7nm zen 2 right around the corner

Don't tell me you don't get a little wet at the thought of installing Gentoo in mere minutes.

the thought does make the dingle tingle tbqh

Who cares
A 30nm haswell has the same single core performance as a 12nm poozen

Intel is dead in the water until they can get 3D stacking to work in consumer products. AMD just pulled ahead super hard, and Intel is too proud to ditch their own fabs. They're the ones in the chase now, and the next big leap after 5nm (which TSMC will probably reach before Intel, assuming the current trend continues) will have to be 3D stacking.

Hopefully this takes Intel down a peg and makes them actually get off their asses and innovate once again. And hopefully it also gives AMD a nice money cushion so they can keep Intel in check and not repeat the entire Bulldozer era again.

>EUV 7nm Zen 3
How big will be the jump compared to current 7nm?

Helo Gustafsons law

>still thinking AMD has any chance of releasing good products
How many times do we have to teach you this lesson, old man?

Impossible to say what the arch itself will change, but on the process side its a moderate reduction in power, moderate uplift in performance, above all its a cost reduction vs current partial immersion patterning.

hard to believe. while bulldozer was a new design with the main work on an new philsophie, zen2 will base on an good foundation and just improve things more.

this is why first ryzen mainboards were shit and the ones we can buy in 6 months will be much better. dont let you fool the last processor generstions from intel were basically all the same, so the mainboards were lo g day fi ished and they had a lot of time to polish the details.

same is happening now at amd.

its like expecting that the samsung galaxy s11+will have worser
stats than an s10+.

Will it have good overclocking? Literally the only aspect in which Intel is still ahead, if the 7nm Zen chips can get close to 5ghz then Intel will receive its final JUSTing.

>Zen 2 will be followed up next year by EUV 7nm Zen3
How about no? It will last for 2-3 years, AMD will make hi-end navi with 7nm EUV first.

it'a bulldozer all over again

Attached: cunt.jpg (1160x984, 308K)

Attached: 1504275173842.png (480x521, 120K)

Literally no data center uses amd though

>trying to figure out if i should build the pc now iwth x470 and just drop in ryzen2 when it comes out or wait another 2-3 months or however long its gonna take for amd to drop it and get new gen mobo alongside it

god damn amd, I'm stuck with this laptop and I want the full desktop now.

I predict 10% higher clocks so 4.6 or 4.7 GHz

Nobody buys Intel.

Only on 5nm dual socket in 2021

This.

>Zen3
they will do one or two refresh and then go die stacking like Intel

I still wonder how die stacking isn't going to cause heat pockets.

>I know exactly what AMD is going to do in the future despite not having disclosed any of this information to investors
Sure thing, kid.

It is. You need insane thermal conductivity and insanely low thermal density ICs for conventional logic stacking to be possible. Intel is only getting away it it by using very small low power cores to supplement their main big cores. This isn't true stacked high performance logic.

Zen 3 was never talked about once
They made it clear that 1 and 2 were different approaches
Its fair to assume they will tweak this designs couple of years instead of pulling something never mentioned out of their ass

>tfw bottlenecked by io when installing gentoo
not sure how i feel desu

Attached: how_the_shinto_exorcise_daemons.jpg (600x776, 96K)

Additional cores don't scale that high. Sorry

Papermaster confirmed at CES that Zen 3 was complete, and Zen 4 was being worked on. As for your stupid assertion, it isn't fair to assume that at all when you have literally no idea what the executives at AMD have planned.

>still objectively worse than intel

Could next year Zen have DDR5 and PCIE5 support?

It can be a zelda person too.

I believe AMD said Zen3 would still be socket AM4, so no.

>7nm amd will barely match Intels 14nm
LOL!

Attached: 1519154211409.jpg (568x612, 66K)

Attached: hell.jpg (500x500, 208K)

You mean 14nm+++++

PCIE 4.0 effectively doubles your available bandwidth, and in most practical use cases it will double available lanes since virtually everything can just be run in 3.0 mode without impacting performance.
PCI-E 5.0 will not be supported by anything or even utilized outside of some obscure HPC configurations for years to come.
Also DDR5 won't have market penetration until 2021, so wanting that is just as fucking retarded.

can somebody tell me if Zen 2 will have better single core performance than the 9900k?

what ever,no different for old ryzen.
and more and more intels cpus going win,wait end of year.

tdp is just that high what is should be with more cores or clocks...none upgrade.!!!

looks same effect ike gpu sector, fury,vega and radeon vii...no different
well radeon vii was 7nm gpu, but not belive it! so lausy efficiency,and still slow!

ok, cputdp also same...there is anyhting new,just different name....i wonder how long amd can sell product with no profit..well looks that no longer...sooner or later its eat hole company and debt givens not tolerance more...end of 2020 is my guess.

hmm. i try really see different from ryzen 2000 series,but no....oh yeah... more cores..but is it only trend or commercial way.
you dont need more than total max 6 cores for gaming,even 4-core is enough. easily.
more than that is waist of cash, no matter how much its cost.

and,i still see not any goodies from amd 32-core cpu...erhh i mean where to use thouse??!! tell me..

200-cores cpu.... so what!

no help gaming,apzz or rendering...useless.

another hand, real pro rendering use always intels cpus...why??

well,buy them amd fans, you keep amd up and pay them callary...youself you give so little back.
except red fine main...lol

Attached: 1523382018743.png (408x450, 35K)

Anyone who tells you anything other than "we can't know for sure until the CPU comes out" is bullshitting you.

Zen2 will be competing with whatever lake comes next, not 9900K

No, they would have shown benchmarks at ces otherwise, like they did with the multicore performance.

Skylake 6: Electric Boogapooinloo

based inteljeet pasta

UMA DELID CIA

Ehm, enjoy your obsolete hardware
for me its Zen 5 + DDR6 + PCIe 6.0

2020

"only" 128 this year :(

I think they mentioned about the possibility to have two of those on the same motherboard, though.

I think Zen3 should be the last one on AM4 no? I've had my i5 since 2016 and am curious about getting the best socket from AMD depending on how things go in my life.

Sadly Nvidia did shit the bed with its rubbish new RTX line-up of gpus since I give 0 fucks about more pretty fire reflections. I just want better performance overall so my 1080 will have to keep chugging for a but longer.

What did he mean by this?

Isn't EUV only supposed to increase 7nm yields rather than bring any significant change to the technology itself? As in, it'll be still the same 7nm just cheaper to make.

>I think Zen3 should be the last one on AM4 no?
I don't think anyone can tell for sure. AMD mentioned support until 2020 which doesn't necessarily have to include 2020, that means Zen 2 could just as well be the last AM4 gen.

Probably Zen2 refresh to be honest. For Zen it was 1st gen in 2017 and then zen+ in 2018. You can probably expect a similar release cycle with Zen2 with 1st gen Zen2 this year and a refresh the year after that. After that it's onto a new socket.

That makes sense in the light of DDR5 introduction in 2020 so it could mean AM4 successor comes out that year. They could pull AM2/2+/3 shit with AM4+ and new CPUs having IMC capable of both DDR4 and DDR5 but why would they really. Sure keeping platform alive for a long time is a noice move towards customers but not necessarily giving them a good standing with their business partners who want to be able to sell more boards and new PCs.

There is supposed to be a 15% area reduction, 10% power reduction, and possibly a 0-10% perf uplift.

I think 3d stacking is Intel's response to AMDs APUs. Integrate as many of a PC components in a single low power package meant to be a jack of all trades.
If you're looking for performance you'll probably have to stay with good old 2d layouts, since at least they wont be heat traps. But if Intel can do 3d, chances are they can and will make 2d in a similar way AMD is doing now.

>I think 3d stacking is Intel's response to AMDs APUs.
But intels are already housefires as it is, won't adding layers just cause them to melt from the inside?

Appreciate Intel's research for consumer fusion reactors, goy.

>10nm
>high clocked large dies
pick one and only one

Attached: tumblr_m3m157ibcf1qe8bfro2_2501.gif (245x314, 825K)

i will never not laugh at this

Attached: 1541538120178_0.jpg (818x693, 91K)

>buy Intel
>suddenly care about socket comparability
based blackpilled poster

>Hopefully this takes Intel down a peg and makes them actually get off their asses and innovate once again.
bwahahaha. not this fucken company. not this time.

Attached: dims.jpg (1600x917, 205K)

>You mean 10+++
Fixed your mistake

What it they don't want to hype people and then blame Amd for it?

> 10nm fails
> Intel releases a 7nm generation using TSMC
> AMD is ded again

Shit will get more complicated now, on a market per market play.
AMD beats intel on laptop, intel beats AMD on server, AMD takes the server back, intel takes the desktop, AMD takes consoles, intel take portables and it keep going into that.

Apparently, they will be clocked lower, but having an enormous cash of memory close gives more performance

>AMD beats intel on laptops
nope. maybe for low power chips and ingegrated graphics, but intel is the only one that has 6-core 45W chips with the best raw cpu performance on the market. Intel has 99% of the laptop market.

I don't know what you mean with "intel beats AMD on server", Epyc is pretty much destroying Xeons on price/performance, and that 64 core zen2 is really promising

>intel take portables
i thought tegra and snapdragon was powering most portables these days?

I read the 15% area reduction is a certain. I don't know if amd won't cheap out and not use that again like they didn't with 12nm. It reads a bit like 7nm vanilla is a bit light on what it's capable of because of the wait for EUV.

Intel's current 3d stack chip is seriously low power. Like below 10w and I don't think anyone knows who the customer for it is yet. My prediction is those tiny stick pcs that plug into hdmi ports when I saw the demo pcb. It won't have enough power or thermal limit to allow it to compete with an actual performance chip (15w+).

It's a joke, but with Keller and Raja at the top with probably unlimited resources they'll pull it together in time. It depends whether it's the first (released) 10nm process or the one after that fully fixes it. It's still huge monster of a corporation and losing market share won't kill it unfortunately.

Intel's low supply compared with 3rd ryzen mobile fixing those power idle issues is actually predicted to have that change this year.

It's just a random pick to show the spirit of the thing, not necessarily the things that will happen.

>It's a joke, but with Keller and Raja at the top with probably unlimited resources they'll pull it together in time. It depends whether it's the first (released) 10nm process or the one after that fully fixes it. It's still huge monster of a corporation and losing market share won't kill it unfortunately.
not on its own but when you drink your own koolaid you tend to fuck yourself up, and intel is clearly not what it used to be before they got all virtuous and diverse.

Zen 2 isn't that big of a jump to the low-midrange though
The 1700 and 2700x are very competitive.
I think adored hyped u lot up too much hence his latest video where he's more realistic.
I think the chips will top out at 4.5-5ghz at significantly reduced power latency and has vastly better arch 25% perf bump across the board is the best case scenario

>on the next generation, PS5 goes with AMD again, Xbox NaN goes with Intel thanks to Keller and nintendo stick with Nvidia

I suspect there's been some internal restructuring, especially with those two being hired and given pretty up top positions. Unless raja is just spending everything he can on making new facilities in india and killing intel a little bit more then breaking away for ati2.

If Zen 2 gets single thread performance comparable to an 8700K and at similar launch prices as last time, they will have a real winner. Reason is gaymers really care about that single thread perf even though for things like video encoding the overall numbers won't be that huge of a difference

So a 2700x oc then? Dunno what benches your looking at.
5% slower is comparable performance
T. 2700x owner

Nah bro, dual ringbus house fires with 2 soldered off cores are the way to go.

Attached: 1528277581392.jpg (1072x650, 126K)

There's no way TSMC can expand to fulfill a contract with Intel along with all their other clients. At that point Samsung would be the only other major fab.

amd would have to shit face fuck itself for that to happen. zen 2 is just a refinement of zen 1 from everything they learned and new improvements and features. it be a monumental achievement of fuck ups to fuck that up when all they have to do is improve upon zen 1. with zen 1 being 7-10% average behind skylake in ipc, zen 2 matching or slightly (+2-4%) surpassing skylake is yugely possible and quite frankly guaranteed unless amd some how went full autistic.

zen 2 should be pretty guud and intel should be terrified and get their 10nm working. and finally release a new processor that has ipc greater than skylake. as everything since skylake has been... skylake.

>Intel should ditch their fabs so white people lose their job and chinks get even richer
ok Chang

Intel 10nm will never reach bulk production.

nice cope lol

TSMC advertises area and power gains.
Plus better uniformity and less anal design rules.

No, and most definetly no.
PCIe5 won't be coming to consumers for the longest while.

>At least you know what you're talking about

"The PCI-SIG expects to ratify the final 1.0 revision in the first quarter of 2019, and the first PCIe 5.0 devices should debut this year. Broader availability should come in 2020"
tomshardware.com/news/pcie-4.0-5.0-pci-sig-specification,38460.html

>consumers

>Broader availability should come in 2020
>PCIe5 won't be coming to consumers for the longest while.

Also you seem to forget lots of mobo manufacturers release early boards with draft specifications which later update to the final. it happened before, it will happen again. so yeah expect to see the first PCIe 5.0 boards appear later this year.

You're not getting it.
PCIe5 isn't designed for consumer applications, it requires a plethora of retimers and new PCB materials to just function properly.

>Intel is dead in the water
Intel will never be dead in the water.

Moving esxi clusters to a different architecture is fraught with too much uncertainty.

WINTEL will never die, because it’s reliable.

No it's you who is not getting it, and you insist that PCIe 5 is some alien tech designed for exotic HPC. first PCIe boards are out in late 2019 and that's that.