http://www.pcworld.com/article/3114 710/hardware/amd-thinks-

Why even run an older version of Windows on brand new hardware? It's a consumer level OS, I can understand different Linux Kernels, but using an old version of windows, even when they implement the same >botnet level features as the new version, what's the point?

AMD doesn't operate a fab.
They're using Global Foundries who acquired IBM's foundry business, all their process IP, and engineers.
IBM developed a bleeding edge 7nm process that GloFo will bring online at the end of next year. IBM's process IP is uniquely distinct from intel's in that is is not bulk silicon based.

Kaby Lake, Summit Ridge, and Raven Ridge are not locked into Windows 10. This is pure clickbait. What is happening is that Microsoft won't have support for new pstate models in older versions of Windows. You can install any OS you want on them, you just won't be able to take advantage of their new ulta low power bullshit for thin laptops.

>>56381920

>Ignore that Zen sucks, trust us, the future is going to be great!

So the AMD cycle continues.

>>56382043
So Zen+ will be okay thanks to IBM and theres no reason to waste on Zen?

>>56382043
>You can install any OS you want on them, you just won't be able to take advantage of their new ulta low power bullshit for thin laptops.
This, and in all likelihood I expect them to update 7 and 8.1 to detect these future CPUs by product ID. As long as you don't update, and anyone with a brain has stopped updating since they tried to add telemetry, they can't do shit.

>>56382073
Theres no official word on anything, but I'd bet Vega money on Zen+ based parts utilizing IBM's 14nm SOI process. Higher frequency at a given target power, not sure on the area scaling, but they'll likely be in the same ballpark. The Zen+ equivalent of Summit Ridge could hit 4ghz with the same 95w TDP.

It would be unprecedented for 7nm to come online and start churning out desktop/enterprise chips for sale in 2018. Regardless Zen+ based parts will not be using 14nm LPP from Samsung, and they will be substantially higher performance. Greater IPC and greater frequency.

>>56382177
Does this mean 2019 for zen+?

quantum leap sure...

File: XeonD perfwatt.png (27KB, 673x294px) Image search: [Google]

27KB, 673x294px

>>56382177
>Vega money
Meant Vegas, but that typo is golden.

>>56382298
If AMD wanted to wait to release Zen+ on IBM's 7nm that would likely be the case, but I don't see that happening.
I expect Zen+ based parts to come out in 2018 utilizing IBM's 14nm SOI FinFET process. Its the process that IBM's own POWER9 chips are being fabbed on.

The arch that comes after Zen+ will be the 7nm part. Zen++ or whatever naming convention they'll switch to.

File: Intel-Coffee-Lake-Roadmap.jpg (58KB, 635x255px) Image search: [Google]

58KB, 635x255px

>>56381920
don't worry so much, Kaby Lake is garbage and Intel won't have 10 nm desktop chips until late 2018 at the absolute earliest now.

>>56381920

Now it's my turn to open another thread with the same subject.

>>56380248

File: 1457055053671.png (29KB, 480x480px) Image search: [Google]

29KB, 480x480px

>7nm AMD chips aren't seen for at least another 3 years
>mfw it's another fake process node just like 14 and 16nm FinFET leaving feature size almost unchanged

yawn

Oh look it's another cycle of amd hype that's destined to end in failure.

so AMD isn't even going to beat Intel's 4-5 gen old parts in performance and won't be releasing the successor to Zen until after coffeelake, cannonlake and icelake?

kek. AMD are truly masters of squandering every opportunity they get.

>>56382857
Its amazing how you managed to pull all of that out of your ass.

>>56383022

read the OP idiot. no zen+ until 7nm means we won't get zen+ until 2019 at the earliest.

>>56383069
I don't need to read anything, I asked you a question, since you're clearly clueless. I explained this plainly a few posts above.
GloFo is bringing IBM's 7nm node online at the end of next year. Zen+ based parts are likely to use their 14nm SOI FinFET process, not their 7nm process.
No one is in anyway implying that Zen+ based parts will be delayed until 7nm starts yielding at volume.

Why are you conflating IBM's process technology with bulk solutions from Samsung and TSMC?

>>56383108
>GloFo is bringing IBM's 7nm node online at the end of next year.

yeah, that might have been the plan a few years ago, but the massive delays to 20nm and 14nm have pretty much ended the feasibility of that. i'd expect 7nm to delayed to 2019 at least just like it's predecessors were, anything else is just wishful thinking or outright fanboyism retard logic.

>>56382857
>>56383069
anon, if you think Intel is getting their non-mobile 10nm chips out much before 2019, you're in for a rough surprise:
>>56382670

also,
> imblying coffeelake and icelank won't be as much shit as skylake and kaby lake

>>56383141
>I have no idea what I'm talking about: The Post
Why would you just blindly post your ill informed guesses as if they have any value whatsoever?

>>56381920
Next week they'll announce that the chip comes with a complementary BJ. I'm not buying any of this.

It's time to nut up or shut up AMD. Ship it and stop fucking around. All this PR bullshit just seems to me like they're hyping up a product they know sucks ass.

>>56383197
>ill informed guesses

How is it a guess? GloFo has been underachieving for over a decade now. If you think 7nm won't end up delayed you are living in some schizo bernie land where everything is free and perfect.

>>56383268
>triggered /pol/tard

>>56383268
Global Foundries *is* IBM's foundry business.
If you can't wrap your head around this stop talking out of your ass. Putting for a guess then asserting as fact at the same time is the definition of living in a schizo bernie land.

>>56383300

What's your point? IBM doesn't do anything innovative when it comes to fab tech. They do the bare minimum required for them to continue producing their housefire enterprise CPUs.

GloFo will most likely be stuck licensing stolen TSMC tech again for 7nm that is destined for consumers, because anything IBM makes will be geared torwards large dies/high clock speed/high leakage.

>>56383336
>IBM doesn't do anything innovative when it comes to fab tech.
>ultra thin body SOI
>SOI based FinFETs
>first commercial ready process with a SiGe channel
>hurrrr IBM doesn't do anything innovative guys

Sounds like you're just a shitposting fanboy who doesn't understand what hes talking about.

File: projection.jpg (36KB, 800x533px) Image search: [Google]

36KB, 800x533px

>>56383374
>anyone who disagrees with me is a fanboy

>>56383398
>oh shit I just got proven wrong
>do I man up and admit my mistake?
>no
>I better shitpost and try to save face
Denying reality to fit your narrative is what makes you a fanboy. Discrediting a company's achievements because they're inconvenient to the position you're trying to argue is what makes you a fanboy. Fanboy.

>>56383430

i don't see any proof in your previous posts, perhaps you should cite some sources buddyboi

>Windows

>>56383445
>still just trying to save face while shitposting
>demonstrably doesn't understand burden of proof fallacy

You made the principle claim that IBM doesn't do anything innovative when it comes to semiconductor fabrication.
I just provided numerous examples of why you're wrong.
You want to keep making the same childish argument, the burden is on you.
You won't though, since you're just a shitposter.

>>56381920
Zen isn't Windows 10 only. Microsoft just won;t be adding kernel support, and cpu manufacturers etc won't be writing chipset software for anything else.

Don't fall for disingenuous clickbait.

>>56383473
>I just provided numerous examples of why you're wrong.

all you did was shitpost some buzzwords, post some real sources AMDrone

>>56383268
>schizo bernie land
Another sad case of a person brainwashed into thinking he has an opinion about things he doesn't even understand.

Keep being a useful idiot. If I could gut the media and use their hollowed out shell to begin drip feeding a re-education campaign, I'd use you for all you're worth too.

>>56383519

so you're a welfare leeching bernie supporter then? no wonder you think AMD can succeed. red must be your favorite color.

File: dynamic power.png (91KB, 1134x778px) Image search: [Google]

91KB, 1134x778px

>>56383495
>desperate, butthurt, tech illiterate fanboy retard
>still, even now, trying to save face by continuing to act like a child
>resorts to throwing out meme insults like AMDrone
>when discussing IBM process IP

Do you realize that you're too stupid to take seriously?
A veritable subhuman.

>>56383568

>reduced to typing out inane insults
>still can't cite any sources

AMD loves welfare-subsidized marketing staff!

>>56383555
>so you're a welfare leeching
No. But thinking of the bigger picture I don;t really mind people sucking a broken system dry.

>bernie supporter then?
Yes. Most of his ideas were the only viable way forward, taking into account the entirety of human history and what we've learned simply does not, and will not, work.

>no wonder you think AMD can succeed.
I don't think anything. That Keller worked on their new chips is a good sign, Global Foundries isn't.

I hope they succeed.

>red must be your favorite color.
I don't have a favorite color, comrade.

>>56383590
You have the source right there, if you're that retarded you should go back to shill on reddit
You didn't even provide any argument in any post

>>56381920
>Intel's Kaby Lake and AMD's Zen processors will only support Windows 10
what a FUD...
https://blogs.windows.com/business/2016/08/11/updates-to-silicon-support-policy-for-windows/

10nm is a bullshit process like 20nm that every major player will drop. AMD is not alone.

And I don't see how it matters if it only supports win 10. Using win 7 and cutting edge hardware is like using a V12 engine mounted in a rusty VW beetle.

File: longD.png (756KB, 860x793px) Image search: [Google]

756KB, 860x793px

>>56386773
nigga that analogy sucks

do you know how fucking fast a v12 engine in a beetle would be?

it would be prone to crashing and a deathtrap but it would be sanic speed