU.

>> No.67106611

so it'd make sense to say that AMD is only going to use TSMC for the stuff that will have higher profit margins, so paying the fee wouldn't hurt them as much?
this would leave TSMC with the enterprise products and GloFo with the consumer ones

>> No.67106684

adding to this:
this only enforces the claim that AMD built two different dies, because if they were the same, it'd mean that TSMC would be fabbing both enterprise and consumer stuff

>> No.67106882

I wonder how tiny would the die size and power consumption of a pentium 4 be it it was built using one of those nodes

>> No.67106920

A single-core one? Probably ridiculous. Modern Intel uarch is basically a P4 with a hefty frontend.

>> No.67107029

Sorry Mr Pot, but my kettle isn't black.

>> No.67107156

> How does this make you feel?
like, «twelve months? 7nm? holy shit i feel old»

>> No.67107447
File: 64 KB, 691x771, 1515145992968.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67107549

I assume you tested with another PSU, different RAM, different GPU, different hard disk/SSD and BIOS? If you did and it still BSOD's then it is either mobo or CPU. Have you spoken to AMD the motherboard manufacturer? Maybe ask AMD for a loan CPU to test?

>> No.67107671

yeah just like am3 didn't cover am3+ !
oh wait I can use an 8350 bulldozer on an old ass am3 board just fine.

>> No.67107672


>muh MHz

Cucktel marketing.

>> No.67107722

>shilled by everyone
>every (((tech reviewer))) praises it
>their blame moore's law when Intel doesn't deliver (instead of blaming management)
>still will not come in 2019

>> No.67107742

No...no no no no no!

>> No.67107775

intel already perfected their 10nm process
you can get the greatest chip that exists today!
buy it guys, there's nothing better than intel!

>> No.67107819

come on anon i'm a idiot but i'm not a brainlet

>> No.67107829

Only certain AM3 boards though. A lot were never updated with FX series support, including the Gigabyte GA-MA790XT-UD4P that I own.

>> No.67107830

>Subject to future roadmap modifications

>> No.67107861

alright then, give one good reason for them to introduce an AM4+ socket as opposed to what they promised

>> No.67107866

And? Guess what we're in 2H 2018 right now.

>> No.67107940

Does it even matter? Literally everyone bought shitty X/B mobos with fake VRMs, if they want anything more powerful than what they have they'll have to upgrade anyway.

>> No.67107991

>fake VRMs
Tell that to the Asus Strix (X chipset), Asus Crosshair 6/7, AsRock Taichi/Fatal1ty equivalent, and Gigavyte Aorus 7.

They can deliver way over 250A of current, The Crosshair 7 can deliver 500A. A 12/16 core Ryzen on the 7nm lithography probably won't go over 250W.

Only MSI's high end boards have shit VRMs, although they have the best B chipset VRMs on boards like the tomahawk.

>> No.67108022

That's not AMDs fault either. That's up to you as a consumer to make choices. I have a 1600 on a high end B board that can handle the 8 cores just fine.

>> No.67108039

Anyone with half a brain bought a Crosshair VI, knowing that it would be the base of their system for years to come on what was always going to be an extremely long-lived platform. The brainlets who paid almost as much for a Taichi can enjoy their shit tier BIOS support.

>> No.67108059

>Tell that to the Asus Strix (X chipset), Asus Crosshair 6/7, AsRock Taichi/Fatal1ty equivalent, and Gigavyte Aorus 7.
I'm obviously not talking about those, which only a minority bought due to "hurr AMD must be cheaper or I go Intel" meme. I'm talking about low-end Asus and Gigabyte B boards (and even some X boards, specially from MSI).

>> No.67108070

The Taichi is still the best for right now.

>> No.67108154
File: 203 KB, 1510x1006, 2018-08-12_lMKOeGlzgPbEzHDB.jpg [View same] [iqdb] [saucenao] [google] [report]


>> No.67108173



>> No.67108223

>Crosshair VI, knowing that it would be the base of their system for years to come on what was always going to be an extremely long-lived platform.

Lol no. The latest BIOS for Crosshair VI has problems whereby the fan stops spinning midway for no reason (resulting in thermal throtting/safety shutdowns/BSODs) and it has not been fixed for more than 3 months. Not everyone is affected by the issue, but it is significant nonetheless. Don't expect long term UEFI support on Asus either.

>I'm obviously not talking about those, which only a minority bought due to "hurr AMD must be cheaper or I go Intel" meme. I'm talking about low-end Asus and Gigabyte B boards (and even some X boards, specially from MSI).

Back when I got my Crosshair 6 and Ryzen 7 the combo only allowed me to get a Z270 Strix-F board with the 7700K(before coffee lake was a thing). The non-bundle price of the 1700 and 7700K was similar in my region(and the 1800X non-bundle is the same price as a 7820X, not kidding you), but the motherboard price difference is phenomenal. A Maximus IX Hero costs the same as the Crosshair VI Extreme. If I went for a Maximus IX Code+7700K, I could also go for a Sabertooth X299 MK 2 and a 7820X(I will definitely choose the latter because I play with Virtualization quite a bit).

IIRC the Intel Strix-E/F boards have pleb-tier Vishay MOSFETs too, whereas the AMD Strix-F boards actually use chad-tier 6 IR Powerstages.

So yes, even those who go for the high-end boards also considered that AMD is cheaper, more sustainable, have better support for more power-hungry CPUs and more value in the long run.

>> No.67108230
File: 62 KB, 680x463, 43213123.jpg [View same] [iqdb] [saucenao] [google] [report]

MFW I have AMD shares

>> No.67108375

>when even toms hardware can't defend intlel


>> No.67108410

>the fake VRM board is still the best right now
No it isn't

>> No.67108438

The VRMs are real(better than crosshair 6, worse than 7), but their BIOS is atrocious

>> No.67108549

>16 phase power phase design
My bad, this was the 6+2 phase board on par with a Strix-F. I'd mistaken it for the Master SLI "12 phase design" 4+2 board that's absolutely bottom of the barrel. Wonder when Ascock will stop lying about their VRM?
on a side note, I still think the Asus Prime Pro is still the pick on bang for buck.

>> No.67108637


Damn that threadraper is good, and it's clear that AMD makes a great profit on them and could easily cut the price, where it's likely that intel is making a loss on those i9.

>> No.67108644

The Taichi has actual doublers.

>> No.67108718

AMD is just doing what they did with TR 1, they priced their flagship the same as intel's flagship of the time (1950X and 7900X)
intel is not losing money on their released HEDT lineup because shareholders demand 60% profit margins, and those big monolithic dies are expensive
I don't think the 28c product will really happen any time soon, maybe only in extremely limited quantities, since it's a extremely well binned xeon 8180 and those things are already pretty expensive as they are

>> No.67109132

They don't need to design two different dies.
Both 7nm nodes from GloFo and TSMC are similar.
>In an article at EETimes, GlobalFoundries admits using similar pitches and SRAM cells to TSMC’s 7nm, precisely because it gives AMD more flexibility. The company’s chief technologist, Gary Patton, told EET’s Rick Merritt that AMD “will have more demand than we have capacity, so I have no issues with that.” https://www.extremetech.com/computing/270191-amd-needs-more-7nm-capacity-than-globalfoundries-can-provide
AMD were able to port Vega 20 to TSMC when GloFo couldn't supply what AMD were demanding.
But considering AMD has a exclusive contract with GloFo with their CPU line, I don't think that will happen.

>> No.67109214

but there's an issue when you're making the same die on two foundries, people will start overclocking those things and getting wildly different results, when you buy a product you'll have to be lucky that you got one of the right foundry
this would not be good PR for AMD

>> No.67109267

eluded, have GloFo manufacture consumer CPUs e.g. Ryzen and Threadripper
and have TSMC manufacture EBYN. Therefore there won't be that problem you suggest.
Besides, enterprise parts are rarely overclocked

>> No.67109313 [DELETED] 

what if I told you that those are also me?

>> No.67109465

well then you kinda contradicted yourself in >>67106611

>> No.67109489

but that would still require different dies, the difference between a die that goes on a EPYC 7601 and one that goes in a r3 1200 is how well binned it is
if there's a single die design you're automatically making all the products in the lineup, not only enterprise or consumer ones

>> No.67109667

No you don't
AMD can port one uarch design over 2 nodes without penalties. Apple did this with A9 processor produced at Samsung and TSMC

>> No.67109723

And it probably AMD will do with Vega 20 too; have tsmc dies in enterprise part and GloFo dies in RX parts. Hence we haven’t seen RX Vega 20 yet but Instinct is out

>> No.67110072

that old gem

>> No.67110192

Based AMD

>> No.67110466

The 9700K will be great for gamers and encoders but nothing else.

>> No.67110484

>AMD will be moving to 7nm
It's not going to happen before intel

>> No.67110511

10nm isn't gonna be out by 2020, 7nm AMD is scheduled for early 2019, but hey maybe 14nm++++++ will hold up in the meantime

>> No.67110516

Lol I fucking bet this is what happened

>> No.67110648

EPYC on 7nm is already sampling to trusted OEM's. They are screwed down tighter than a 3yo's asshole on NDA's.

>> No.67110691

Thicker rivers but the rivers are closer together.

Won't matter because Apple and Dell want small nm processors so they don't THROTTLE like Intel and their i9

>> No.67110908

>doesnt check the date
>july 31 2018
>thinks amd will publicly post a roadmap that a week after will throw it away


>> No.67110929


>> No.67111206

>Intel have always been the most aggressive with chip density
And they didn’t learn from the failings of initial 14nm production which was 2.5x density of 22nm, so they decided to aim for 2.7x density for 10nm? They cucked themselves

>> No.67111439
File: 52 KB, 569x347, fni2w2bre.png [View same] [iqdb] [saucenao] [google] [report]


>> No.67111981

yeye 7 nm from company who just rename chip name and deploy it on market

>> No.67111986

whaaaaat? already? I thought they were just downgraded to neutral.

got a link?

>> No.67112898
File: 87 KB, 757x627, 1529567124804.jpg [View same] [iqdb] [saucenao] [google] [report]

Mommy Su and the AMD Team broke them.

>> No.67113697

>implying we weren't doomed the moment electricity was used
>implying we don't already live in a cyberpunk dystopia considering everyone looks at screens which tell them what to think

>> No.67113797

This is single-thread performance, the AMD processors listed come with 4x the amount of threads of most of the listed Intel processors.

is more accurate for real life usage

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