Intel Corporation Building "Revolutionary" New Processor

>>61205364
>here's only so much room for Intel to crank up the frequencies at which its processors operate and the amount of work that these chips can get done per clock cycle hasn't gone up by that much over the last several years.

Yeah thats why this is happening.

>Intel creates itanium 2
>HP goes bankrupt
>AMD released new x86 extension with the same improved capabilities whilst maintaining backwards compatibility with their new EPYC-FU processors
>Intel bolts amd-FU extension on to repeat lake
>intel starts work on core4 architecture.

We're on the good timeline again.

>>61205364
>AMD btfo
What happened the last few times Intel tried to introduce a new architecture to beat competition?

>Itanium lol
>Bentium IV Bibelines :DDD
>x86 in phones

>>61205641
yeah thats what i think will happen too but still research > no research

You're not funny.

Too late really. AMD will easily capitalize on Intel slacking up to 2020 at least in the desktop space.

>>61205641
x86 phones would have succeeded if they included the ability to use full windows 10

File: The pussie destroyer.png (105KB, 255x345px) Image search: [Google]

105KB, 255x345px

>>61205364

>>61205743
Has anyone got an interview with him since the Ryzen launch?

>>61205364
>within the next four to five years

Wait for Intel

>>61205364
who would have thought competition would wake intel the fuck up

they have been shafting their customers for years with marginal improvements at jew prices

>>61205761
He got a private one with Intel

>>61205717
If amd manages to deliver high clock on 7nm Intel is fucked big

>>61205364
im sure when this releases in 5 years itll be midly impressive compared to 7nm zen2

>>61205733
x86 isn't as power efficient as ARM. what windows 10 programs would you even want to use on a small slow phone

>>61205820
Dwarf Fort

>>61205808
She's cute

>>61205846
>she

File: OwO what's this.jpg (17KB, 540x255px) Image search: [Google]

17KB, 540x255px

>>61205808

>>61205864
>hurry durr durr muh she hahaha look raddat im so original

>>61205364
This would've been a great idea .. 4 years ago.

>>61205364
competition is a wonderful thing :^)

>>61205733
that's the thing, they had to remove a lot of stuff out of the actual x86 chips to make them as low energy as mobile required
so much so they snipped out what windows needs

>>61205900
Are you blind or a literal faggot? Virgin maybe?

>>61205932
Virgins have higher IQ, subhuman.

>AyyMD stole 10% market share in only 1 quarter!
>Jewtel B T F O
Meanwhile in reality
https://seekingalpha.com/article/4085523-amd-gain-10-percent-market-share-quarter

>>61205900
there's a bulge in there, anon

File: last shred of hope intel wojak.png (84KB, 653x726px) Image search: [Google]

84KB, 653x726px

>>61206278
>i-it was only 10% of the enthusiast market
>intel craptops still rule, b-btfo

File: 1497297720642.jpg (119KB, 1000x1000px) Image search: [Google]

119KB, 1000x1000px

>>61206278
>https://seekingalpha.com/article/4085523-amd-gain-10-percent-market-share-quarter

damn the damage control is real

>intel gets the contract keller
>top 10 meme bretrayals

>>61206478
He's too busy working for Tesla. Going to put truckers out of work.

>>61205364
>Ashart Outassa

Now, since the goddamn scalpers and miners are out, I want nothing more but a good processor with a reasonable iGPU.
That's all.
I don't even care about HD FULL graphics.
Nobody needs chromatic aberration, nobody needs film grains.
Nobody needs more antialiasing if the resolution is big enough.
Just an iGPU that works good for the next 2-3 years, even 1st year is okay.

>>61206465
You dumb fuck
Seekingalpha is a hard dickrider of AMD
https://seekingalpha.com/symbol/AMD
>Ignore The 'Sell' Call On Advanced Micro Devices
>Why I'm Staying Long AMD Despite Suggestions The Stock Is Relatively Overvalued
>AMD: Vega's Advantage
kek at this one
>Wells Fargo: AMD Epyc offers better performance than Intel Broadwell
>AMD Ryzen Pro Unexpected Launch Indicates Surprises Ahead
>The Real Reason Why AMD Is Going To Surprise In Q2
>Advanced Micro Devices New Super Chip Should Frighten Intel
>Why Advanced Micro Devices, Inc. (AMD) Stock Might be a Great Pick

>>61205820
Muh gayms.

>>61206278
>"all the ryzen chips are actually enthusiast class"
>"and according to these sourceless claim, enthusiast chips are a third of the market"
>"also desktops are a third of the market too, according to this other sourceless claim"
>"our conclusion contradicts this claim of a 10% market gain, obviously that gain is non existant"

>>61206545
They're right on every point, though. In fact, he's saying making excuses for why AMD hasn't been taking away 10% market share "because Ryzen only just came out".

And if you think he's an Intel shill, just read the other articles he did on AMD. He's a clear AMD shill through-and-through.

>>61205364
Is this new chip will be good for MEGAtasking?

Goodbye MMX and SSE.
Goodbye x86.

>>61207308
It's still x86

Unless they're moving from Von Neumann, don't expect much other than some kind of new memory rewamp, maybe a new interface, in any case, Von Neumann's bottleneck is memory access and has been for a while.

>>61205820
Office, chrome, facebook, an actual browser.

>>61207331
It's gonna be a Processor-in-memory architecture

Just 5 years. Just Wait.

>>61205364
>Early 2017
Intel isn't afraid of Ryzen, it's going to coast on it's sales anyway
>Mid 2017
Intel is trying to create Itanium 2, not in response to it's competitors.

>>61205808
What's up with those traps on /g/?...
Besides, in 5 years aren't we going to rock zen 3s already?

>>61205364
I hope they succeed. Thank you based AMD for shaking Jews to the point they actually try again.

Intel will already be bankrupt in 2020
AMD is the new CPU innovator

THANK YOU BASED INTEL

AYYMD IS FINISHED & BANKRUPT

AYYMDPOORFAGS CONFIRMED ON SUICIDE WATCH

>>61207530
>AMD is the new CPU innovator
When weren't they? I mean faildozer was a fucking failure was it was still a departure from ordinary ways.
Also AMD is thousand times better than Intel at pioneering memory types.

File: Sandwich.webm (3MB, 480x270px) Image search: [Google]

3MB, 480x270px

>>61207583
All bulldozer really needed was one FPU per core, and it would have done pretty alright overall. Also better stock coolers naturally.

I still want one of those metal CPU boxes though.

the FX chips weren't that bad at least when comparing multi-threaded performance to an i5. But AMD immediately started work on Zen after it's launch most likely.

>>61205364
On a scale of Pentuim 4 to Itanium how bad is going to be /g/?

>>61207610
It was not only FPU. Bigger high latency caches and shared front-end were a mistakes.
That's why Zen uses ordinary cores and not wacky "modules" of any kind.
Also Interlagos was fucking boring since it was a Magny-Cours with two Zambezi instead of two Istambuls.
Also Intel fell for bigger L2 meme in SKL-X and it backfired in some workloads.

>>61207631
If it's Oregon team? It's Tejas-tier single socket holocaust. If it's da joos - it's another P6 iteration with parts fused off for bigger vectors and moar registers.

>>61207651
I did some quick cinebench testing at one point using the One Core per module BIOs setting and it certainly would have helped.

>>61205784
I'm scared, what if the jews go full panic mode, call Keller and give him unlimited budget for a processor that save their asses from the curry flavored shoa?

>>61207693
No, the joos coerced Compaq into killing Alpha series. Killing the jew is totally personal for him.

File: intel.jpg (22KB, 511x165px) Image search: [Google]

22KB, 511x165px

>>61207610
>sandwich wasn't covered in processors
Disappointed.
>>61207671
pic related

Reminder that Compaq axed Alpha in favour of Itanium.

>>61205820
>x86 isn't as power efficient as ARM
This is pretty much bullshit. There's virtually no overhead to the architecture, it's all down to chip design and software. ARM could take a spot on the desktop from if anyone actually took the time to develop an ARM chip for higher TDPs and the Intel Atom chips could have been successful had they not been hamstrung by a full desktop OS on tablets and what was essentially an emulation layer on phones.

Neither of those things will ever happen but that doesn't mean that ARM is inherently more power efficient or x86 inherently better at absolute performance.

>>61207825
I wonder how low can you scale Zen before performance becomes pitiful.

>>61207958
I meant power envelopes.
Everyone knows Zen and IF shit on everything made before in terms of raw coarz scaling.

File: 1471434695013.jpg (55KB, 849x656px) Image search: [Google]

55KB, 849x656px

>>61205364
>four to five years of the same old shit
It's the era of AMD after all.

>>61207651
Bigger l2 cache was required by big players who will have better performance in database and other enterprise loads.
It's another show how this arch is just xeon rebadged in a poor way.

>>61208003
And with R&D money (a lot of R&D money) AMD can cook a lot of wonderful stuff.
Heck, they pioneered fuckton of shit with their peanuts R&D money.

File: m1q9yw7ugp2z.png (10KB, 583x344px) Image search: [Google]

10KB, 583x344px

>>61208027
Not him, but I just want to drive the point home.

>>61208058
That's, like, hilarious.

>>61208058
R&D in diversity and OY VEY!

>>61208058
$12 birrion dorra worth of R&D money and yet these faggots delivered failure that was Optane?
The fuck.
Just donate this money to AMD they will make working memory for you.

>>61205364
>expected 2021-2022
Way too late. 2019 will truly start the downfall of Intel. They can't keep up with process shrinks.

>>61208122
tfw half of intels R&D is spent on buying AMD products and inflating their stocks to keep themselves from being cucked by monopoly laws

>>61208151
Zen2 +10% ipc + 7nm + 5ghz + 6cores per ccx. Intel is fucked.

>>61208155
Like really, they have all the money in the world, fabs and partnership with Micron and yet they still failed to deliver on Optane. THEY CANT MAKE A FUCKING MEMORY STANDARD WITH $12 GORROLION DORRA R&D budget.

>>61208217
this is exactly what they wanted..
>release shit that's better than most other shit, but still shit
>release slightly improved shit the next year while either keeping the price stupidly high, or raise it
literally intel 101..
just look at how the release rebranded cpus are different models albeit being exactly the same..

>>61208160
Add on 7nm Navi for godly APUs. We're gonna have GTX 1050 performance at the least on a single chip.

>>61205468
>Intel creates itanium 2
>AMD reveals K12 wasn't killed off, still being worked on and the age of ARM has come as Windows has 99% efficient translation layer for x86 code
>Intel dies

>>61207825
That's not entirely true. It's known that x86 isn't a perfectly optimizable ISA whereas ARM is, so it's automatically less efficient by design.

>>61208403
Optane is shit. Also promised Optane-based NVDIMMs (Apache Pass) disappeared somewhere.
>>61208541
Or RISC-V. Anything.

>>61208151
>They can't keep up with process shrinks.

Wut?

Everybody uses the exact same machines to make chips (build by ASML).
Sure there are some differences in processes, but they can all achieve similar resolutions.

>>61208585
He means the process shrinks alone won't save Intel.

>>61208585
You can't just die shrink an existing process and hope it works perfectly. Intel has already fucked up 10nm either ways.

>>61207313
Nothing x86 could be called "revolutionary" at this point.

Isn't AMD,IBM , Samsung and Global Foundries working together on 5nm process on graphene ?

>>61209452
Since what, 2010? There has been no updated news on that.

>>61205364
Finally, a gay processor

>>61205364
>based on a job listing on Intel's website
Fucking lel.
>>holy fucking shit how do we recover from epyc?
>>>we should just hire a guy and blame him for the company going bankrupt in 5 years

>>61209452
no, graphmeme is still decades away, they built a silicon one

>>61205364

>Four or five years.
>i11
>i13
>i15

>>61205364
>"""""Revolutionary"""""

File: amd1.jpg (71KB, 552x661px) Image search: [Google]

71KB, 552x661px

>>61205364
>Core 2 incoming

File: sanic.jpg (1MB, 3671x2000px) Image search: [Google]

1MB, 3671x2000px

>4 cores mainstream
>3% IPC gain
>4% less power
>12% improved IGP
Revolutionary

>>61209494
>>61209660
Samsung still holds one of the largest, if not the largest, portfolio on graphene related patents, so we'll see. It probably is at least a decade away though.

File: Itanic.jpg (104KB, 672x371px) Image search: [Google]

104KB, 672x371px

>>61210687
>Itanic 2.0 incoming

>>61205364
Why (for that matter HOW) is Nvidia's stock higher than both AMD and Intel's COMBINED?

>>61210745
best marketing would be my guess

>>61205468
>EPYC-FU
I laffed

>>61210745
>>61210754

stock price depends on the number of stocks a company has issued

do you guys not actually know what a stock is lmao

>>61208160
>5ghz
>AMD fags believe this.

File: 1498818178499.gif (2MB, 389x279px) Image search: [Google]

2MB, 389x279px

>>61205364
>5years

>>61212505
4.5 base is realistic
intel will do that with 10nm+++ in 3-4 years

>>61205364
>intel tries to kill of X86 by making a new Itanium
>new arch entirely centered about vectorization
>uses code morphing to handle legacy instructions

>AMD goes full ARM with software emulation for X86 software support

>>61205364
>>AyyyyMD IS A JOKE

intel is building a new core in about 5 years

intel will just be bankrupt by then

>>61205761
why he still calls himself jim keller? we all know he is kevin sorbo

File: arm-windows-10-pc-sdn-02.jpg (301KB, 1000x620px) Image search: [Google]

301KB, 1000x620px

>>61205364
They are attempting to compete with ARM because they know its just a matter of time before ARM fucks AMD/Intel in the ass, Windows PC's were already announced with ARM processors.

This has nothing to do with AyyMD pajeets.

Intel knows this, AMD knows this and most important of all, MICROSOFT knows this.

>4-5 years.
Well it better be photonic or something else just as revolutionary, because in 5 years silicon is about as dead as it can be.

>>61212638
you do know that keller worked a lot on the k12 right? he basicly said so in 2014
k12 is set to be the first high end arm cpu

>>61212713
Your point is?

File: 1499022139550.jpg (9KB, 480x360px) Image search: [Google]

9KB, 480x360px

TWWWOOOO BING BUSSSSS

>>61205364
Pootel can't produce good cpu architectures.

>>61208585
14nm was delayed, 10nm is being delayed and things don't look good for 7nm either. Meanwhile, GloFo is promising a 7nm process(which is better than Intel's 10nm process) in mass production before 2019.

>>61205364
The egos in Intel must be so big that they cannot do anything right without smashing things around. It's going to be a mess in there for a very long while.

>>61208160
And still no IGPU.

>>61212967
Go the poorfag way and buy G4560 then

>>61205641
Don't forget iAPX 432, XScale, i860, and i960

>>61205364
AMD scared them shitless.

>>61208541
>>61208581
>Or RISC-V.
I'm to lazy to read tons of shit to understand what your one about. Can you please explain abit?

>>61205364
>hire poo in loos, independent womin who need no man, and other "diversity genders" because of affirmative action quota and other SJW reasoning
>wonder why your company doing so shit job

>>61210745
What is market cap.

You need to educate yourself.

>>61206542
This is like cherry picking benchmarks.

SeekingAlpha was giving sell guidance on AMD stock while it shot up from 2 dollars to 10 in a week. For every positive article on AMD there are five negative articles at that site.

>>61210736
Why are the propellers on the front?

>>61213458
It trying to fly.

>>61213410
SeekingAlpha is actual unironic stock manipulation. Its one big inside joke. A bunch of their analysts were only being followed by other writers for SeekingAlpha. One giant click circle but for investing. Trying to manipulate stocks and make millions by pennies at a time.

>>61205374
no you're right it's definitely a conspiracy, malice before incompetence after all.

That "article" is fucking nothing. Literal air.

Fuck it would be good if Intel got their shit together. You just know they won't though. Too much love for shorterm shekels.

>>61205733

>windows 10

Yeah because that's the OS everyone has been cheering for these past few years.

>>61208058
What if Intel RnD expenses were simply some form of money laundering scheme?

>>61208058

How do you fuck up this badly? Spending over $12 BN while delivering barely anything?

>>61214294
>>61214339
Intel's R&D expenditure covers their own fabs, modems, and all kinds of tertiary shit from companies they acquire.

File: [HorribleSubs] Sakura Quest - 12 [720p].mkv_snapshot_20.16_[2017.07.01_16.36.07].jpg (65KB, 603x603px) Image search: [Google]

65KB, 603x603px

>>61205364
>only starting now
I wonder how long it'd take to make it to the market and how sore Intel's ass would be after the beating from AMD.

>>61214921
Pink Aoi.

File: A2_Sermiconductors_INTC_2017-RD-Spending.png (240KB, 660x417px) Image search: [Google]

240KB, 660x417px

>>61208058
Most of Intel's R&D goes into non-CPU technology because the CPU business alone can't pay for the fabs anymore. Most of that other business is low margin with lots of competition though.

Intel painted themselves into a corner by not spinning off their fabs like everyone else did.

>>61214974
>Intel painted themselves into a corner by not spinning off their fabs like everyone else did
Intel is literally finished for eternity. It'll only get worse for them from now on.

>>61214974
Man Intel's failed venture into IoT it probably the most hilarious thing of the past years.

>>61214339
intel makes more than processors
Also AMD just buys intels fabs for pennies once intel shrinks their die

>>61215077
The rest are shitty memory and good NICs. That's totally not worth $12gorillion in R&D money.

File: image042.gif (38KB, 425x603px) Image search: [Google]

38KB, 425x603px

The last time Intel designed a processor core from scratch, that wasn't just an iterative P6 design, they came up with Netburst and the Itanic

can't wait to see what they come up with this time

>netburst 2.0
calling it already

>>61215148
Man NetBurst was so bad in so many ways.
What were they thinking designing that?

>Itanium II right after Prescott 2.0

History truly is repeating itself.

>>61215094
>server assembly
>fiber optics
>Wifi
Plus they're the only other company other than samsung who actually Pioneers advancements in SSD tech

>>61215223
You totally forgot Toshiba but whatever.
>>61215204
He he they need another P6 derivative to save their ass.

>>61215259
Real question would be if they Return the Pentium to prominence.

isnt it kind of pointless without a graphene-base to work with?

>>61215281
The real question is in what way will Intel fuck up their new uarch.

>>61215281
never.

>>61215335
moar hardware drm

>>61215358
Who cares.
I mean performance-wise.
Bibelines? Moar housefire-enabling vectors?

>>61215378
Intel Pentium Neo: Elephant's Foot Edition

>>61215335

Mini-ITX sized processor dies.

File: 1272121144334.jpg (76KB, 640x480px) Image search: [Google]

76KB, 640x480px

>>61215168
MOAR GIGAHERTZ

>>61215335
All processors welded into mobos. This is good for the customer you see because then they need to buy only one part instead of painful research into sockets and whatnot that fucking nerd stuff. You red it here first.

File: 1178120120185.jpg (416KB, 1386x2070px) Image search: [Google]

416KB, 1386x2070px

>>61215168
RAMBUS. Yes, RAMBUS. A RAM type that upped the megahertz and lowered the pin/trace count. Its problem was high latencies that kicked P6 in the nuts. So Intel set out to design a CPU that could tolerate high RAM latencies. It became the Pentium 4.

>>61205364
right now what Intel needs is better interconnect, not a new uarch

>>61215648
Ebinmail will save them (no)!

>>61214974
>mobile down 55%
Intel failure

>IoT up 80%
failure to happen, they're arriving in force extremely late to the party

>non-volatile memory up 40%
Intel failure that already happened, but they don't know it yet.

>data center up 25%
The only sensible investment in that whole picture

>>61205468
>EPYC-FU

I'd buy that if they actually used it as a product brand.

>>61215168
>What were they thinking designing that?

AMD was the first to break the GHz barrier, they needed something to counter that.
That, plus RDRAM.

It actually wasn't that bad early on; what I really wonder is how did they fuck up Prescott so bad.

>>61215648
It'll be interesting to see how well the mesh will work after Intel irons out the bugs in it.

>>61215693
> It actually wasn't that bad early on
Yes, Northwood was pretty fucking decent.
>I really wonder is how did they fuck up Prescott so bad.
Somebody get me that pic.

>>61215476
No, they'll just put HBM next to the die. That way you won't need to worry about RAM compatibility and installing additional memory sticks. You just plug in 1 thing and you are done.

It also means you need to buy an entirely new cpu if you want to upgrade your memory. i3 with 4gb ram, i5 with 8gb ram, i7 with 16gb ram.

Tejas and its hundred step pipeline will finally return at 7 nm!

>>61215701
won't matter if they still need to fab massive dies to compete with EPYC. Intel needs MCMs.

AMD just needs to glue zeppelin dies together and they have over 80% yields on those. if 7nm fixes the thermal wall issue then there won't be a single Intel product winning in its price range.

File: 1435497588761.png (973KB, 801x1500px) Image search: [Google]

973KB, 801x1500px

>>61215708
This? :DDDDDDDDDD

>>61215788
Yes!
BIIIIIIIIIIIIIIIIIBELINESSSSSSSSSSSSSSS :--------DDDDDDDDDDDDDD

:v)

>>61214974
>IoT up 80%
HOLY MOTHER OF KEK

>>61212789
point is amd is making high end arm cpu. So AMD is kinda related even if point is intel is getting cucked by ARM, but AMD actually will continue compete with its own custom ARM K12.

>>61213671
>no you're right it's definitely a conspiracy,
as if we've forgotten the bullshit intel has pulled before.

good try kike.

>>61214974
they are literally bleeding money at this point, by 2021, intel might actually be bankrupt.

>>61216420
Hell no they dont, but they are big meaty pinyata for AMD to punch and get money from.

>>61216420
Nah. If AMD didn't die from Bulllozer I doubt Intel will be dead and bankrupt™ in 5 years.
They might spin out their fabs though.
Though on the other hand AMD didn't have retarded shareholders that basically required them to maintain 60%

Unless its a photonic based cpu using the calabi-yau manifold no one will cares.

>>61205364
>crank up the fabs
>same core with higher frequency
>Codename Gehenna
>Watercooler mandatory
Watch out AMD, the kikes are onto something

File: new_intel_gehenna.jpg (30KB, 474x356px) Image search: [Google]

30KB, 474x356px

>>61216732
Forgot pic

>>61216732
That's i9. Not their new core.

>>61216838
No.
This will be the new i11 FireLake.
Screencap this.

>>61216838
> Introducing Intel Core(R) i12
> River Styx architecture
> Fluid cooling integrated into die
> 4c/8t running at 7.6Ghz
> Requires Intel X999 chipset
> VRMs are modular and requires replacement approximately every 3 months
> CPU no longer uses PSU for power, plugs directly into 110/220V mains
> CPU lid is glass and uses patented technology to produce light without requiring LEDs

>>61217042
Kek

>>61214130
Windows 10 was made to run on super low compute devices, you can also optimize it further yourself by disabling a bunch of the useless services.

>>61207367
I'll believe that kind of thing when fabrication techniques make micro-channel direct-stack water cooling a reality. Some sources have been saying Samsung is finalizing early GAA designs for their 7/5nm position, we will wait and see.

>>61217042
>Not x666

>>61217042
> CPU lid is glass and uses patented technology to produce light without requiring LEDs
> Not Aerogel, so it doesn't leak heat out of the package

>>61205364
This is why I wanted Ryzen. Now these Jews are actually frightened and they will put some work back again. Thanks AMD. Competition greatly benefits end users.

>>61217810
That bad part is that every actually new Intel uarch ended up a disaster.

>>61217838
Shit happens, but CPU architectures tend to improve a lot unless you make a downright pile of shit like Netburst or Bulldozer.

>>61205641
Atom phones were genuinely good, though. Intel would have dominated the market if they had taken it seriously in 2010 instead of waiting half a decade.

>>61205820
This is objectively wrong. If ARM was more efficient, you'd see it deployed in supercomputers where individual core speed is irrelevant (because you can always add a few thousand more) and efficiency is paramount. Just because ARM is only used in low-power devices doesn't mean it's inherently more efficient.

>>61205929
Explain how I'm posting from (desktop) Windows on a Cherry Trail device, then.

>>61217867
No, i mean every new Intel's attempt at making something really new (Itanic/Shitburst/Larrashit) ended up being a disaster.

>>61217891
>Larrabee
>Shit
Neck yourself, faggot.

>>61217960
>not shit
Come on faggot.

>>61206278
>seeking alpha

>>61218227

I agree, they are nowhere near as reliable as Motley Fool.

>Intel has provided increased performance while keeping power consumption in check.
"No."

File: DD2qQLnUwAEUnh6.jpg (70KB, 581x974px) Image search: [Google]

70KB, 581x974px

>>61214974
>Intel spends ~$30B on pic related

>>61205364
The core redisign is for a while. It's a good time to buy AMD because they're almost on par with current Intel on an IPC basis and they'll be on AM4 for a while, but this new core from Intel will be a nice leap up from the AM4 CPUs on the 14nm process whenever it finally comes out

>>61218510
https://www.youtube.com/watch?v=oE0sddhCIdE

>>61220132
Isn't Zen 2 going to be 7mm already?

>>61220946
Yes.
They wre going to have a Zen+ at 14nm, but the IBM process development is advancing so quickly that they're going straight to Zen2 at 7nm.

>>61212713
K12 was shelved m8

File: dis gon b gud.gif (973KB, 350x188px) Image search: [Google]

973KB, 350x188px

>>61220966
So... AMD is going to hop into 7nm before Intel even has (good) 10nm? And it's going to improve it by 2020 (roadmap)?
dued

>>61221094
But attention. Zen2 will be with a supposedly "inferior" 7nm process (DUV). The real fun starts after that with 7nm EUV.

>>61221117
7nm EUV (first gen of it at least) is in no way better than 7nm DUV. Just less complex litho thus better yields.

>>61205364
>yfw they copy the infinity fabric

>>61221241
Just like they did copy HyperTransport.

>>61215012
>paint themselves into corner by spinning off fans
Why would spinning it off save them? Is the fab business bleeding money? How would they have their chips manufactured without their own fabs?

>>61218510
and diversity.

>>61221281
Because then Intel would have somebody that wasn't them and their retard architects to blame for shitty yields.
And yes, fabs are money pits, and Intel won't lease their fabs out to anyone else.

>>61221189
>Just less complex litho thus better yields.

DUV/EUV trades tons of extra masking steps (quad patterning bullshit) vs. more painful optical setup (pulsed tin droplet vaporization lasing, liquid cool dielectric mirrors, etc).

It's hard to call one way unambiguously more/less complicated, since # stages and equipment complexity aren't exactly directly comparable.

>>61221448
I don't think it can be any worse than quad motherfucking patterning.

>>61221094
>>61221117
GloFo 7nm isn't actually 7nm or anywhere near close

Sadly you are proof that their bullshit works

>>61221550
Hello, Brian! No, BEOL scaling is totally not everything node shrink is all about.
Sometimes you want that juicy power target.
And 7nm LP is denser than Intel's 10nm anyway.

>>61221550
It's GloFlo 7nm that is actually a 9nm or so compared to Intel's process.
Every foundry has their own measurements.

But GloFlo's 7nm (DUV) is supposedly slightly better than Intel's 10nm.
Nothing has been said about 7nm EUV, but it's supposedly better than 7nm DUV, so it should be better than Intel's 10nm too.

>>61221585
You're telling me intel has a worse fab than anyone else? Highly unlikely man

File: screen-shot-2012-05-14-at-9.35.09-am.png (123KB, 547x261px) Image search: [Google]

123KB, 547x261px

>>61221550
10nm by 2015.

the author is a well known diehard intel shill

>>61221616
They are about to lose their fab advantage. Deal with it.

>>61221616
You won't believe me how awesome Nazi IBM wizardry is.
>>61221615
GloFo's 7nm DUV is 7.9 AFAIK.

>>61221632
https://www.fool.com/investing/2017/03/06/intel-corporation-expresses-doubt-about-10-nano-ra.aspx

Well that was a worthwhile 2 clicks to see what else the guy wrote.

>>61221585
>>61221615
I literally insult people and they reply like this

what's going on with /g/

>>61221616
That's exactly what's happening. Intel stayed for far too long on 14nm while having a shitload of problems with the 10nm process. Be reminded, Cannonlake with 10nm was supposed to have been on the shelves one or two years ago.

They've lost the big advantage they had. And it's even more problematic for Intel since GloFlo, Samsung and IBM decided to join forces to fuck them over.

Intel needs to get their 10nm out of the door fast and without problems.

File: the call that saved amd.jpg (60KB, 1000x581px) Image search: [Google]

60KB, 1000x581px

>>61213671
>no you're right it's definitely a conspiracy, malice before incompetence after all.

Yep, intel just out of the blue decided to make a new arch

>>61221711
Quick rundown?

>>61214974
Wow, desktop PCs have really become a niche.

Anyway, how can they spend tens of billions of dollars on research and still suck? This is beyond me.

You spend 30bn and only get small, marginal increments in performance?

>>61221750
It's because they dragged their feet on making a new architecture.

>>61221241
Good luck with that. I won't be surprised if Nvidia will manage to create Fabricworks™ before Intel.
Inb4 Intel combines bingbus and infinity fabric into some unholy abomination that just Werks.

>>61221750
Their fabs are still >5 years ahead of TSMC and the Common Platform, who are in turn >5 years ahead of anyone else. They're essentially subsidizing the development of production machinery for everyone else.

>>61206513
My E2-3200 is still working great.

>>61221810
Wut? Intel is years behind Common Platorm right now. Especially when it comes to R&D.
>>61221809
NVIDIA will face a truckload of troubles since they have zero experience in dealing with complicated NUMA designs.

>>61221750
they threw away their baytrail to chinese to get a foot on the tablet market. some neckbeards got it and bought a couple of teclast windows tablet. nice and eveything but after two years they are starting from scratch again, those teclast are now old, and since they stopped subsidizing chinese tablet makers with cheap atom they basically threw away their money in their previous attempt.

they also stopped developing their atom lineup, so well, this is even worse from an investing perspective.

they are now leasing some of their fabs to some shit tier chinese arm chipmaker, but never heard anything more out of this.

they pledged alliance with other chinese chipmaker in hope to reduce cost and subsidize the developing of some of their model/lte atom chips for phone, and their sofia iot platform. again little results i suppose.

their croatian ceo seems like is bunching from project to project with not clear stategies and just by throwing an awful lot of money at everything that could grow. unfortunately they missed their biggest train when smarthphones rised up and they kept jerking off to the awfully absurdly large margins they had on their desktop but more importantly server and laptop cpus.

>>61221491
they literally can't get power up high enough, which makes shit take longer too.

> have to spray microscopic molten tin droplets every few milliseconds, then zap them in fucking midair to get EUV pulses
> normal shit can't reflect EUV, so you need special mirrors made of dozen of alternating layers of thicknesses perfectly one half the fucking wavelength of the already short wavelengths
> those unobtanium mirrors still only reflect something like 70% of the light that hits them. each.
> they absorb so much of the light that they need water cooling to not have their surfaces ablated away immediately
> and they still wear out pretty fast from just having that much low-wavelength radiation pounding them

I'm sure they'll get something to work well enough eventually, but it's a complete clusterfuck of engineering problems.

>>61221889
It's been years and we're only starting adoption of EUV. Yes, it's ass. Yes, even ASML knows it's fucking ass right now. But what can't be fixed with fuckton of engineering man-hours?

>>61221889
Just use "conflict materials" and you're good to go.

>>61205743
>"The pussie destroyer"
>Not "Certified Shit Wrecker"

You had one job anon

File: they-dont-think-it-be-like-it-is-but-it-do.jpg (166KB, 523x720px) Image search: [Google]

166KB, 523x720px

>>61221889
>that technobabble
>wait, it's real

>>61221860
>bunching from project to project with not clear stategies and just by throwing an awful lot of money at everything that could grow.
This seems to be common in super large corporations. Microsoft and Google seem to have the same problem, they dump money into shit that goes no where and then cancel anything that doesn't immediately take off.

>>61207704
>Compaq into killing Alpha series
;_;7
I still cry everytime
Such a great architecture completely gone
Now we are stuck with this X86 bullshit

>>61205784
I guess he must have smelled the desperation during it, which is why he ended up at Tesla instead.

>>61222033
You can thank our favourite merchants at Intel for coercing Compaq into adopting Itanic and killing Alpha.
Fuck Intel.

File: amd great again.png (418KB, 627x627px) Image search: [Google]

418KB, 627x627px

>>61205364

>>61221889
It just blows my mind that blasting tin droplets in midair with a high powered laser is the preferred way to even start any manufacturing process.

>>61222041
this has been posted on /g/ before, but if you are a CE person, this is a great watch
https://www.youtube.com/watch?v=klg1FtHADso

>>61205641

We should also mentions AMD's lulldozer as a failure. It seems to me that the ones who stray away from P6 too far get roasted miserably. Programmers optimized desktop programs only for this fucking architecture.

>>61206278
>seeking alpha
unpaid interns
t. finance

>>61205364
>Ashart Outassa
America is really cucked

Is intel a CIA project?

This is based on a job listing?

Well no shit Intel talks about projects. All of these companies have roadmaps, something down the road

But first we need something tangible or it's just another Larabee

>>61205364
>inb4 it's Intel's bulldozer

>>61205364
nice propaganda lol

I'm guessing some Altera FPGA on the same die as Intel CPU much like Xilinx Zynq, but with much more powerful CPU.

Very few people will figure out how to fully utilize the FPGA function, and by the time people do, Intel management will panic and just kill the feature instead of letting it breathe find an audience.